Article 4981 of alt.sys.pdp10:
Path: news3.best.com!news2.best.com!feed1.news.rcn.net!rcn!chnws02.mediaone.net!24.128.1.101!chnws05.ne.mediaone.net!24.128.44.7!wbnws01.ne.mediaone.net.POSTED!not-for-mail
Newsgroups: alt.folklore.computers,alt.sys.pdp10
Subject: Re: How in Hell did the Great Unix to NT Migration begin??
References: <372d88f7$0$490@news.zetnet.co.uk> <7gkgdq$pj9$5@antiochus.ultra.net>
X-Newsreader: NN version 6.5.0 CURRENT #119
From: werme@werme.ne.mediaone.net (Ric Werme)
Lines: 29
Message-ID: <2IsX2.28538$tY1.18718@wbnws01.ne.mediaone.net>
Date: Tue, 04 May 1999 02:12:14 GMT
NNTP-Posting-Host: 24.128.109.10
X-Trace: wbnws01.ne.mediaone.net 925783934 24.128.109.10 (Mon, 03 May 1999 22:12:14 EDT)
NNTP-Posting-Date: Mon, 03 May 1999 22:12:14 EDT
Xref: news3.best.com alt.folklore.computers:129468 alt.sys.pdp10:4981

jmfbahciv@aol.com writes:

>In article <372d88f7$0$490@news.zetnet.co.uk>, lisard@zetnet.co.uk wrote:
>>
>>
>>On 1999-05-03 weiner@world.std.com(SamWeiner) said:
>>   :It should be noted BLISS was descended from BCPL and used for a
>>   :number of TOPS-10/-20 programs including FORTRAN-10, RMS-10/-20,
>>   :Datatrieve-20, etc.  A few customers used it as well despite the
>>   :lack of support from Digital.
>>   :Sam
>>   :PS Yes Barb, I've heard your rant before but I still liked it.
>>
>>I haven't. Please, do tell...

><grin>  Can't....unless you allow swears.  BLISS of any flavor
>rapidly became a swear in one of my jobs.  Of course, BLISS
>was a swear to anybody who had to support it, too.

After I found Ed Taft's PDP-10 I/O package he wrote at Harvard, I started
rewriting some of my BLISS code in MACRO-10.  With I/O hassles solved, fast
compile times, and a wonder macro language, why use BLISS?

I later named the package Tulip, it's on a DECUS tape.  Need to write a WWW
page for it someday.
--
Ric Werme                  | http://people.ne.mediaone.net/werme
werme@nospam.mediaone.net  | http://www.cyberportal.net/werme
      ^^^^^^^ delete


Article 4983 of alt.sys.pdp10:
Path: news3.best.com!news1.best.com!uninett.no!Norway.EU.net!not-for-mail
From: blarsen@infostream.no (Bjørn Hell Larsen)
Newsgroups: alt.folklore.computers,alt.sys.pdp10
Subject: Re: How in Hell did the Great Unix to NT Migration begin??
Date: 04 May 1999 11:04:13 +0200
Lines: 20
Message-ID: <u6769mcj6.fsf@infostream.no>
References: <372d88f7$0$490@news.zetnet.co.uk> <7gkgdq$pj9$5@antiochus.ultra.net> <2IsX2.28538$tY1.18718@wbnws01.ne.mediaone.net>
NNTP-Posting-Host: aubert.svg.infostream.no
X-Newsreader: Gnus v5.5/Emacs 20.3
Xref: news3.best.com alt.folklore.computers:129478 alt.sys.pdp10:4983


werme@werme.ne.mediaone.net (Ric Werme) writes:

> After I found Ed Taft's PDP-10 I/O package he wrote at Harvard, I started
> rewriting some of my BLISS code in MACRO-10.  With I/O hassles solved, fast
> compile times, and a wonder macro language, why use BLISS?
> 
> I later named the package Tulip, it's on a DECUS tape.  Need to write a WWW
> page for it someday.

You wrote Tulip? In that case, here's a fan letter. I used stuff in Tulip
regularly for several years, it was a great library.

The documentation was great, especially the section that explained the
coroutine support in Tulip. One of the more hilarious documents I've
ever read.


Bjorn



Article 5017 of alt.sys.pdp10:
Path: news3.best.com!news2.best.com!newsfeed.berkeley.edu!newsfeed.enteract.com!netnews.com!chnws02.mediaone.net!24.128.1.101!chnws05.ne.mediaone.net!24.128.44.7!wbnws01.ne.mediaone.net.POSTED!not-for-mail
Newsgroups: alt.folklore.computers,alt.sys.pdp10
Subject: Re: How in Hell did the Great Unix to NT Migration begin??
References: <372d88f7$0$490@news.zetnet.co.uk> <7gkgdq$pj9$5@antiochus.ultra.net>
     <2IsX2.28538$tY1.18718@wbnws01.ne.mediaone.net> <u6769mcj6.fsf@infostream.no>
X-Newsreader: NN version 6.5.0 CURRENT #119
From: werme@werme.ne.mediaone.net (Ric Werme)
Lines: 54
Message-ID: <vvOX2.1222$kE6.916@wbnws01.ne.mediaone.net>
Date: Wed, 05 May 1999 03:00:43 GMT
NNTP-Posting-Host: 24.128.109.10
X-Complaints-To: abuse@rr.com
X-Trace: wbnws01.ne.mediaone.net 925873243 24.128.109.10 (Tue, 04 May 1999 23:00:43 EDT)
NNTP-Posting-Date: Tue, 04 May 1999 23:00:43 EDT
Organization: Road Runner
Xref: news3.best.com alt.folklore.computers:129543 alt.sys.pdp10:5017

blarsen@infostream.no writes:


>werme@werme.ne.mediaone.net (Ric Werme) writes:

>> After I found Ed Taft's PDP-10 I/O package he wrote at Harvard, I started
>> rewriting some of my BLISS code in MACRO-10.  With I/O hassles solved, fast
>> compile times, and a wonderful macro language, why use BLISS?

>> I later named the package Tulip, it's on a DECUS tape.  Need to write a WWW
>> page for it someday.

>You wrote Tulip? In that case, here's a fan letter. I used stuff in Tulip
>regularly for several years, it was a great library.

I wrote just the documentation!  Well, I think I added four
instructions to the parser.  Ed didn't write any documentation, so I
figured I should, if only to repay him for the time it saved me.

>The documentation was great, especially the section that explained the
>coroutine support in Tulip. One of the more hilarious documents I've
>ever read.

Also one of the first pieces of prose I _wanted_ to write, or at least
to write well.  One thing I learned from it was a) boy, do a lot of
typos slip in, and b) after fixing everything I could find I was now a
great proofreader.  The latter can be more hindrance than help -
beyond a certain % of misspellings in a USENET post, I hit the n key,
no matter how interesting the post appeared.  I can spot typos in
things I'm not reading.  When I started typing this I noticed I had
written "wonder macro language".  A few months ago I was reading one
of Bob Metcalfe's columns in the back of Infoworld about networking in
schools.  I was the first to point out to him that he had written
"pubic education".  I don't remember anything else about the column,
it may not have been on networking, but the school system itself.

Yeah, I had fun writing it.  Let's see, I had a couple challenges in it.
One was what was the item in the index that conveyed no useful information.
That was the start of the index.  There was also something about that
coroutine code.  I think the second coroutine did a jump to the JSP
in the first to save an instruction.  (You really have to see that on
paper for it to make any sense.)

I do have the documentation and sample programs.  I will get them on
the WWW, but not for a while.

I do have a new page on the WWW you might enjoy -
http://people.ne.mediaone.net/werme/deimos.html

	-Ric Werme
--
Ric Werme                  | http://people.ne.mediaone.net/werme
werme@nospam.mediaone.net  | http://www.cyberportal.net/werme
      ^^^^^^^ delete


Article 5345 of alt.sys.pdp10:
Path: news3.best.com!news2.best.com!news.maxwell.syr.edu!ptdnetP!newsgate.ptd.net!news.cc.ukans.edu!srvr1.engin.umich.edu!news.tc.cornell.edu!news.graphics.cornell.edu!news
From: westin*nospam@graphics.cornell.edu (Stephen H. Westin)
Newsgroups: alt.sys.pdp10
Subject: Re: BLISS-10
Date: 30 Jun 1999 10:05:22 -0400
Organization: Cornell University Program of Computer Graphics
Lines: 46
Sender: westin@DIESEL
Message-ID: <un1xhiy1p.fsf@graphics.cornell.edu>
References: <7kg08v$iht$1@news.latrobe.edu.au> 	 <7kh5pj$fe0@dfw-ixnews7.ix.netcom.com><7kigdi$49t$3@autumn.news.rcn.net> 	 <ALDERSON.99Jun20154524@netcom17.netcom.com> <376e6250.0@news.kivex.com> <7l94re$9ms$1@autumn.news.rcn.net>
NNTP-Posting-Host: diesel.graphics.cornell.edu
X-Newsreader: Gnus v5.5/Emacs 20.3
Xref: news3.best.com alt.sys.pdp10:5345


It seems to me that most of the 36-bit folks were not fans of
BLISS. There are two main reasons for using a higher-level language.

o Programmer productivity: by using a language that maps better to
  a programmer's thought processes or the problem at hand, we can
  get the same work done in less time. Or more work in the same time,
  etc.

o Portability: by avoiding machine language, we have some hope of
  separating the code from the hardware details of a single machine.

In which ways did BLISS fail in these promises? I can think of several
possibilities.

o By failing to improve on the expressiveness of machine code, it
  didn't really improve productivity. I've never used BLISS and don't
  know much about it; could the experts comment on this?

o By including too much of the peculiar details of a particular
  machine architecture (word length, or byte addressability), it
  failed to be portable. As, say, Multics would be a pain to port
  because of all the explicit PL/1 declarations of 36-bit integers.

o By failing to be reliable, it negated any benefits that might
  otherwise accrue.

o By generating code so inefficient that it wasn't practical to use.

o Maybe the 36-bit machines were just so different that it wasn't
  practical to port code (especially system code) between them and the
  16/32-bit byte-oriented architectures.

Anyway, it seems that C succeeded in a big way where BLISS failed. I
can think of a couple of reasons for that, like the proprietary
single-vendor nature of BLISS, and that it has largely succeeded on
16/32/64-bit byte oriented machines that probably ease the porting
load.

Any comments from those who know a lot more than I?

-- 
-Stephen H. Westin
Any information or opinions in this message are mine: they do not
represent the position of Cornell University or any of its sponsors.



Article 5348 of alt.sys.pdp10:
Path: news3.best.com!news2.best.com!news.maxwell.syr.edu!logbridge.uoregon.edu!news.u.washington.edu!Tomobiki-Cho.CAC.Washington.EDU!mrc
From: Mark Crispin <mrc@CAC.Washington.EDU>
Newsgroups: alt.sys.pdp10
Subject: Re: BLISS-10
Date: Wed, 30 Jun 1999 12:39:46 -0700
Organization: Networks & Distributed Computing
Lines: 24
Message-ID: <Pine.NXT.4.20.9906301223560.29451-100000@Tomobiki-Cho.CAC.Washington.EDU>
References: <7kg08v$iht$1@news.latrobe.edu.au>   <7kh5pj$fe0@dfw-ixnews7.ix.netcom.com><7kigdi$49t$3@autumn.news.rcn.net>   <ALDERSON.99Jun20154524@netcom17.netcom.com> <376e6250.0@news.kivex.com> <7l94re$9ms$1@autumn.news.rcn.net> <un1xhiy1p.fsf@graphics.cornell.edu>
NNTP-Posting-Host: tomobiki-cho.cac.washington.edu
Mime-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
X-Trace: nntp1.u.washington.edu 930771592 21766 (None) 140.142.17.37
X-Complaints-To: help@cac.washington.edu
NNTP-Posting-User: att
To: "Stephen H. Westin" <westin@graphics.cornell.edu>
In-Reply-To: <un1xhiy1p.fsf@graphics.cornell.edu>
Xref: news3.best.com alt.sys.pdp10:5348

The problem with BLISS on the PDP-10 was that it was much easier to write
an application in PDP-10 assembly language than BLISS.  The only PDP-10
operating system which had system calls that were remotely ameniable for
high-level languages was ITS (the .CALL UUO).  In all other operating
systems, setting up a system call required that you knew the precise
accumulator and memory values; this was trivial in assembly language but a
nightmare in a high level language.

C succeeded where BLISS failed for many reasons.  C's syntax is much more
reasonable, and it certainly much easier to write applications in C than
in assembly language on these gruesome hardware architectures being built
today.

I spent 15 years programming PDP-10 assembly language, and a subsequent 11
years programming in C.  I would not think of writing in assembly language
on an Alpha, Pentium, etc.  On the other hand, to this day, I can still
write a PDP-10 program better in assembly language than in C.

-- Mark --

* RCW 19.190 notice: This email address is located in Washington State.	*
* Unsolicited commercial email may be billed $500 per message.		*
Science does not emerge from voting, party politics, or public debate.



Article 5349 of alt.sys.pdp10:
Path: news3.best.com!news2.best.com!feed1.news.rcn.net!rcn!news.idt.net!nyd.news.ans.net!newsxfer.itd.umich.edu!srvr1.engin.umich.edu!news.tc.cornell.edu!news3.cac.psu.edu!not-for-mail
From: viro@weyl.math.psu.edu (Alexander Viro)
Newsgroups: alt.sys.pdp10
Subject: Re: BLISS-10
Date: 30 Jun 1999 16:41:05 -0400
Organization: -ENOENT
Lines: 29
Message-ID: <7ldvd1$6p6@weyl.math.psu.edu>
References: <7kg08v$iht$1@news.latrobe.edu.au> <7l94re$9ms$1@autumn.news.rcn.net> <un1xhiy1p.fsf@graphics.cornell.edu> <Pine.NXT.4.20.9906301223560.29451-100000@Tomobiki-Cho.CAC.Washington.EDU>
NNTP-Posting-Host: weyl.math.psu.edu
Xref: news3.best.com alt.sys.pdp10:5349

In article <Pine.NXT.4.20.9906301223560.29451-100000@Tomobiki-Cho.CAC.Washington.EDU>,
Mark Crispin  <mrc@CAC.Washington.EDU> wrote:
>The problem with BLISS on the PDP-10 was that it was much easier to write
>an application in PDP-10 assembly language than BLISS.  The only PDP-10
>operating system which had system calls that were remotely ameniable for
>high-level languages was ITS (the .CALL UUO).  In all other operating
>systems, setting up a system call required that you knew the precise
>accumulator and memory values; this was trivial in assembly language but a
>nightmare in a high level language.

Could you elaborate on that? If anything, library written in assembler
usually isolates such stuff pretty well. Not all Unices implement system
calls as calls to invalid addresses - normally it's "set registers and cause
a particular trap". Mixing pathes used for syscalls and paging is not
too bright idea... C doesn't know about PDP-11 "trap" or x86 "int $0x80"
either...

>C succeeded where BLISS failed for many reasons.  C's syntax is much more
>reasonable, and it certainly much easier to write applications in C than
>in assembly language on these gruesome hardware architectures being built
>today.

ObBLISSsyntax: what kind of pervert came up with idea of _ and . usage
in BLISS? In C usage of . is more bearable, but using it in place of
unary *... Mind-boggling. Especially nice on dirty printouts...

-- 
"You're one of those condescending Unix computer users!"
"Here's a nickel, kid.  Get yourself a better computer" - Dilbert.


Article 5351 of alt.sys.pdp10:
Path: news3.best.com!news2.best.com!newsfeed.berkeley.edu!enews.sgi.com!harbinger.cc.monash.edu.au!news.latrobe.edu.au!lucifer.its.latrobe.edu.au!cchd
From: cchd@lucifer.its.latrobe.edu.au (Huw Davies)
Newsgroups: alt.sys.pdp10
Subject: Re: BLISS-10
Date: 1 Jul 1999 01:44:20 GMT
Organization: La Trobe University
Lines: 38
Message-ID: <7leh5k$blq$2@news.latrobe.edu.au>
References: <7kg08v$iht$1@news.latrobe.edu.au> <7l94re$9ms$1@autumn.news.rcn.net> <un1xhiy1p.fsf@graphics.cornell.edu> <Pine.NXT.4.20.9906301223560.29451-100000@Tomobiki-Cho.CAC.Washington.EDU> <7ldvd1$6p6@weyl.math.psu.edu>
NNTP-Posting-Host: lucifer.its.latrobe.edu.au
X-Trace: news.latrobe.edu.au 930793460 11962 131.172.12.11 (1 Jul 1999 01:44:20 GMT)
X-Complaints-To: usenet@news.latrobe.edu.au
NNTP-Posting-Date: 1 Jul 1999 01:44:20 GMT
X-Newsreader: TIN [version 1.2 PL2]
Xref: news3.best.com alt.sys.pdp10:5351

Alexander Viro (viro@weyl.math.psu.edu) wrote:

: ObBLISSsyntax: what kind of pervert came up with idea of _ and . usage
: in BLISS? In C usage of . is more bearable, but using it in place of
: unary *... Mind-boggling. Especially nice on dirty printouts...

Well of course, until they changed the standard, _ printed out as <-
(but one character). The idea of . was for consistancy. In most programming
languges, a variable name means either its address or its value depending
on placement. For example, in fortran

    i = i + 1

the i has different meanings. Typically this is expressed as the lvalue (the
address on the left hand side of the =) and the rvalue (the value, on the
right). In Bliss, everything is an lvalue, to get the rvalue, preceed with .

Thus, in Bliss

    i = i + 1

is completely valid, it just doesn't do what you might expect. (It adds one
to the address in memory where i is stored and stores the result in the
memory location pointed to by i). To get "normal" behaviour write

    i = .i + 1

I much prefer this approach to that used by c, but it's a personal thing!
I find things like

    .i = .j

fun to right and easy to understand....
--
 Huw Davies                      | e-mail: Huw.Davies@latrobe.edu.au
 Information Technology Services | Phone: +61 3 9479 1550  Fax: +61 3 9479 1999
 La Trobe University             | "If God had wanted soccer to be played
 Melbourne Australia 3083        | in the air, the sky would be green" 


Article 5353 of alt.sys.pdp10:
From: viro@weyl.math.psu.edu (Alexander Viro)
Newsgroups: alt.sys.pdp10
Subject: Re: BLISS-10
Date: 30 Jun 1999 23:44:09 -0400
Organization: -ENOENT
Lines: 26
Message-ID: <7leo69$843@weyl.math.psu.edu>
References: <7kg08v$iht$1@news.latrobe.edu.au> <Pine.NXT.4.20.9906301223560.29451-100000@Tomobiki-Cho.CAC.Washington.EDU> <7ldvd1$6p6@weyl.math.psu.edu> <7leh5k$blq$2@news.latrobe.edu.au>
NNTP-Posting-Host: weyl.math.psu.edu
Path: news3.best.com!news2.best.com!news.maxwell.syr.edu!news.syr.edu!newsstand.cit.cornell.edu!news.tc.cornell.edu!news3.cac.psu.edu!not-for-mail
Xref: news3.best.com alt.sys.pdp10:5353

In article <7leh5k$blq$2@news.latrobe.edu.au>,
Huw Davies <cchd@lucifer.its.latrobe.edu.au> wrote:
>Alexander Viro (viro@weyl.math.psu.edu) wrote:
>
>: ObBLISSsyntax: what kind of pervert came up with idea of _ and . usage
>: in BLISS? In C usage of . is more bearable, but using it in place of
>: unary *... Mind-boggling. Especially nice on dirty printouts...
>
>Well of course, until they changed the standard, _ printed out as <-
>(but one character). The idea of . was for consistancy. In most programming
>languges, a variable name means either its address or its value depending
>on placement. For example, in fortran
[snip the explanation of explicit vs. implicit dereferencing]

Sorry, but... It's a different question. FWIW I prefer C way - IMHO it's a nice
balance between bare-bones BLISS rules and convoluted mess of A-68, but
that's a personal thing. I was actually asking about a different thing -
choice of operation symbol. I know what '.' means in BLISS and I know the
ideology behind it (no implicit rules, no l- and r-values), but dot is the
worst character for such usage - way too easy to miss it in dirty printout.
***q=**p; is not the easiest thing to read, but ....Q_...P is much nastier.
And as any child of BCPL BLISS is weak-typed, so...

-- 
"You're one of those condescending Unix computer users!"
"Here's a nickel, kid.  Get yourself a better computer" - Dilbert.


Article 5356 of alt.sys.pdp10:
Path: news3.best.com!news2.best.com!news.maxwell.syr.edu!news.idt.net!nyd.news.ans.net!newsxfer.itd.umich.edu!news.itd.umich.edu!sarr
Newsgroups: alt.sys.pdp10
References: <7kg08v$iht$1@news.latrobe.edu.au> <7l94re$9ms$1@autumn.news.rcn.net> <un1xhiy1p.fsf@graphics.cornell.edu> <Pine.NXT.4.20.9906301223560.29451-100000@Tomobiki-Cho.CAC.Washington.EDU>
Reply-To: sarr@umich.edu
Organization: University of Michigan
Subject: Re: BLISS-10
From: sarr@stick.us.itd.umich.edu (Sarr J. Blumson)
Lines: 23
Message-ID: <PtMe3.1014$ce5.19829@news.itd.umich.edu>
Date: Thu, 01 Jul 1999 16:14:39 GMT
NNTP-Posting-Host: 141.211.165.85
X-Trace: news.itd.umich.edu 930845679 141.211.165.85 (Thu, 01 Jul 1999 12:14:39 EDT)
NNTP-Posting-Date: Thu, 01 Jul 1999 12:14:39 EDT
Xref: news3.best.com alt.sys.pdp10:5356

In article <Pine.NXT.4.20.9906301223560.29451-100000@Tomobiki-Cho.CAC.Washington.EDU>,
Mark Crispin  <mrc@CAC.Washington.EDU> wrote:
>
>I spent 15 years programming PDP-10 assembly language, and a subsequent 11
>years programming in C.  I would not think of writing in assembly language
>on an Alpha, Pentium, etc.  On the other hand, to this day, I can still
>write a PDP-10 program better in assembly language than in C.

Having ported Ingres (the version that came with BSD 4.1 [?], not the
commercial version) to TOPS-10, the one thing that really bit me was that
there was no good choice for representing (char *) on the 10.  The compiler
I was using stored them as a byte count from 0, which made pointer arithmetic
easy but made dereferencing a pointer, or casting it to a non char pointer,
expensive.  Storing them as byte pointers would have done the opposite.

Of course this was aggravated by the tendency of the original Ingres crew to
declare _all_ pointers as char *.

-- 
--------
Sarr Blumson                    sarr@umich.edu
voice: +1 734 764 0253          home: +1 734 665 9591
ITD, University of Michigan     http://www-personal.umich.edu/~sarr/


Article 5361 of alt.sys.pdp10:
From: "Phil Gagner" <gagner@svg-law.com>
Newsgroups: alt.sys.pdp10
References: <7kg08v$iht$1@news.latrobe.edu.au> <7l94re$9ms$1@autumn.news.rcn.net> <un1xhiy1p.fsf@graphics.cornell.edu> <Pine.NXT.4.20.9906301223560.29451-100000@Tomobiki-Cho.CAC.Washington.EDU> <PtMe3.1014$ce5.19829@news.itd.umich.edu>
Subject: Re: BLISS-10
Date: Fri, 2 Jul 1999 16:49:37 -0400
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 5.00.2314.1300
X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2314.1300
NNTP-Posting-Host: pc162.svg-law.com
Message-ID: <377d1819.0@news.kivex.com>
Path: news3.best.com!news2.best.com!news.maxwell.syr.edu!dispose.news.demon.net!demon!news-lond.gip.net!news.gsl.net!gip.net!nntp.news.xara.net!xara.net!tank.news.pipex.net!pipex!uunet!zur.uu.net!ffx.uu.net!news2.one.net!news.kivex.com!pc162.svg-law.com
Lines: 43
Xref: news3.best.com alt.sys.pdp10:5361

This is verbatim from HAKMEM, and unless I miss the point of this post, it's
not that hard...
It's one of those little tricks that comes from indexed address arithmetic.

ITEM 165 (Freiberg):
A byte pointer can be converted to a character address < 2^18 by MULI A,<#
bytes/word> followed by SUBI B,1-<# b/w>(A). To get full word character
address, use SUB into a magic table.

Sarr J. Blumson <sarr@stick.us.itd.umich.edu> wrote in message
news:PtMe3.1014$ce5.19829@news.itd.umich.edu...
> In article
<Pine.NXT.4.20.9906301223560.29451-100000@Tomobiki-Cho.CAC.Washington.EDU>,
> Mark Crispin  <mrc@CAC.Washington.EDU> wrote:
> >
> >I spent 15 years programming PDP-10 assembly language, and a subsequent
11
> >years programming in C.  I would not think of writing in assembly
language
> >on an Alpha, Pentium, etc.  On the other hand, to this day, I can still
> >write a PDP-10 program better in assembly language than in C.
>
> Having ported Ingres (the version that came with BSD 4.1 [?], not the
> commercial version) to TOPS-10, the one thing that really bit me was that
> there was no good choice for representing (char *) on the 10.  The
compiler
> I was using stored them as a byte count from 0, which made pointer
arithmetic
> easy but made dereferencing a pointer, or casting it to a non char
pointer,
> expensive.  Storing them as byte pointers would have done the opposite.
>
> Of course this was aggravated by the tendency of the original Ingres crew
to
> declare _all_ pointers as char *.
>
> --
> --------
> Sarr Blumson                    sarr@umich.edu
> voice: +1 734 764 0253          home: +1 734 665 9591
> ITD, University of Michigan     http://www-personal.umich.edu/~sarr/




Article 5350 of alt.sys.pdp10:
Path: news3.best.com!news2.best.com!newsfeed.berkeley.edu!enews.sgi.com!harbinger.cc.monash.edu.au!news.latrobe.edu.au!lucifer.its.latrobe.edu.au!cchd
From: cchd@lucifer.its.latrobe.edu.au (Huw Davies)
Newsgroups: alt.sys.pdp10
Subject: Re: BLISS-10
Date: 1 Jul 1999 01:37:26 GMT
Organization: La Trobe University
Lines: 25
Message-ID: <7legom$blq$1@news.latrobe.edu.au>
References: <7kg08v$iht$1@news.latrobe.edu.au>   <7kh5pj$fe0@dfw-ixnews7.ix.netcom.com><7kigdi$49t$3@autumn.news.rcn.net>   <ALDERSON.99Jun20154524@netcom17.netcom.com> <376e6250.0@news.kivex.com> <7l94re$9ms$1@autumn.news.rcn.net> <un1xhiy1p.fsf@graphics.cornell.edu>
NNTP-Posting-Host: lucifer.its.latrobe.edu.au
X-Trace: news.latrobe.edu.au 930793046 11962 131.172.12.11 (1 Jul 1999 01:37:26 GMT)
X-Complaints-To: usenet@news.latrobe.edu.au
NNTP-Posting-Date: 1 Jul 1999 01:37:26 GMT
X-Newsreader: TIN [version 1.2 PL2]
Xref: news3.best.com alt.sys.pdp10:5350

Stephen H. Westin (westin*nospam@graphics.cornell.edu) wrote:

I did a fair amount of BLISS-10 code (mainly for recreational purposes) as
well as lots of MACRO-10. At the time the learning curve of BLISS was quite
high (in terms of getting the compiler to generate the code that I wanted
rather than its view) but in the end I was reasonably happy with the
two larger projects I wrote - a user mode version of SYSDPY and a translation
of a BCPL program used to produce cross reference listings of BCPL programs.
The reason for the second program was a project to compare "systems 
programming" languages on the DECsystem-10.

: Anyway, it seems that C succeeded in a big way where BLISS failed. I
: can think of a couple of reasons for that, like the proprietary
: single-vendor nature of BLISS, and that it has largely succeeded on
: 16/32/64-bit byte oriented machines that probably ease the porting
: load.

Anyone who has tried to port c programs from either 16bit systems to 32bits
or 32bits to 64bits will find that many programmers make assumptions about
the size of pointers and integers that aren't true....
--
 Huw Davies                      | e-mail: Huw.Davies@latrobe.edu.au
 Information Technology Services | Phone: +61 3 9479 1550  Fax: +61 3 9479 1999
 La Trobe University             | "If God had wanted soccer to be played
 Melbourne Australia 3083        | in the air, the sky would be green" 


Article 5355 of alt.sys.pdp10:
Path: news3.best.com!news2.best.com!newsfeed.berkeley.edu!newsfeed1.earthlink.net!uunet!lax.uu.net!ffx.uu.net!in5.uu.net!news-master.cisco.com!not-for-mail
From: Bill Westfield <billw@flipper.cisco.com>
Newsgroups: alt.sys.pdp10
Subject: Re: BLISS-10
Date: 30 Jun 1999 23:25:59 -0700
Organization: Cisco Systems, Inc.
Lines: 42
Message-ID: <54aetg6g3s.fsf@flipper.cisco.com>
References: <7kg08v$iht$1@news.latrobe.edu.au>
	<7kh5pj$fe0@dfw-ixnews7.ix.netcom.com><7kigdi$49t$3@autumn.news.rcn.net>
	<ALDERSON.99Jun20154524@netcom17.netcom.com>
	<376e6250.0@news.kivex.com> <7l94re$9ms$1@autumn.news.rcn.net>
	<un1xhiy1p.fsf@graphics.cornell.edu>
NNTP-Posting-Host: flipper.cisco.com
X-Newsreader: Gnus v5.3/Emacs 19.34
Xref: news3.best.com alt.sys.pdp10:5355

    o By failing to be reliable, it negated any benefits that might
      otherwise accrue.

That's for sure.


    o By generating code so inefficient that it wasn't practical to use.

I dunno.  Other horrendously bad code generators were/are successful.  This
isn't nearly as bad as the reliability issue above.  For the most part, code
efficiency is only interesting "near the edge", where most programs AREN'T.
If a language catches on, the code generation gets better.


    o Maybe the 36-bit machines were just so different that it wasn't
      practical to port code (especially system code) between them and the
      16/32-bit byte-oriented architectures


Could be.  More like "the target audience for BLISS already was proficient
in assembly language, and therefore recoiled in horror."  You get people to
write in high level languages because its HARD to write assembler (for one
reason or another.)  Macro wasn't HARD, especially for the would-be converts
to bliss.


    Anyway, it seems that C succeeded in a big way where BLISS failed. I
    can think of a couple of reasons for that, like the proprietary
    single-vendor nature of BLISS...

It's worth noticing that C didn't "succeed in a big way" until well AFTER
the PDP10s were thrashing around in their death throes.  As late as 1984,
MAC-OS was released, programmed mainly in Pascal (?)  Plenty of other high
level languages have failed (APL, Modula2, Pascal (for all practical
purposes, I think), OCCAM, etc, etc) I suppose the real question is "why did
languages like fortran, cobol, and C *succeed*", rather than "why did BLISS
fail."

BillW

-- 
(remove spam food from return address)


Article 5374 of alt.sys.pdp10:
Path: news3.best.com!news2.best.com!newsfeed.berkeley.edu!logbridge.uoregon.edu!news.indiana.edu!news.iupui.edu!haystack!mhwood
From: "Mark H. Wood" <mhwood@haystack.Bogusnet.IN.US>
Newsgroups: alt.sys.pdp10
Subject: Re: BLISS-10
Date: Sun, 4 Jul 1999 11:50:53 GMT
Organization: La Petite Hackerie
Lines: 59
Message-ID: <FECGwt.7rJ@Ameritech.net>
References: <7kg08v$iht$1@news.latrobe.edu.au>   <7kh5pj$fe0@dfw-ixnews7.ix.netcom.com> <7kigdi$49t$3@autumn.news.rcn.net>   <ALDERSON.99Jun20154524@netcom17.netcom.com> <376e6250.0@news.kivex.com> <7l94re$9ms$1@autumn.news.rcn.net> <un1xhiy1p.fsf@graphics.cornell.edu>
NNTP-Posting-Host: mhw.ulib.iupui.edu
User-Agent: tin/pre-1.4-19990517 ("Psychonaut") (UNIX) (Linux/2.2.5 (i586))
Xref: news3.best.com alt.sys.pdp10:5374

Stephen H. Westin <westin*nospam@graphics.cornell.edu> wrote:
> In which ways did BLISS fail in these promises? I can think of several
> possibilities.

> o By failing to improve on the expressiveness of machine code, it
>   didn't really improve productivity. I've never used BLISS and don't
>   know much about it; could the experts comment on this?

That could be it.  I suspect that it was simply too powerful for the
sort of things it made me want to do with it.  But that could be just
me.

> o By including too much of the peculiar details of a particular
>   machine architecture (word length, or byte addressability), it
>   failed to be portable. As, say, Multics would be a pain to port
>   because of all the explicit PL/1 declarations of 36-bit integers.

Like the three-branch IF in FORTRAN?  Or car/cdr in Lisp?  Both are
straight out of the IBM 70x architecture.

BTW take a look at Bliss-11 (used quite a bit on the CMMP), or Common
Bliss (which was supposed to enable portability between 16-, 32-, and
36-bit architectures, only one of which has the funky byte pointers in
hardware).  Bliss was only tied to the PDP-10 if you chose to think of
it that way.

> o By failing to be reliable, it negated any benefits that might
>   otherwise accrue.

Do you mean that the compiler was unreliable, or that the language
made programmers unreliable?  I don't recall either one being
especially true of Bliss.

> o By generating code so inefficient that it wasn't practical to use.

On the contrary.  I thought I was a pretty good PDP-10 assembly
programmer until I started studying assembly listings from BLIS10 and
saw some of the things *it* did.

> o Maybe the 36-bit machines were just so different that it wasn't
>   practical to port code (especially system code) between them and the
>   16/32-bit byte-oriented architectures.

Again, FORTRAN grew up on 36-bit machines (from IBM) and made the
transition without serious problems.

The biggest problem I recall is that the learning curve was nearly a
step function.  You didn't have to learn so many things, but they were
hard lessons, at least for me, and you had to learn them *all* before
you could do much.  And, as somebody else has said, assembly on the
PDP-10 was simply too easy for most people to bother with "system
programming" HLLs.  On TOPS-20 I wrote a mail user-agent that was
pretty much a clone of the non-network portions of DECmail-MS,
entirely in MACRO, and looking back I don't see how it would have been
significantly easier in any other language we had then.
-- 
-- 
Mark H. Wood, radical centrist     OpenPGP ID 876A8B75     mhwood@ameritech.net
Why do we always draw organization charts upside down?


Article 5376 of alt.sys.pdp10:
Path: news3.best.com!news2.best.com!news.maxwell.syr.edu!ix.netcom.com!news
From: pfarrell@netcom.com (Pat Farrell)
Newsgroups: alt.sys.pdp10
Subject: Re: BLISS-10
Date: Mon, 05 Jul 1999 23:07:16 GMT
Organization: Netcom
Lines: 116
Message-ID: <7lrdiu$oqm@dfw-ixnews7.ix.netcom.com>
References: <7kg08v$iht$1@news.latrobe.edu.au> 	 <7kh5pj$fe0@dfw-ixnews7.ix.netcom.com><7kigdi$49t$3@autumn.news.rcn.net> 	 <ALDERSON.99Jun20154524@netcom17.netcom.com> <376e6250.0@news.kivex.com> <7l94re$9ms$1@autumn.news.rcn.net> <un1xhiy1p.fsf@graphics.cornell.edu>
NNTP-Posting-Host: netcom9.netcom.com
X-NETCOM-Date: Mon Jul 05  6:02:54 PM CDT 1999
NNTP-Posting-User: pfarrell
X-Newsreader: Forte Free Agent 1.0.82
Xref: news3.best.com alt.sys.pdp10:5376

westin*nospam@graphics.cornell.edu (Stephen H. Westin) wrote:
>It seems to me that most of the 36-bit folks were not fans of
>BLISS. There are two main reasons for using a higher-level language.
>o Programmer productivity: >o Portability:

>In which ways did BLISS fail in these promises? I can think of several
>possibilities.

> I've never used BLISS and don't   know much about it; could the experts comment on this?

It is clear from your post that you didn't use it.

Expressiveness of the language is not important, or we'd
all be writing in APL.

First, you should at least know that there were two
separate Bliss compilers in 36-bit land. Bliss-10 and Bliss-36.
Bliss-10 was the first language, and was oritented strongly towards
TOPS-10 (and later Tops-20. Bliss-36 was supposed to be
part of the Common Bliss family that was capable of working
accross LCG, Vaxen, and PDP-11 families.

Sometimes I wonder if I was the only person who started a 
project in Bliss-36 and was happy.

Bliss-10/36  could "express" any machine construct, it could
simply emit machine codes if you couldn't figure out what
would cause it to generate what you wanted. While it was a bit
baroque, I don't think it was worse than the mess that Tops-20
Exec bacame with all the pseudo-HLL constructs/macros
that folks did to "improve" the codng styles.

>o By including too much of the peculiar details of a particular
>  machine architecture (word length, or byte addressability), it
>  failed to be portable

Two problems here:
1) there wasn't much interest in portability back then. KLs were
a damn sight more interesting and capable than Vaxen or, shudder,
PDP-11s. We wrote truely portable code, but it cost a bunch and
we never made back the investment.

2) writing portable code is really expensive. Writing applications
portably is expensive.  Writing system-level utilities and 
languages accross architectures is seriously hard.
It has only become practical recently, when improvements in
speeds and memory sizes make "wasteful" techniques such
as virtual machines affordable.

While interpreted languages are not new (Smalltalk-80 gives a hint
of its age in its name), they were not successful until fairly
recently. I don't want to step into a holy war, but I don't
consider most C code to be portable. Java and Smalltalk
allow fairly portable code to be written, but they have
limits and were not feasible back when KLs were feasible.

>o By failing to be reliable, it negated any benefits that might
>  otherwise accrue.

Reliable? Huh? Do you have examples to back up this assertion?
I found Bliss to be highly reliable. We wrote a DBMS package in
Bliss-36 that ran unmodified for more than 15 years. And it used
what were at the time, techniques that were on the forefront
of the art, such as user mode extended addressing, OWGBP, etc.

>o By generating code so inefficient that it wasn't practical to use.

The Bliss-36 code generator was decent, and the peephole 
optimizer was fine. Sure you could beat it with hand tweaked
code, but it generated faster and more compact code than
a beginning Macro-10/20 programmer. I had many examples
of this from my staff.

>o Maybe the 36-bit machines were just so different that it wasn't
>  practical to port code (especially system code) between them and the
>  16/32-bit byte-oriented architectures.

We did it. The economics didn't work, but that isn't a fault if the
language or the concept of High Level Languages. Problems
that were solvable by PDP-11s in 1979 were not addressed
by folks with the resources to own a KL or two.

>Anyway, it seems that C succeeded in a big way where BLISS failed. I
>can think of a couple of reasons for that, like the proprietary
>single-vendor nature of BLISS, and that it has largely succeeded on
>16/32/64-bit byte oriented machines that probably ease the porting
>load.

There are as many beliefs behind the success of C over Bliss as
programmers.

I subscribe to the theory that C succeeded because it was a small
language that worked well on small machines such as the original IBM
PC. Hackers couldn't afford KLs or even Vaxen, but they could
afford an IBM PC or clone. They needed a language. C fit and
was available. So a market appeared and grew.

Bliss wouldn't have fit in an original PC, even the BLiss-16 compilers
of the early 80s expected you to have a Vaxen or KL to cross compile
on.

Bliss was invented to solve similar problems as C, in the same
timeframe. They are  closely parallel languages, altho I much
prefer some of Bliss' constructs over C's. But like in other
areas of technology, the superior product does not always
succeed. I think of Bliss as the Beta of late 70s programming.
Better than the competion that beat it in the marketplace.
But still a loser because it lost.

All just IMHO, of course.

Pat

Pat Farrell                           pfarrell@netcom.com
PGP key on the usual servers



Article 5378 of alt.sys.pdp10:
Path: news3.best.com!news2.best.com!feed1.news.rcn.net!rcn!d14
From: jmfbahciv@aol.com
Newsgroups: alt.sys.pdp10
Subject: Re: BLISS-10
Date: Tue, 06 Jul 99 09:11:21 GMT
Organization: UltraNet Communications, Inc.
Lines: 64
Message-ID: <7lso11$77o$1@autumn.news.rcn.net>
References: <7kg08v$iht$1@news.latrobe.edu.au> 	 <7kh5pj$fe0@dfw-ixnews7.ix.netcom.com><7kigdi$49t$3@autumn.news.rcn.net> 	 <ALDERSON.99Jun20154524@netcom17.netcom.com> <376e6250.0@news.kivex.com> <7l94re$9ms$1@autumn.news.rcn.net> <un1xhiy1p.fsf@graphics.cornell.edu> <7lrdiu$oqm@dfw-ixnews7.ix.netcom.com>
X-Trace: Wf3K9XC+0DwJJJ9BUwMk0rBf5DrqYX8ztH8FLV0Gs8Y=
X-Complaints-To: abuse@rcn.com
NNTP-Posting-Date: 6 Jul 1999 11:07:13 GMT
X-Newsreader:  News Xpress Version 1.0 Beta #4
Xref: news3.best.com alt.sys.pdp10:5378

In article <7lrdiu$oqm@dfw-ixnews7.ix.netcom.com>,
   pfarrell@netcom.com (Pat Farrell) wrote:
>westin*nospam@graphics.cornell.edu (Stephen H. Westin) wrote:
>>It seems to me that most of the 36-bit folks were not fans of
>>BLISS. There are two main reasons for using a higher-level language.
>>o Programmer productivity: >o Portability:
>
>>In which ways did BLISS fail in these promises? I can think of several
>>possibilities.
>
>> I've never used BLISS and don't   know much about it; could the experts 
comment on this?
>
>It is clear from your post that you didn't use it.
>
>Expressiveness of the language is not important, or we'd
>all be writing in APL.
>
>First, you should at least know that there were two
>separate Bliss compilers in 36-bit land. Bliss-10 and Bliss-36.
>Bliss-10 was the first language, and was oritented strongly towards
>TOPS-10 (and later Tops-20. Bliss-36 was supposed to be
>part of the Common Bliss family that was capable of working
>accross LCG, Vaxen, and PDP-11 families.
>
>Sometimes I wonder if I was the only person who started a 
>project in Bliss-36 and was happy.

Probably :-) none of my guys were happy.  What kind of project was it?
I more curious about the level of the software, i.e., OS
level, user mode level, application level, etc. rather than
the specs of the project.

<snip>

>>Anyway, it seems that C succeeded in a big way where BLISS failed. I
>>can think of a couple of reasons for that, like the proprietary
>>single-vendor nature of BLISS, and that it has largely succeeded on
>>16/32/64-bit byte oriented machines that probably ease the porting
>>load.
>
>There are as many beliefs behind the success of C over Bliss as
>programmers.
>
>I subscribe to the theory that C succeeded because it was a small
>language that worked well on small machines such as the original IBM
>PC. Hackers couldn't afford KLs or even Vaxen, but they could
>afford an IBM PC or clone. They needed a language. C fit and
>was available. So a market appeared and grew.
<snip>

One of the problems of any flavor of the blisses on the -10
is that a build almost required a stand-alone -10 that
didn't crash for 12 hours (that was not our environment).
It was so much faster to build stuff written in MACRO-10
than BLISS-xx.  So a developer who wrote in assembly
would rarely have to wait a day for a build turn-around
time.  Building anything, and I mean ANYTHING, with BLISS
meant that the IBM card turnaround time mentality became
the norm.  That wasn't our style.

/BAH

Subtract a hundred and four for e-mail.


Article 5385 of alt.sys.pdp10:
Path: news3.best.com!news1.best.com!su-news-hub1.bbnplanet.com!news.gtei.net!logbridge.uoregon.edu!news-peer.gip.net!news.gsl.net!gip.net!ix.netcom.com!news
From: pfarrell@netcom.com (Pat Farrell)
Newsgroups: alt.sys.pdp10
Subject: Re: BLISS-10
Date: Wed, 07 Jul 1999 03:43:06 GMT
Organization: Netcom
Lines: 61
Message-ID: <7lui3e$123@dfw-ixnews21.ix.netcom.com>
References: <7kg08v$iht$1@news.latrobe.edu.au> 	 <7kh5pj$fe0@dfw-ixnews7.ix.netcom.com><7kigdi$49t$3@autumn.news.rcn.net> 	 <ALDERSON.99Jun20154524@netcom17.netcom.com> <376e6250.0@news.kivex.com> <7l94re$9ms$1@autumn.news.rcn.net> <un1xhiy1p.fsf@graphics.cornell.edu> <7lrdiu$oqm@dfw-ixnews7.ix.netcom.com> <7lso11$77o$1@autumn.news.rcn.net>
NNTP-Posting-Host: netcom10.netcom.com
X-NETCOM-Date: Tue Jul 06 10:38:22 PM CDT 1999
NNTP-Posting-User: pfarrell
X-Newsreader: Forte Free Agent 1.0.82
Xref: news3.best.com alt.sys.pdp10:5385

jmfbahciv@aol.com wrote:
>>Sometimes I wonder if I was the only person who started a 
>>project in Bliss-36 and was happy.
>Probably :-) none of my guys were happy.  What kind of project was it?

We actually did two.

First, we wrote a DBMS package.
Specifically because System 1022 had a hard stop limit of 18-bits of
records per database. We needed more than that, and we ended up
writing our own. Ended up with 700,000 records filling an RP07.
Wow, that was huge back then.

The second was a client/server application that used DEC-Pro350
(PDP-11) "personal computers" as intellegent programmable terminals
for our KLs. We also had a version that used a Vaxen or multi-user
PDP-11 as an intellegent concentrator for Vt100s. This was 
before "decnet phase 3" made interoperability and/or 
internetworking between DEC product lines feasible :-)

>I more curious about the level of the software, i.e., OS
>level, user mode level, application level, etc. rather than
>the specs of the project.

Gee, That was along time ago, so the specifics are fuzzy.
We were a Tops-20 source site, running the "current" monitors.
We jumped on whatever Tops-20 first had user mode extended
addressing. The hardware were six KL, all of which
ended up as B models, usually 2060s or 2065s.

The code was used as tools for applications, most of which
were written in Cobol with System 1022 as the datastore.

I'm not sure what else you're currious about. Let me know,
and if I remember...

>One of the problems of any flavor of the blisses on the -10
>is that a build almost required a stand-alone -10 that
>didn't crash for 12 hours (that was not our environment).

Funny, We didn't see this problem. Of course, we weren't building
a monitor. But we constantly used the usual edit/compile/debug cycle,
and it was effective for Bliss-36. 

For the crossported stuff, where we had to run Bliss-16 on the KL,
the download and TKB crud from PDP-11 land added tons of time
to the cycle. But I think the inept PDP-11 debugging tools cost
as much as the compilers themselves.

I'm not claiming Bliss-36 was as fast as Macro-20, but it was not
a problem for us.

We didn't run our KLs are grossly overloaded as education sites.
We had users paying for the machines, and they complained about
performance when we had 50 or so users simultaneously.

Pat

Pat Farrell                           pfarrell@netcom.com
PGP key on the usual servers



Article 5396 of alt.sys.pdp10:
Path: news3.best.com!news2.best.com!news.maxwell.syr.edu!newspeer1.nac.net!news.mv.net!not-for-mail
From: Dan Murphy <dlm@opost.com>
Newsgroups: alt.sys.pdp10
Subject: Re: BLISS-10
Date: 09 Jul 1999 23:12:56 -0400
Organization: MV Communications, Inc.
Lines: 54
Message-ID: <m27lo9qjt3.fsf@opost.com>
References: <7kg08v$iht$1@news.latrobe.edu.au> <376e6250.0@news.kivex.com> <7l94re$9ms$1@autumn.news.rcn.net> <un1xhiy1p.fsf@graphics.cornell.edu> <377e1a77.0@news.wizvax.net>
NNTP-Posting-Host: opost.com
X-Trace: pyrite.mv.net 931576385 7723 207.22.41.2 (10 Jul 1999 03:13:05 GMT)
X-Complaints-To: abuse@mv.com
NNTP-Posting-Date: 10 Jul 1999 03:13:05 GMT
X-Newsreader: Gnus v5.5/Emacs 20.3
Xref: news3.best.com alt.sys.pdp10:5396

wilson@dbit.com (John Wilson) writes:

> In article <un1xhiy1p.fsf@graphics.cornell.edu>,
> 
> Anyway, I've often heard people refer to C as a glorified assembler -- I *wish*
> that were true!

    My gracious!  I pop in here after a long absense and find... a
    compiler religious war!  Some things never change...
    
    I was one of the violent opponents of Bliss in the late '70s when
    it was moving into DEC and was finally declared to be the
    "official" implementation language by Gordon Bell.  It's not that
    I opposed using an HLL (where the 'H' here is debatable), but I
    really thought Bliss didn't make the grade.  It really was little
    more than a glorified assembler and did little that I couldn't do
    in MACRO-10 macros.

    Bliss was available and basically free, and DEC was too cheap and
    short-sighted to think beyond that.

    And yes, I subsequently used Bliss a lot.  Thank GoD!  Otherwise,
    in those circumstances, I'd have had to program in VAX Macro (gag,
    wretch).  Ironically, PDP-10 assembler programming was so easy
    (compared to most), especially with macros to handle stack and
    structure stuff, that we were lulled into a false sense of not
    really needing a more "higher level" language.
    
    Bad mistake.

    If TOPS-20 had been written (or recoded) into an HLL, it could be
    where Unix is today.  It was a better OS in the days when both
    were new, but Unix won out for two reasons: ran on cheap machines
    compared to PDP-10's, and (2) portable in a practical sense --
    i.e. was ported to lots of other machines during critical
    junctures.  Shit.  If we'd even have recoded in Bliss, we might
    have had a chance.
    
    I knew the writing was on the wall when the ARPA community, which
    had earlier adopted TENEX/TOPS-20, declared in the late '70s that
    it was going to Unix, and that DEC could take VMS and shove it.

    C was good enough, and "the perfect is the enemy of the good".
    All the "much better" languages are dust.  Yes, C and Unix
    supported each other in an amazing example of symbiosis.  Each
    would likely have disappeared without the other.
    
    Saying C is not portable is ratshit.  That it is not perfectly
    portable is exactly why it succeeded.  Those who waited for the
    perfectly portable language are still waiting.
    
    
    dlm
    


Article 5399 of alt.sys.pdp10:
Path: news3.best.com!news2.best.com!news1.best.com!nntp.primenet.com!nntp.gctr.net!nuq-peer.news.verio.net!nntp2.cerf.net!nntp3.cerf.net!news.sdsc.edu!news.tc.cornell.edu!news3.cac.psu.edu!not-for-mail
From: viro@weyl.math.psu.edu (Alexander Viro)
Newsgroups: alt.sys.pdp10
Subject: Re: BLISS-10
Date: 10 Jul 1999 02:07:10 -0400
Organization: -ENOENT
Lines: 38
Message-ID: <7m6nue$ai3@weyl.math.psu.edu>
References: <7kg08v$iht$1@news.latrobe.edu.au> <un1xhiy1p.fsf@graphics.cornell.edu> <377e1a77.0@news.wizvax.net> <m27lo9qjt3.fsf@opost.com>
NNTP-Posting-Host: weyl.math.psu.edu
Xref: news3.best.com alt.sys.pdp10:5399

In article <m27lo9qjt3.fsf@opost.com>, Dan Murphy  <dlm@opost.com> wrote:
>wilson@dbit.com (John Wilson) writes:
[snip]
>    If TOPS-20 had been written (or recoded) into an HLL, it could be
>    where Unix is today.  It was a better OS in the days when both
>    were new, but Unix won out for two reasons: ran on cheap machines
>    compared to PDP-10's, and (2) portable in a practical sense --
>    i.e. was ported to lots of other machines during critical

(3) had been licensed to many places (UCB was probably the most essential)
(4) had an extremely lucky leak - Lions' Book.
(5) was not controlled by the single hardware vendor.

>    junctures.  Shit.  If we'd even have recoded in Bliss, we might
>    have had a chance.

Commercial success or not, but there were at least 3 reimplementations of
UNIX kernel (Coherent, Minix and Linux). I've never seen the guts of Coherent,
but API was pretty close to v7. Minix guts are very different, but API is
again v7-ish. Linux... well, we are very close to 4.4BSD, both guts and API.
Were there any attempts to write a TOPS-20 look-alike? XKL uses modified
TOPS-20 sources. Were there any independent projects? 4BSD is not exactly
what I mean - AFAIK many *very* nice ideas came from TOPS-20/TENEX, but...

Is there *any* text that would describe the TOPS-20 internals on the level
of Daemon Book? Printed, on-line - something? 

ObInternals: could somebody give a description of the file locking in TOPS-20
(information other systems are also welcome, indeed). I mean *technical*
description - for $DEITY sake, to hell with advocacy. Maybe the best way would
be to reply by email - I'll summarize and post to the group. (Reason: current
cleanup of our file locking. And yes, I'm aware that it sucks. Those who came
up with POSIX locks semantics should be forced to suck several putrefied
whales through the straw. Daily. I'm sure that they will manage).

-- 
I've written small amts of c++, what more familiarity with objects do you want?
	L. Detweiler on c.o.l.d.s, after lenghty drivel re developing OO-OS.


Article 5409 of alt.sys.pdp10:
Path: news3.best.com!news2.best.com!feed1.news.rcn.net!rcn!not-for-mail
From: "Alan H. Martin" <AMartin@MA.UltraNet.Com>
Newsgroups: alt.sys.pdp10
Subject: Re: BLISS-10
Date: Sat, 10 Jul 1999 10:51:36 -0400
Lines: 8
Message-ID: <37875DF8.E1AC8E3F@MA.UltraNet.Com>
References: <7kg08v$iht$1@news.latrobe.edu.au> <7kh5pj$fe0@dfw-ixnews7.ix.netcom.com>
		<7kigdi$49t$3@autumn.news.rcn.net> <ALDERSON.99Jun20154524@netcom17.netcom.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
X-Trace: w5/C3Ig/VQ7XCCsWCMwXySsPDFCYNCUUR+fA4ba5RpY=
X-Complaints-To: abuse@rcn.com
NNTP-Posting-Date: 10 Jul 1999 14:51:53 GMT
X-Accept-Language:  en,en-US,en-GB,es
X-Mailer:  Mozilla 4.6 [en] (Win95; U)
Xref: news3.best.com alt.sys.pdp10:5409

"Richard M. Alderson III" wrote:

> ...  Bliss-36 cost *lots* of money when it came out as a product; ...

$13,800.00, as I recall.
				/AHM
-- 
Alan Howard Martin				AMartin@MA.UltraNet.Com


Article 5429 of alt.sys.pdp10:
Path: news3.best.com!news1.best.com!144.212.100.101.MISMATCH!newsfeed.mathworks.com!news-out.cwix.com!newsfeed.cwix.com!204.59.152.222!news-peer.gip.net!news.gsl.net!gip.net!ix.netcom.com!nntp.ix.netcom.com!alderson
From: alderson@netcom10.netcom.com (Richard M. Alderson III)
Newsgroups: alt.sys.pdp10
Subject: Re: BLISS-10
Date: 12 Jul 1999 18:50:07 GMT
Organization: NETCOM On-line services
Lines: 16
Message-ID: <ALDERSON.99Jul12115007@netcom10.netcom.com>
References: <7l94re$9ms$1@autumn.news.rcn.net> <7kg08v$iht$1@news.latrobe.edu.au>
	 <7kh5pj$fe0@dfw-ixnews7.ix.netcom.com> <7kigdi$49t$3@autumn.news.rcn.net>
	 <ALDERSON.99Jun20154524@netcom17.netcom.com> <376e6250.0@news.kivex.com>
Reply-To: alderson@netcom.com
NNTP-Posting-Host: netcom10.netcom.com
X-NETCOM-Date: Mon Jul 12  1:50:07 PM CDT 1999
NNTP-Posting-User: alderson
In-reply-to: "Larry S. Samberg"'s message of Mon, 28 Jun 1999 20:42:59 -0400
Xref: news3.best.com alt.sys.pdp10:5429

[ First posted on Friday, 9 July 1999.  Forgive me if it is a duplicate on your
  system, as it never showed up on Netcom.
  --rma ]

In article <7l94re$9ms$1@autumn.news.rcn.net> "Larry S. Samberg" <larry-samberg@sonoma-systems.com> writes:

>Phil, 
>Wasn't most of Typeset-10 written in BLISS?

I asked Valdeane Alusic about this this morning; she said "No."  She did say,
after further reflection, that people might have thought so because that group
was the first in DEC to use "structured programming" for a project.
--
Rich Alderson			Last LOTS Tops-20 Systems Programmer, 1984-1991
				Current maintainer, MIT TECO EMACS (v. 170)
last name @ XKL dot COM		Chief systems administrator, XKL LLC, 1998-now


Article 5422 of alt.sys.pdp10:
Path: news3.best.com!news2.best.com!feed1.news.rcn.net!rcn!not-for-mail
From: "Alan H. Martin" <AMartin@MA.UltraNet.Com>
Newsgroups: alt.sys.pdp10
Subject: Re: BLISS-10
Date: Sun, 11 Jul 1999 17:01:29 -0400
Lines: 23
Message-ID: <37890629.71FB82A6@MA.UltraNet.Com>
References: <7kg08v$iht$1@news.latrobe.edu.au> <7kh5pj$fe0@dfw-ixnews7.ix.netcom.com> <7kigdi$49t$3@autumn.news.rcn.net>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
X-Trace: bq6qMCJa1K/9tbqFQ3xdt6LYDzVJ5GXQVYysOE+dxrs=
X-Complaints-To: abuse@rcn.com
NNTP-Posting-Date: 11 Jul 1999 21:01:47 GMT
X-Accept-Language:  en,en-US,en-GB,es
X-Mailer:  Mozilla 4.6 [en] (Win95; U)
Xref: news3.best.com alt.sys.pdp10:5422

jmfbahciv@aol.com wrote:
> 
> In article <7kh5pj$fe0@dfw-ixnews7.ix.netcom.com>,
>    bugs@freebsd.netcom.com (Mark Hittinger) wrote:
...
> >Bliss-11 may have been Decus, but a source license
> >for Bliss-32 was probably a lot harder to get :-).  Bliss-10 came with
> >source code in the tradition of 10-ness and was not Decus.
> 
> I don't think we shipped the sources.   I'd have to look at the
> SPDs to verify this.  But I don't think there was an SPD for
> BLISS-10, hence the customer couldn't "order" it.

BLISS-10 sources were definitely available to the field, although it probably
was only supported for building supported products written in BLISS-10.  I
can't remember if the BLISS-10 sources initially shipped just with FORTRAN-10
or what.
				/AHM
P. S.  Speaking of BLISS sources, BLISS-11 is online at
ftp://iecc.com/pub/file/bliss.tar.gz .  Its optimizer is better than
BLISS-10's - rather closer to Common BLISS's quality.
-- 
Alan Howard Martin				AMartin@MA.UltraNet.Com


Article 5423 of alt.sys.pdp10:
Path: news3.best.com!news2.best.com!feed1.news.rcn.net!rcn!d1
From: jmfbahciv@aol.com
Newsgroups: alt.sys.pdp10
Subject: Re: BLISS-10
Date: Mon, 12 Jul 99 09:36:16 GMT
Organization: UltraNet Communications, Inc.
Lines: 40
Message-ID: <7mcjpe$k1p$1@autumn.news.rcn.net>
References: <7kg08v$iht$1@news.latrobe.edu.au> <7kh5pj$fe0@dfw-ixnews7.ix.netcom.com> <7kigdi$49t$3@autumn.news.rcn.net> <37890629.71FB82A6@MA.UltraNet.Com>
X-Trace: T1uS8SIA4f1gRliStc6FDdyyZGOozvXE2TVR8yaBBvg=
X-Complaints-To: abuse@rcn.com
NNTP-Posting-Date: 12 Jul 1999 11:33:02 GMT
X-Newsreader:  News Xpress Version 1.0 Beta #4
Xref: news3.best.com alt.sys.pdp10:5423

In article <37890629.71FB82A6@MA.UltraNet.Com>,
   "Alan H. Martin" <AMartin@MA.UltraNet.Com> wrote:
>jmfbahciv@aol.com wrote:
>> 
>> In article <7kh5pj$fe0@dfw-ixnews7.ix.netcom.com>,
>>    bugs@freebsd.netcom.com (Mark Hittinger) wrote:
>....
>> >Bliss-11 may have been Decus, but a source license
>> >for Bliss-32 was probably a lot harder to get :-).  Bliss-10 came with
>> >source code in the tradition of 10-ness and was not Decus.
>> 
>> I don't think we shipped the sources.   I'd have to look at the
>> SPDs to verify this.  But I don't think there was an SPD for
>> BLISS-10, hence the customer couldn't "order" it.
>
>BLISS-10 sources were definitely available to the field, although it 
probably
>was only supported for building supported products written in BLISS-10.

Alan, think.  One didn't need BLISS-10 sources in order to build
products written in BLISS-10.

>  I
>can't remember if the BLISS-10 sources initially shipped just with 
FORTRAN-10
>or what.

The _Bliss_ sources weren't shipped on the tape; just the *.EXEs
and (I can't remember this) I suspect LIBs, etc.  The Fortran
tape was a mess since it also insisted on shipping RELs, UNVs and
EXEs that belonged on the CUSP tape.  One of my knockdowns with
Sara was to get that stuff off all the language tapes so that
there would no longer be any version skews of supported products.
It took the monitor group about 5 years to solve that problem
along with eliminating 22 CUSP update tapes.

/BAH


Subtract a hundred and four for e-mail.


Article 5419 of alt.sys.pdp10:
Path: news3.best.com!news2.best.com!feed1.news.rcn.net!rcn!not-for-mail
From: "Alan H. Martin" <AMartin@MA.UltraNet.Com>
Newsgroups: alt.sys.pdp10
Subject: Re: BLISS-10
Date: Sun, 11 Jul 1999 14:07:15 -0400
Lines: 110
Message-ID: <3788DD53.6B3F6CE7@MA.UltraNet.Com>
References: <7kg08v$iht$1@news.latrobe.edu.au> <7l94re$9ms$1@autumn.news.rcn.net> <un1xhiy1p.fsf@graphics.cornell.edu> <Pine.NXT.4.20.9906301223560.29451-100000@Tomobiki-Cho.CAC.Washington.EDU> <7ldvd1$6p6@weyl.math.psu.edu> <7leh5k$blq$2@news.latrobe.edu.au>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
X-Trace: l0eDdFuSA5Bsc9siQrr51kH1BAtwvsXnDbs9MRtW0Wk=
X-Complaints-To: abuse@rcn.com
NNTP-Posting-Date: 11 Jul 1999 18:07:32 GMT
X-Accept-Language:  en,en-US,en-GB,es
X-Mailer:  Mozilla 4.6 [en] (Win95; U)
Xref: news3.best.com alt.sys.pdp10:5419

Huw Davies wrote:
> 
>... In Bliss, everything is an lvalue, ...

(Except for BIND or LITERAL identifiers).


>... to get the rvalue, preceed with .
> 
> Thus, in Bliss
> 
>     i = i + 1
> 
> is completely valid, it just doesn't do what you might expect. (It adds one
> to the address in memory where i is stored and stores the result in the
> memory location pointed to by i). To get "normal" behaviour write
> 
>     i = .i + 1
> 
> I much prefer this approach to that used by c, but it's a personal thing!
> I find things like
> 
>     .i = .j
> 
> fun to right and easy to understand....

This one might put hair on your chest:

"
GLOBAL ROUTINE IOBLD (NODEDATA,DEFUNIT,UNITFLAG)= ! [1465] New

!----------------------------------------------------------------------
!This routine expects a pointer in STK[0] to:
!	pointer to:
!		unit expression
!		format expression
...
!	option
!	iolist
!----------------------------------------------------------------------
!
! For ENCODE and DECODE, the action routine that parses the keyword list
! guarantees that unit, format, and variable are all present.

BEGIN
...
	REGISTER BASE R1:R2;
...
	! Offsets into semantic block built by KEYSCAN

	STRUCTURE RBASE [I,J,K,L] =
		  CASE .I OF SET
   %0%		     (\.RBASE +.J)<.K,.L>;
   %1%		    (@\.RBASE +.J)<.K,.L>
		  TES;

	BIND RBASE QUNIT = 0<FULL,R2>:
		   QFMT = 1<FULL,R2>:
		   ...

	R1 = .STK[0];			! Get pointer to args
	R2 = .R1[ELMNT];
...
	IF .QUNIT EQL 0
	...

	! Check UNIT.  Legal forms are *, integer expression,
	! character variable or array element or substring,
	! or character array name.

	IF .QUNIT^(-18) EQL ASTERISK
	THEN
	...
	ELSE IF .QUNIT[VALTYPE] EQL CHARACTER
	THEN
	...
"

Note all three dereference operators (atsign/backslash/dot) used in a row in
the definition of structure RBASE.


@y<p,s,x,i> would generate:
	MOVE	AC,y		; Ignore p, s, x, i

\y<p,s,x,i> would do:
	MOVE	AC,@(i)y(x)	; Ignore p, s

.y<p,s,x,i> would do:
	LDB	AC,[POINT p,s,@(i)y(x)]
(or	MOVE	AC,@(i)y(x) if P == 0, S == 36)

The %1% case arm used @, because it was known if one wanted to indirect one
more time, the fetched value was just an 18-bit address.  So, it wouldn't work
to treat it as a byte pointer with a general-case LDB, and it didn't contain
indexing or indirection, so there was no point chewing on it with a full-blown
effective address calculation.

Both case arms used \, because the table was known to contain indirect words
which indexed off of some register named R2, and so you needed the generated
code to indirect through the table entry.


I didn't write the code - I just had the honor of pre-reviewing it before the
project meeting.  (The author didn't want to have to pull me out of the
ceiling by the ankles if it had been sprung on me unawares).  The code was
necessary and sufficient; we shipped it; it works.
				/AHM
-- 
Alan Howard Martin				AMartin@MA.UltraNet.Com


Article 5420 of alt.sys.pdp10:
Path: news3.best.com!news1.best.com!su-news-hub1.bbnplanet.com!news.gtei.net!logbridge.uoregon.edu!newspeer1.nac.net!netnews.com!feed1.news.rcn.net!rcn!not-for-mail
From: "Alan H. Martin" <AMartin@MA.UltraNet.Com>
Newsgroups: alt.sys.pdp10
Subject: Re: BLISS-10
Date: Sun, 11 Jul 1999 15:19:12 -0400
Lines: 155
Message-ID: <3788EE30.3B7E7B0F@MA.UltraNet.Com>
References: <7kg08v$iht$1@news.latrobe.edu.au>   <7kh5pj$fe0@dfw-ixnews7.ix.netcom.com><7kigdi$49t$3@autumn.news.rcn.net>   <ALDERSON.99Jun20154524@netcom17.netcom.com> <376e6250.0@news.kivex.com> <7l94re$9ms$1@autumn.news.rcn.net> <un1xhiy1p.fsf@graphics.cornell.edu> <Pine.NXT.4.20.9906301223560.29451-100000@Tomobiki-Cho.CAC.Washington.EDU>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
X-Trace: pHAaNvlmKh8/y4F4Ia5n3YpHplF2/55jYtEncIX/TgU=
X-Complaints-To: abuse@rcn.com
NNTP-Posting-Date: 11 Jul 1999 19:19:30 GMT
X-Accept-Language:  en,en-US,en-GB,es
X-Mailer:  Mozilla 4.6 [en] (Win95; U)
Xref: news3.best.com alt.sys.pdp10:5420

Mark Crispin wrote:
> 
> The problem with BLISS on the PDP-10 was that it was much easier to write
> an application in PDP-10 assembly language than BLISS.  The only PDP-10
> operating system which had system calls that were remotely ameniable for
> high-level languages was ITS (the .CALL UUO).  In all other operating
> systems, setting up a system call required that you knew the precise
> accumulator and memory values; this was trivial in assembly language but a
> nightmare in a high level language.

It was trivial in BLISS-36 to define custom linkages for Tops-20 JSY which put
input and output parameters into the proper registers.  For instance, DFOUT
took AC1-AC4 as inputs, left stuff in AC1 and AC4, and had a single skip
return on success.

LINKAGE
	J_1234_14_1 = JSYS(1, 2, 3, 4; 1, 4) : SKIP(1);
BIND ROUTINE
	DFOUT = %O'235' : J_1234_14;

So, a DFOUT of the contents of variables A, B, C and D, which put the output
parameters into variables Y and Z, and which evaluated true if the call
skipped, would be written:

	IF NOT DFOUT(.A, .B, .C, .D; Y, Z)
	THEN error


On TOPS-10, you constructed the arg block in memory, stuck its address into
the AC of your choice, and executed the CALLI of your choice with a MACHOP or
MACHSKIP builtin.

With BLISS-10, you declared the CALLI opcode as a MACHOP and called them as
functions (perhaps in an IFSKIP statement).


Here's some BLISS-36 to print the lowseg and hiseg filespec of a TOPS-10 job. 
Comments might help, but I can still reverse engineer all the fields of those
system calls from the code.  A tighter coding style, a subroutine for pulling
a job or hiseg's path, and a printf would have come in handy, too.  (It would
really rock with C++ wrappers for the CALLIs).

ROUTINE PNAMES_DSP(JOB) : NOVALUE=
BEGIN

REGISTER
	T1,T2;
LOCAL
	PATH_BLOCK:STR_PATH;

	T1=GET_TAB(.JOB,$GTRDV,0);
	IF .T1 NEQ 0
	 THEN
	  BEGIN
		SIXWRD(T1);
		OUTASZ(':');
	  END;
	T1=GET_TAB(.JOB,$GTRFN,%SIXBIT '??????');
	SIXOUT(6,T1);
	PATH_BLOCK[PATH_PPN]=GET_TAB(.JOB,$GTRDI,0);
	IF .PATH_BLOCK[PATH_PPN] NEQ 0
	 THEN
	  BEGIN
		T1=0;
		DO
		  BEGIN
			T2=(PATH_BLOCK[PATH_SFD]+.T1)=
			 GET_TAB(.JOB,$GTRS0+.T1,0);
			T1=.T1+1
		  END
		 UNTIL .T2 EQL 0;
		DIR_DSP(PATH_BLOCK)
	  END;

	IF .POINTR(HISEG,SN_SPY) NEQ 1
	 THEN
	  BEGIN
		IF .HISEG<RH> NEQ 0
		 THEN
		  BEGIN
			OUTASZ(' + ');
			T1=GET_TAB(.HISEG<RH>,$GTDEV,%SIXBIT '??????');
			IF .T1 NEQ 0
			 THEN
			  BEGIN
				SIXWRD(T1);
				OUTASZ(':')
			  END;
			IF .POINTR(HISEG,SN_SHR)
			 THEN T1=GET_TAB(.HISEG<RH>,$GTPRG,%SIXBIT '??????')
			 ELSE T1=%SIXBIT '(Priv)';
			SIXOUT(6,T1);
			T1=PATH_BLOCK[PATH_PPN]=GET_TAB(.HISEG<RH>,$GTPPN,0);
			IF .T1<LH> NEQ 0
			 THEN
			  BEGIN
				T1=0;
				DO
				  BEGIN
					T2=(PATH_BLOCK[PATH_SFD]+.T1)=
					 GET_TAB(.HISEG<RH>,$GTRS0+.T1,0);
					T1=.T1+1
				  END
				 UNTIL .T2 EQL 0;
				DIR_DSP(PATH_BLOCK)
			  END;
			OUTASZ(' ')
		  END
	  END
	 ELSE
	  BEGIN
		OUTASZ(' + SPY:');
		T1=(.HISEG<RH>+1)^-9;
		DECOUT(T1);
		OUTASZ('P ')
	  END
END;

BTW, those SIXOUT and DECOUT calls are invoking LUUOs in-line.


> C succeeded where BLISS failed for many reasons.  C's syntax is much more
> reasonable, and it certainly much easier to write applications in C than
> in assembly language on these gruesome hardware architectures being built
> today.

Common BLISS could be written very portably between 10's, 11's, VAXen and
Alphas, but engineering almost never paid the extra investment it required -
extra rules to follow every step of the way.  Those decisions are still
bearing fruit.

Two examples of portability shortcuts: VAX DEBUG was written with only
half-hearted use of the character-handling functions which abstracted away
PDP-10 byte pointers from 11/VAX byte addressing.  So, a port to the -10 would
have required looking at every line of code to see where CH$ builtins had to
be added.  And I suspect that a lot of BLISS-32 code *knew* that a fullword
was 32 bits, which must have been just lovely to port to Alpha.

There was no silver bullet to learning BLISS.  The Fortran-10 group once hired
3 or 4 people in a row who learned BLISS-10 in college, but that was an
aberration.  A vanishingly small user community outside the company guaranteed
that you had to train almost everyone (sometimes without even in-house course
materials), and wait the months for them to come fully up to speed.  That kept
costs too high once C really became the rage.

A lot of BLISS programmers have been just as opposed to C as any BLISS
opposition by MACRO programmers we see here.  Widespread C use inside Digital
seems to postdate the advent of X/DECwindows.  I don't know how much change
was caused by it, but many products did have to write or rewrite GUIs. 
Maintaining and using in-house BLISS bindings that shadowed the various X APIs
didn't really catch on.  So, a critical mass of people got dipped in C sooner
or later (and more and more people came in the door already knowing C).
				/AHM
-- 
Alan Howard Martin				AMartin@MA.UltraNet.Com


Article 5421 of alt.sys.pdp10:
Path: news3.best.com!news2.best.com!feed1.news.rcn.net!rcn!not-for-mail
From: "Alan H. Martin" <AMartin@MA.UltraNet.Com>
Newsgroups: alt.sys.pdp10
Subject: Re: BLISS-10
Date: Sun, 11 Jul 1999 16:49:18 -0400
Lines: 43
Message-ID: <3789034E.1DCE46B1@MA.UltraNet.Com>
References: <7kg08v$iht$1@news.latrobe.edu.au> <37706667.3038342@netnews.worldnet.att.net> <Pine.NXT.4.20.9906222324400.20271-100000@Tomobiki-Cho.CAC.Washington.EDU> <QCgc3.2767$Wr1.95137@wbnws01.ne.mediaone.net>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
X-Trace: iHKGllPsYd3F949Jjs+X+7XFUpYH5e8Dv+P+wEzR9Q8=
X-Complaints-To: abuse@rcn.com
NNTP-Posting-Date: 11 Jul 1999 20:49:36 GMT
X-Accept-Language:  en,en-US,en-GB,es
X-Mailer:  Mozilla 4.6 [en] (Win95; U)
Xref: news3.best.com alt.sys.pdp10:5421

Ric Werme wrote:
> 
> Mark Crispin <mrc@CAC.Washington.EDU> writes:
> 
> >As I recall, the code generated by the BLISS compiler was horrible.  Most
> >C compilers on x86, Alpha, etc. can generate better code than a human can
> >write assembly by hand.  The situation was the exact opposite on the
> >PDP-10.
> 
> I guess I don't remember the code quality.  I do recall keeping tabs
> on the object file size - it jumped up if I forgot a "." in front of a
> variable.  One problem - BLISS wanted to pass arguments on the stack.
> Assembler programmers use registers and on the -10 use certain
> registers for pointers to things that needed to be passed to many
> routines within a module, so the argument passing cost for them was
> zero.

BLISS-10's code quality was nothing to write home about.  BLISS-36's was
pretty good, especially for the time.  (I suspect it was tighter than the
MACRO code in the average random cusp).

As far as reserving special registers goes, BLISS-10 had some features for it,
and BLISS-36 was very good at it.  BLISS-10 would let you state where the
stack pointer, etc. lived, and let you reserve some registers to be declared
as variables.  BLISS-36 would let you declare that specific routines of yours
passed certain arguments in certain registers.  Furthermore I recall that if
you told it (as I posted elsewhere today), "this JSYS takes ACs 1, and 2 as
input, and leaves something in AC1 as a result", and said your own routine
took a byte pointer in AC1, then code like this:

ROUTINE MY_ROUTINE(BUFFER_BP) : MY_ROUTINE_LINKAGE=
BEGIN
	...
	THE_JSYS(.BUFFER_BP, .THIS_STRING; BUFFER_BP);
	THE_JSYS(.BUFFER_BP, .THAT_STRING; BUFFER_BP);
	...
END;

would allocate the buffer byte pointer to AC1 and leave it there for all the
calls, which were moving it along as they wrote to the in-core buffer.
				/AHM
-- 
Alan Howard Martin				AMartin@MA.UltraNet.Com


Article 5431 of alt.sys.pdp10:
Path: news3.best.com!news2.best.com!news.maxwell.syr.edu!feed1.news.rcn.net!rcn!not-for-mail
From: "Alan H. Martin" <AMartin@MA.UltraNet.Com>
Newsgroups: alt.sys.pdp10
Subject: Re: BLISS-10
Date: Mon, 12 Jul 1999 22:38:19 -0400
Lines: 61
Message-ID: <378AA69B.D8772FC0@MA.UltraNet.Com>
References: <7kg08v$iht$1@news.latrobe.edu.au> <7kh5pj$fe0@dfw-ixnews7.ix.netcom.com> <7kigdi$49t$3@autumn.news.rcn.net> <37890629.71FB82A6@MA.UltraNet.Com> <7mcjpe$k1p$1@autumn.news.rcn.net>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
X-Trace: To339Uy55OW1soIwrNYumWYfYp1APK4df4gqXx0ZZ30=
X-Complaints-To: abuse@rcn.com
NNTP-Posting-Date: 13 Jul 1999 02:38:40 GMT
X-Accept-Language:  en,en-US,en-GB,es
X-Mailer:  Mozilla 4.6 [en] (Win95; U)
Xref: news3.best.com alt.sys.pdp10:5431

jmfbahciv@aol.com wrote:
> 
> In article <37890629.71FB82A6@MA.UltraNet.Com>,
>    "Alan H. Martin" <AMartin@MA.UltraNet.Com> wrote:
> >jmfbahciv@aol.com wrote:
> >>
> >> In article <7kh5pj$fe0@dfw-ixnews7.ix.netcom.com>,
> >>    bugs@freebsd.netcom.com (Mark Hittinger) wrote:
> >....
> >> >Bliss-11 may have been Decus, but a source license
> >> >for Bliss-32 was probably a lot harder to get :-).  Bliss-10 came with
> >> >source code in the tradition of 10-ness and was not Decus.
> >>
> >> I don't think we shipped the sources.   I'd have to look at the
> >> SPDs to verify this.  But I don't think there was an SPD for
> >> BLISS-10, hence the customer couldn't "order" it.
> >
> >BLISS-10 sources were definitely available to the field, although it
> probably
> >was only supported for building supported products written in BLISS-10.
> 
> Alan, think.  One didn't need BLISS-10 sources in order to build
> products written in BLISS-10.

0.  I entered a couple of dozen DCE edits to Fortran-10 from the Software
Dispatch while I was in the field.

1.  BLISS-10 sources were definitely available to the field.

2.  BLISS-10 was probably only supported to build supported products written
in BLISS-10.  (Just like PA1050 would only accept bugs encountered while
running supported compatibility mode cusps).  However by 1981, BLISS-10 bugs
would probably have been dumped in the Fortran group's lap, and we would have
just issued a permanent restriction w/workaround.  It was a non-goal to touch
a hair on BLISS-10's head.

I once met the last BLISS-10 maintainer, but I can't remember his name...


> >  I
> >can't remember if the BLISS-10 sources initially shipped just with
> FORTRAN-10
> >or what.
> 
> The _Bliss_ sources weren't shipped on the tape; just the *.EXEs
> and (I can't remember this) I suspect LIBs, etc.  ...

Quite plausible.  I don't remember how the BLISS-10 sources were available to
the field.


>... The Fortran
> tape was a mess since it also insisted on shipping RELs, UNVs and
> EXEs that belonged on the CUSP tape.  One of my knockdowns with
> Sara was to get that stuff off all the language tapes so that
> there would no longer be any version skews of supported products.

Might have been before my time.
				/AHM
-- 
Alan Howard Martin				AMartin@MA.UltraNet.Com


Article 5435 of alt.sys.pdp10:
Path: news3.best.com!news2.best.com!feed1.news.rcn.net!rcn!not-for-mail
From: "Alan H. Martin" <AMartin@MA.UltraNet.Com>
Newsgroups: alt.sys.pdp10
Subject: Re: BLISS-10
Date: Tue, 13 Jul 1999 00:20:13 -0400
Lines: 79
Message-ID: <378ABE7D.69E03940@MA.UltraNet.Com>
References: <7kg08v$iht$1@news.latrobe.edu.au> 	 <7kh5pj$fe0@dfw-ixnews7.ix.netcom.com><7kigdi$49t$3@autumn.news.rcn.net> 	 <ALDERSON.99Jun20154524@netcom17.netcom.com> <376e6250.0@news.kivex.com> <7l94re$9ms$1@autumn.news.rcn.net> <un1xhiy1p.fsf@graphics.cornell.edu>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
X-Trace: dy/zVqvLozSu/VZ3HhN7Bvzk9/iqorzCpMqKMHnHQnw=
X-Complaints-To: abuse@rcn.com
NNTP-Posting-Date: 13 Jul 1999 04:20:33 GMT
X-Accept-Language:  en,en-US,en-GB,es
X-Mailer:  Mozilla 4.6 [en] (Win95; U)
Xref: news3.best.com alt.sys.pdp10:5435

"Stephen H. Westin" wrote:
> 
> It seems to me that most of the 36-bit folks were not fans of
> BLISS. There are two main reasons for using a higher-level language.
> 
> o Programmer productivity: by using a language that maps better to
>   a programmer's thought processes or the problem at hand, we can
>   get the same work done in less time. Or more work in the same time,
>   etc.
> 
> o Portability: by avoiding machine language, we have some hope of
>   separating the code from the hardware details of a single machine.
> 
> In which ways did BLISS fail in these promises? I can think of several
> possibilities.
...
> o By failing to be reliable, it negated any benefits that might
>   otherwise accrue.

Between 1977 and 1986, I can recall seeing evidence of 3 bugs in BLISS-10. 
(It had plenty of bug fixes, but these were all the ones I saw evidence of). 
The evidence was one piece of graffiti, a revelation found by reading
BLISS-10's sources, and an actual bug encountered while coding:

1.  This comment in FORTRAN-10:

"
%(****FEB 23,1972 - THE FOLLOWING BLOCK WAS INSERTED TO
	PREVENT A BLISS BUG THAT DELETED CODE . THIS BLOCK FORCES
	BLISS TO USE 2 TEMP REGS***)%
BEGIN
	OWN T,T1,T2,T3;
	T = 1; T1 = 2; T2 = 3; T3 = 4;
END;
"

We never saw another example of it, and for all we know the bug was fixed.


2.  One of the two typos in FORTRAN-10's PEEPOP.BLI was dutifully copied into
BLISS-10's H3PEEP.BLI (a comparison of an instruction's Y field symbol table
pointer, instead of the Y field's value, to another instruction's register
number).  It would erroneously go ahead and change this:

	MOVEM	R1,R2
	OP	R2,ANY
	MOVE	R1,R2
to this:
	MOVEM	R1,R2
	OP	R2,ANY

even though R2 had been changed by OP, and so R1 did really need to be
reloaded. 
I never heard of anyone encountering the bug; we didn't try and fix it.  (The
other bug, a missing dot, was silently fixed in BLISS-10 without being
reported back to FORTRAN-10, until I spotted it while reading code).


3.  This would make BLISS-10 punt:

MACRO
	OUTASZ(STR)=OUTSTR(PASZ STR)$;

	OUTASZ('~');

The value for tilde was used as an escape by BLISS's macro processor, and it
couldn't tolerate it as a macro actual.  The one place we wanted to pass a
tilde to a macro (the place in FINE's help screen for M-~), we just expanded
the macro in-place.

Of course, I saw many dozens of people use BASIC-10 for 3 1/2 years in high
school, and never once saw BASIC screw anyone there.  So by that metric, BLISS
was incredibly unreliable.


BLISS-36?  I should have been so lucky as to have had a job using it.
				/AHM
-- 
Alan Howard Martin				AMartin@MA.UltraNet.Com


Article 5437 of alt.sys.pdp10:
Path: news3.best.com!news2.best.com!feed1.news.rcn.net!rcn!nntp.abs.net!newshub2.home.com!news.home.com!newshub1-work.home.com!null!albaugh
From: albaugh@agames.com (Mike Albaugh)
Newsgroups: alt.sys.pdp10
Subject: Re: BLISS-10
Date: 13 Jul 1999 17:46:43 GMT
Organization: Atari Games Corporation
Lines: 24
Message-ID: <7mfu23$7ic$1@null.agames.com>
References: <7kg08v$iht$1@news.latrobe.edu.au> <Pine.NXT.4.20.9906241055470.22648-100000@Tomobiki-Cho.CAC.Washington.EDU> <7ku5qp$d62$1@null.agames.com> <378AAC9A.308D1AF3@MA.UltraNet.Com>
NNTP-Posting-Host: java.agames.com
X-Trace: null.agames.com 931888003 7756 192.245.83.156 (13 Jul 1999 17:46:43 GMT)
X-Complaints-To: abuse@agames.com
NNTP-Posting-Date: 13 Jul 1999 17:46:43 GMT
X-Newsreader: TIN [version 1.2 PL2]
Xref: news3.best.com alt.sys.pdp10:5437

Alan H. Martin (AMartin@MA.UltraNet.Com) wrote:
: Mike Albaugh wrote:
: > 

: >         Ob. Bliss reference: at least two coin-operated video-games
: > were writen in Bliss. One giant "attaperson" to the one who can
: > name them... :-)

: Doesn't sound like there are any takers - what's the answer?

	APB (top-down "Cop Game") and 720 (The original Skateboard game,
which used the phrase "Skate or Die", but is otherwise not related
to the unauthorized sequel of that name :-) Both written in Bliss-16
and run on T-11s. The relevance to alt.sys.PDP10 is pretty low, until
I also mention that when DEC came around trying to sell us J-11 chipsets,
we were already pretty tired of "stupid MMU tricks" (The games used
a homebrew banking scheme. noit a real DEC MMU), and unwilling to
sell the new single-chip Vaxen (not that we could really have used
them, ...), so I asked the salesdroid about the possibility of a
single-chip KI10 equivalent. He just about swallowed his tongue, but
claimed he'd "Ask about it". Good thing I didn't hold my breath.

					Mike
| albaugh@agames.com, speaking only for myself


Article 5449 of alt.sys.pdp10:
Path: news3.best.com!news2.best.com!newsfeed.berkeley.edu!su-news-hub1.bbnplanet.com!paloalto-snf1.gtei.net!news.gtei.net!news1.digital.com!pa.dec.com!sniff.shr.dec.com!chuck.enet.dec.com!otoole
From: otoole@chuck.enet.dec.com (Chuck O'Toole)
Newsgroups: alt.sys.pdp10
Subject: Re: BLISS-10
Date: 14 Jul 1999 13:43:46 GMT
Organization: Compaq Computer Corporation, Shrewsbury MA.
Lines: 44
Distribution: world
Message-ID: <7mi46i$b2e$1@sniff.shr.dec.com>
References: <7kg08v$iht$1@news.latrobe.edu.au> <7kh5pj$fe0@dfw-ixnews7.ix.netcom.com> <7kigdi$49t$3@autumn.news.rcn.net> <37890629.71FB82A6@MA.UltraNet.Com> <7mcjpe$k1p$1@autumn.news.rcn.net> <378AA69B.D8772FC0@MA.UltraNet.Com>
Reply-To: otoole@chuck.enet.dec.com (Chuck O'Toole)
NNTP-Posting-Host: chuck.shr.dec.com
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
X-Newsreader: mxrn 6.18-31
Xref: news3.best.com alt.sys.pdp10:5449


In article <378AA69B.D8772FC0@MA.UltraNet.Com>, "Alan H. Martin"
<AMartin@MA.UltraNet.Com> writes:

  [snip...]

|>2.  BLISS-10 was probably only supported to build supported products
|>written
|>in BLISS-10.  (Just like PA1050 would only accept bugs encountered while
|>running supported compatibility mode cusps).  However by 1981, BLISS-10
|>bugs
|>would probably have been dumped in the Fortran group's lap, and we would
|>have
|>just issued a permanent restriction w/workaround.  It was a non-goal to
|>touch
|>a hair on BLISS-10's head.
|>
  [snip...]

"the horror...the horror..."  [Brando]

Ah, Bliss-10.  In late '75 I inherited MCS-10 (Message Control System
- DEC's answer to transaction processing at the time).  The majority
of the messaging logic was written in Bliss-10, about a 4 inch thick
listing on old 132 column fan fold printer paper.

After I finally got it to work (and ship!), with JMF's help of course,
a new version of Bliss-10 came out and when I compiled MCS with it, MCS
completely stopped working.  Some construct I was using stopped generating
the proper code.  Rather than re-qualify and test that much Bliss code, I
simply started shipping the older Bliss .EXE on the MCS tape.  Ruffled a
few feathers but it was the only way I was ever going to get off the MCS
project, I didn't want that to become my life's work.

That was my only encounter with Bliss, ever.  Scarred me for life.  I stuck
with assembly (-10, -11, VAX) until the mid 80's when I had to learn Pascal
(don't ask)
-- 

/cdo
-----
Disclaimer:  The opinions expressed are my own and
do not necessarily represent those of my employer,
loved ones, or passing acquaintances.


Article 5452 of alt.sys.pdp10:
Path: news3.best.com!news1.best.com!nntp.primenet.com!nntp.gctr.net!newspeer1.nac.net!netnews.com!feed1.news.rcn.net!rcn!not-for-mail
From: "Alan H. Martin" <AMartin@MA.UltraNet.Com>
Newsgroups: alt.sys.pdp10
Subject: Re: BLISS-10
Date: Wed, 14 Jul 1999 21:50:17 -0400
Lines: 45
Message-ID: <378D3E59.2CE7A0C5@MA.UltraNet.Com>
References: <7kg08v$iht$1@news.latrobe.edu.au> <7kh5pj$fe0@dfw-ixnews7.ix.netcom.com> <7kigdi$49t$3@autumn.news.rcn.net> <37890629.71FB82A6@MA.UltraNet.Com> <7mcjpe$k1p$1@autumn.news.rcn.net> <378AA69B.D8772FC0@MA.UltraNet.Com> <7mht44$ns2$1@autumn.news.rcn.net>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
X-Trace: SBUM7VhsmKx359PZgH95zEhwhhpBYHBwdElh0ZAiQko=
X-Complaints-To: abuse@rcn.com
NNTP-Posting-Date: 15 Jul 1999 01:50:53 GMT
X-Accept-Language:  en,en-US,en-GB,es
X-Mailer:  Mozilla 4.6 [en] (Win95; U)
Xref: news3.best.com alt.sys.pdp10:5452

jmfbahciv@aol.com wrote:
> 
> In article <378AA69B.D8772FC0@MA.UltraNet.Com>,
>    "Alan H. Martin" <AMartin@MA.UltraNet.Com> wrote:
> >jmfbahciv@aol.com wrote:
...
> >> Alan, think.  One didn't need BLISS-10 sources in order to build
> >> products written in BLISS-10.
> >
> >0.  I entered a couple of dozen DCE edits to Fortran-10 from the Software
> >Dispatch while I was in the field.
> 
> DCE?  What's that?  ...

DCE was not only a Fortran-10 maintainer, but probably had to sign off on all
of *your* bug fixes.


>... Patches to Fortran-10 are not patches to
> Bliss.

It means I have experience patching Fortran-10 in the field.  Have you?


> >However by 1981, BLISS-10 bugs
> >would probably have been dumped in the Fortran group's lap,
> 
> Nope.  There was an official Bliss maintainer but he worked
> closely with the Fortran group.

The last BLISS-10 build was 19-Jan-78.


> >
> >I once met the last BLISS-10 maintainer, but I can't remember his name...
> 
> What do mean once met.  He was at the Fortran meetings quite often.
> Eric Magrath.

I know who Eric is; I met him on or before 4-Jun-81.  This was a younger guy,
probably with a less conservative haircut.  I think he was visiting the office
area.
				/ahm
-- 
Alan Howard Martin				AMartin@MA.UltraNet.Com


Article 5451 of alt.sys.pdp10:
Path: news3.best.com!news1.best.com!newshub.northeast.verio.net!howland.erols.net!outgoing.news.rcn.net.MISMATCH!feed1.news.rcn.net!rcn!not-for-mail
From: "Alan H. Martin" <AMartin@MA.UltraNet.Com>
Newsgroups: alt.sys.pdp10
Subject: Re: BLISS-10
Date: Wed, 14 Jul 1999 21:08:34 -0400
Lines: 21
Message-ID: <378D3492.FAC48090@MA.UltraNet.Com>
References: <7kg08v$iht$1@news.latrobe.edu.au> 	 <7kh5pj$fe0@dfw-ixnews7.ix.netcom.com><7kigdi$49t$3@autumn.news.rcn.net> 	 <ALDERSON.99Jun20154524@netcom17.netcom.com> <376e6250.0@news.kivex.com> <7l94re$9ms$1@autumn.news.rcn.net> <un1xhiy1p.fsf@graphics.cornell.edu> <378ABE7D.69E03940@MA.UltraNet.Com> <7mht84$ns2$2@autumn.news.rcn.net>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
X-Trace: JQajaqLftdvVYNkeA4dqmYFZAiA5HW9VYZ8kc/Cmewo=
X-Complaints-To: abuse@rcn.com
NNTP-Posting-Date: 15 Jul 1999 01:09:02 GMT
X-Accept-Language:  en,en-US,en-GB,es
X-Mailer:  Mozilla 4.6 [en] (Win95; U)
Xref: news3.best.com alt.sys.pdp10:5451

jmfbahciv@aol.com wrote:
> 
> In article <378ABE7D.69E03940@MA.UltraNet.Com>,
>    "Alan H. Martin" <AMartin@MA.UltraNet.Com> wrote:

> >Between 1977 and 1986, I can recall seeing evidence of 3 bugs in BLISS-10.
...
> >1.  This comment in FORTRAN-10:
> >
> >"
> >%(****FEB 23,1972 - THE FOLLOWING BLOCK WAS INSERTED TO
> >       PREVENT A BLISS BUG THAT DELETED CODE . THIS BLOCK FORCES
> >       BLISS TO USE 2 TEMP REGS***)%
> <snip>
> 
> Now look at the date.  ...

During V1 development or field test.
				/AHM
-- 
Alan Howard Martin				AMartin@MA.UltraNet.Com


Article 5458 of alt.sys.pdp10:
Path: news3.best.com!news1.best.com!144.212.100.101.MISMATCH!newsfeed.mathworks.com!solaris.cc.vt.edu!news.vt.edu!news.cc.ukans.edu!srvr1.engin.umich.edu!news.tc.cornell.edu!news.graphics.cornell.edu!news
From: westin*nospam@graphics.cornell.edu (Stephen H. Westin)
Newsgroups: alt.sys.pdp10
Subject: Re: BLISS-10
Date: 15 Jul 1999 09:44:18 -0400
Organization: Cornell University Program of Computer Graphics
Lines: 56
Sender: westin@DIESEL
Message-ID: <uyagif2od.fsf@graphics.cornell.edu>
References: <7kg08v$iht$1@news.latrobe.edu.au> 	 <7kh5pj$fe0@dfw-ixnews7.ix.netcom.com><7kigdi$49t$3@autumn.news.rcn.net> 	 <ALDERSON.99Jun20154524@netcom17.netcom.com> <376e6250.0@news.kivex.com> <7l94re$9ms$1@autumn.news.rcn.net> <un1xhiy1p.fsf@graphics.cornell.edu> <7ldaoo$el2$1@autumn.news.rcn.net> <ulnd1ir14.fsf@graphics.cornell.edu> <7mkevn$82j$3@autumn.news.rcn.net>
NNTP-Posting-Host: diesel.graphics.cornell.edu
X-Newsreader: Gnus v5.5/Emacs 20.3
Xref: news3.best.com alt.sys.pdp10:5458

jmfbahciv@aol.com writes:

> In article <ulnd1ir14.fsf@graphics.cornell.edu>,
>    westin*nospam@graphics.cornell.edu (Stephen H. Westin) wrote:
> >jmfbahciv@aol.com writes:
> >
> >> In article <un1xhiy1p.fsf@graphics.cornell.edu>,
> >>    westin*nospam@graphics.cornell.edu (Stephen H. Westin) wrote:
> >> >
> >> >It seems to me that most of the 36-bit folks were not fans of
> >> >BLISS. There are two main reasons for using a higher-level language.
> >> >
> >> >o Programmer productivity: by using a language that maps better to
> >> >  a programmer's thought processes or the problem at hand, we can
> >> >  get the same work done in less time. Or more work in the same time,
> >> >  etc.
> >> 
> >> Phooey.
> >
> >Well, I had hoped for a somewhat more extensive discussion than
> >that. 
> <snip>
> 
> Stephen,  Have you been following the stories? 

Well, your article was the first I saw in reply to mine, so at the
point I posted that, there were no "stories" to follow. A more
extensive discussion has, in fact, ensued. Possibly before your
message quoted above, given the asychronous nature of USENET.

> Hidden in them
> are a lot of the reasons that Bliss was a pig.

Well I get the impression that you just don't trust any
compiler. Certainly developing a system using tools (including
compilers) that themselves are under development is a real problem; I
get the impression that this is what happened with BLISS on the 36-bit
machines, as the compiler either would generate bad code, forcing
workarounds, or would suddenly change its code generation, breaking
existing code. It's not, I think, a fundamental flaw in the concept of
using higher-level languages. I suspect that this factor had a lot to
do with the relative ease of developing Unix compared to Multics:
since C is such a simpler language than PL/1, a functional, reliable
compiler was much easier to develop.

I also get the impression that moving from an assembly-oriented
environment is a real problem, as cultural assumptions have already
worked their way deep into the system design. Arbitrary calling
conventions are hard to accomodate in a higher level language, for
instance, and this apparently made BLISS something of an outsider in
TOPS-10.

-- 
-Stephen H. Westin
Any information or opinions in this message are mine: they do not
represent the position of Cornell University or any of its sponsors.


Article 5467 of alt.sys.pdp10:
Path: news3.best.com!news2.best.com!feed1.news.rcn.net!rcn!not-for-mail
From: "Alan H. Martin" <AMartin@MA.UltraNet.Com>
Newsgroups: alt.sys.pdp10
Subject: Re: BLISS-10
Date: Sat, 17 Jul 1999 12:36:36 -0400
Lines: 19
Message-ID: <3790B114.20E7987@MA.UltraNet.Com>
References: <7kg08v$iht$1@news.latrobe.edu.au> 	 <7kh5pj$fe0@dfw-ixnews7.ix.netcom.com><7kigdi$49t$3@autumn.news.rcn.net> 	 <ALDERSON.99Jun20154524@netcom17.netcom.com> <376e6250.0@news.kivex.com> <7l94re$9ms$1@autumn.news.rcn.net> <un1xhiy1p.fsf@graphics.cornell.edu> <7lrdiu$oqm@dfw-ixnews7.ix.netcom.com> <7lso11$77o$1@autumn.news.rcn.net>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
X-Trace: L9N37jjp83EnLSMFwAULj2cwen4EZPadN0T0mkutJEM=
X-Complaints-To: abuse@rcn.com
NNTP-Posting-Date: 17 Jul 1999 16:37:05 GMT
X-Accept-Language:  en,en-US,en-GB,es
X-Mailer:  Mozilla 4.6 [en] (Win95; U)
Xref: news3.best.com alt.sys.pdp10:5467

jmfbahciv@aol.com wrote:

> One of the problems of any flavor of the blisses on the -10
> is that a build almost required a stand-alone -10 that
> didn't crash for 12 hours (that was not our environment).
> It was so much faster to build stuff written in MACRO-10
> than BLISS-xx.  So a developer who wrote in assembly
> would rarely have to wait a day for a build turn-around
> time.  Building anything, and I mean ANYTHING, with BLISS
> meant that the IBM card turnaround time mentality became
> the norm.  That wasn't our style.

Building 350K lines of the FORTRAN compiler took an order of magnitude
less time than that on KL2137 (let alone MRFORT).  A single module
rebuild meant typing in the compile and link lines; no script - just a
.CCL file for the modules on the link line.
				/AHM
-- 
Alan Howard Martin				AMartin@MA.UltraNet.Com


Article 5466 of alt.sys.pdp10:
Path: news3.best.com!news2.best.com!news.maxwell.syr.edu!feed1.news.rcn.net!rcn!not-for-mail
From: "Alan H. Martin" <AMartin@MA.UltraNet.Com>
Newsgroups: alt.sys.pdp10
Subject: Re: BLISS-10
Date: Sat, 17 Jul 1999 12:23:35 -0400
Lines: 31
Message-ID: <3790AE07.5087C480@MA.UltraNet.Com>
References: <7kg08v$iht$1@news.latrobe.edu.au> <37706667.3038342@netnews.worldnet.att.net> <Pine.NXT.4.20.9906222324400.20271-100000@Tomobiki-Cho.CAC.Washington.EDU>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
X-Trace: 4cPM/G26chvbdzT1O7FnvHhJLz6R+2IRQVLzuyKUooI=
X-Complaints-To: abuse@rcn.com
NNTP-Posting-Date: 17 Jul 1999 16:24:01 GMT
X-Accept-Language:  en,en-US,en-GB,es
X-Mailer:  Mozilla 4.6 [en] (Win95; U)
Xref: news3.best.com alt.sys.pdp10:5466

Mark Crispin wrote:
> 
> > BLISS was never looked upon very favorably in 36-bit-land for a couple
> > of reasons.  One is that MACRO was the "true language" of 36-bit
> > implementation;  anything done in BLISS was looked at askance by most
> > 36-bitters.
> 
> For a good reason.  Anything written in BLISS was pretty much
> unmaintainable by the vast majority of 36-bitters.  ...

One might get that impression from reading some of this string.


> As I recall, the code generated by the BLISS compiler was horrible.  Most
> C compilers on x86, Alpha, etc. can generate better code than a human can
> write assembly by hand.  The situation was the exact opposite on the
> PDP-10.

BLISS-10's code quality was nothing to write home about - better than
ALGOL-10 worse than FORTRAN-10 and perhaps KCC-10; perhaps on a par with
F40.  (Note the irony: without BLISS-10 to implement FORTRAN-10, I
wonder whether there would ever have been an optimizing replacement for
F40.  Without an optimizing FORTRAN, there would have been a lot less
demand for KL's).

I think BLISS-36 would surprise us on occasion.  (I can remember it
recognizing a common subexpression over such a great distance that we
hadn't even realized there were two uses of the same expression).
				/AHM
-- 
Alan Howard Martin				AMartin@MA.UltraNet.Com


Article 5475 of alt.sys.pdp10:
Path: news3.best.com!news2.best.com!news.maxwell.syr.edu!nntp.primenet.com!nntp.gctr.net!nuq-peer.news.verio.net!ord-feed.news.verio.net!feed.news.verio.net!mozo.cc.purdue.edu!news.iupui.edu!haystack!mhwood
From: "Mark H. Wood" <mhwood@haystack.Bogusnet.IN.US>
Newsgroups: alt.sys.pdp10
Subject: Re: BLISS-10
Date: Sat, 17 Jul 1999 14:39:18 GMT
Organization: La Petite Hackerie
Lines: 30
Message-ID: <FF0rDI.FHy@Ameritech.net>
References: <7kg08v$iht$1@news.latrobe.edu.au> <7kh5pj$fe0@dfw-ixnews7.ix.netcom.com> <7kigdi$49t$3@autumn.news.rcn.net> <37890629.71FB82A6@MA.UltraNet.Com> <7mcjpe$k1p$1@autumn.news.rcn.net>
NNTP-Posting-Host: mhw.ulib.iupui.edu
User-Agent: tin/pre-1.4-19990517 ("Psychonaut") (UNIX) (Linux/2.2.5 (i586))
Xref: news3.best.com alt.sys.pdp10:5475

jmfbahciv@aol.com wrote:
> In article <37890629.71FB82A6@MA.UltraNet.Com>,
>    "Alan H. Martin" <AMartin@MA.UltraNet.Com> wrote:
>>jmfbahciv@aol.com wrote:
>>> 
>>> In article <7kh5pj$fe0@dfw-ixnews7.ix.netcom.com>,
>>>    bugs@freebsd.netcom.com (Mark Hittinger) wrote:
>>....
>>> >Bliss-11 may have been Decus, but a source license
>>> >for Bliss-32 was probably a lot harder to get :-).  Bliss-10 came with
>>> >source code in the tradition of 10-ness and was not Decus.
>>> 
>>> I don't think we shipped the sources.   I'd have to look at the
>>> SPDs to verify this.  But I don't think there was an SPD for
>>> BLISS-10, hence the customer couldn't "order" it.
>>
>>BLISS-10 sources were definitely available to the field, although it 
> probably
>>was only supported for building supported products written in BLISS-10.

> Alan, think.  One didn't need BLISS-10 sources in order to build
> products written in BLISS-10.

Well, I don't know how we got BLISS-10 sources but at least one
customer had 'em, because I crudely hacked in something we needed
w.r.t. JSYS when we moved from TOPS-10 to TOPS-20.
-- 
-- 
Mark H. Wood, radical centrist     OpenPGP ID 876A8B75     mhwood@ameritech.net
Why do we always draw organization charts upside down?


Article 5479 of alt.sys.pdp10:
Path: news3.best.com!news2.best.com!feed1.news.rcn.net!rcn!d2
From: jmfbahciv@aol.com
Newsgroups: alt.sys.pdp10
Subject: Re: BLISS-10
Date: Mon, 19 Jul 99 09:07:20 GMT
Organization: UltraNet Communications, Inc.
Lines: 53
Message-ID: <7mv0pe$52h$1@autumn.news.rcn.net>
References: <7kg08v$iht$1@news.latrobe.edu.au> <7kh5pj$fe0@dfw-ixnews7.ix.netcom.com> <7kigdi$49t$3@autumn.news.rcn.net> <37890629.71FB82A6@MA.UltraNet.Com> <7mcjpe$k1p$1@autumn.news.rcn.net> <FF0rDI.FHy@Ameritech.net>
X-Trace: P+kOviWeRzTtOdxEXz1Wd9akgtFjrBPl7fZhNsrXMR4=
X-Complaints-To: abuse@rcn.com
NNTP-Posting-Date: 19 Jul 1999 11:05:18 GMT
X-Newsreader:  News Xpress Version 1.0 Beta #4
Xref: news3.best.com alt.sys.pdp10:5479

In article <FF0rDI.FHy@Ameritech.net>,
   "Mark H. Wood" <mhwood@haystack.Bogusnet.IN.US> wrote:
>jmfbahciv@aol.com wrote:
>> In article <37890629.71FB82A6@MA.UltraNet.Com>,
>>    "Alan H. Martin" <AMartin@MA.UltraNet.Com> wrote:
>>>jmfbahciv@aol.com wrote:
>>>> 
>>>> In article <7kh5pj$fe0@dfw-ixnews7.ix.netcom.com>,
>>>>    bugs@freebsd.netcom.com (Mark Hittinger) wrote:
>>>....
>>>> >Bliss-11 may have been Decus, but a source license
>>>> >for Bliss-32 was probably a lot harder to get :-).  Bliss-10 came 
with
>>>> >source code in the tradition of 10-ness and was not Decus.
>>>> 
>>>> I don't think we shipped the sources.   I'd have to look at the
>>>> SPDs to verify this.  But I don't think there was an SPD for
>>>> BLISS-10, hence the customer couldn't "order" it.
>>>
>>>BLISS-10 sources were definitely available to the field, although it 
>> probably
>>>was only supported for building supported products written in BLISS-10.
>
>> Alan, think.  One didn't need BLISS-10 sources in order to build
>> products written in BLISS-10.
>
>Well, I don't know how we got BLISS-10 sources but at least one
>customer had 'em, because I crudely hacked in something we needed
>w.r.t. JSYS when we moved from TOPS-10 to TOPS-20.

I was commenting on the edit history that was posted.
I originally stated that I didn't know if there was 
a tape available for customers to get it by "honest"
means.  A lot of stuff was carried into customer sites
and left there under the table, depending on the customer's
needs.  I do know that Magee never built a BLISS tape (she
would have been complaining to me for hundreds of hours).
So that means that a tape wasn't submitted to SDC from
1976 or 1977 on.  I also do know that, even though a 
release of Bliss-10 was highly desirable (this was in 1982),
the logistics of doing a release ala the Project Notebook
was impossible.

One of the things that I started, was a formal review of
the SDC paperwork with each monitor release.  A part of that
review involved going over all of the SPDs to assess the
impact and, more importantly, the deimpact of anything
that would affect the tape submission.  I don't recall
seeing a Bliss SPD.

/BAH

Subtract a hundred and four for e-mail.


Article 5477 of alt.sys.pdp10:
Path: news3.best.com!news1.best.com!newshub.northeast.verio.net!netnews.com!feed1.news.rcn.net!rcn!not-for-mail
From: "Alan H. Martin" <AMartin@MA.UltraNet.Com>
Newsgroups: alt.sys.pdp10
Subject: Re: BLISS-10
Date: Sun, 18 Jul 1999 10:14:57 -0400
Lines: 24
Message-ID: <3791E161.2D1D82FE@MA.UltraNet.Com>
References: <7kg08v$iht$1@news.latrobe.edu.au> <37706667.3038342@netnews.worldnet.att.net> <Pine.NXT.4.20.9906222324400.20271-100000@Tomobiki-Cho.CAC.Washington.EDU> <3790AE07.5087C480@MA.UltraNet.Com> <7msaq4$mtr$2@autumn.news.rcn.net>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
X-Trace: M/H/KmvRBALE5cLR7UAe6p1sUs7odyLkfjUKs0UW/68=
X-Complaints-To: abuse@rcn.com
NNTP-Posting-Date: 18 Jul 1999 14:15:27 GMT
X-Accept-Language:  en,en-US,en-GB,es
X-Mailer:  Mozilla 4.6 [en] (Win95; U)
Xref: news3.best.com alt.sys.pdp10:5477

jmfbahciv@aol.com wrote:
> 
> In article <3790AE07.5087C480@MA.UltraNet.Com>,
>    "Alan H. Martin" <AMartin@MA.UltraNet.Com> wrote:
> >
> >....  (Note the irony: without BLISS-10 to implement FORTRAN-10, I
> >wonder whether there would ever have been an optimizing replacement for
> >F40.  Without an optimizing FORTRAN, there would have been a lot less
> >demand for KL's).
> 
> Why do you think that?

DEC invested money in replacing F40 with FORTRAN-10/20 because the
customers said it was necessary (just like those TU7x's).  The last I
heard, FORTRAN-10/20 had the highest PDP-10 market penetration of any
DEC layered product.

A product like FORTRAN-10/20 couldn't have been developed for nearly the
same cost in people, time and money without some implementation language
like BLISS-10.  What BLISS substitutes were already available on the -10
in the early 70's?
				/AHM
-- 
Alan Howard Martin				AMartin@MA.UltraNet.Com


Article 5481 of alt.sys.pdp10:
Path: news3.best.com!news1.best.com!news2.best.com!feed1.news.rcn.net!rcn!d2
From: jmfbahciv@aol.com
Newsgroups: alt.sys.pdp10
Subject: Re: BLISS-10
Date: Mon, 19 Jul 99 09:17:18 GMT
Organization: UltraNet Communications, Inc.
Lines: 40
Message-ID: <7mv1bu$52h$3@autumn.news.rcn.net>
References: <7kg08v$iht$1@news.latrobe.edu.au> <37706667.3038342@netnews.worldnet.att.net> <Pine.NXT.4.20.9906222324400.20271-100000@Tomobiki-Cho.CAC.Washington.EDU> <3790AE07.5087C480@MA.UltraNet.Com> <7msaq4$mtr$2@autumn.news.rcn.net> <3791E161.2D1D82FE@MA.UltraNet.Com>
X-Trace: JLbrQl37/XJ/lI3gWtVfPZEYzfg8puRBHiWGHBjxBI0=
X-Complaints-To: abuse@rcn.com
NNTP-Posting-Date: 19 Jul 1999 11:15:10 GMT
X-Newsreader:  News Xpress Version 1.0 Beta #4
Xref: news3.best.com alt.sys.pdp10:5481

In article <3791E161.2D1D82FE@MA.UltraNet.Com>,
   "Alan H. Martin" <AMartin@MA.UltraNet.Com> wrote:
>jmfbahciv@aol.com wrote:
>> 
>> In article <3790AE07.5087C480@MA.UltraNet.Com>,
>>    "Alan H. Martin" <AMartin@MA.UltraNet.Com> wrote:
>> >
>> >....  (Note the irony: without BLISS-10 to implement FORTRAN-10, I
>> >wonder whether there would ever have been an optimizing replacement for
>> >F40.  Without an optimizing FORTRAN, there would have been a lot less
>> >demand for KL's).
>> 
>> Why do you think that?
>
>DEC invested money in replacing F40 with FORTRAN-10/20 because the
>customers said it was necessary (just like those TU7x's).  The last I
>heard, FORTRAN-10/20 had the highest PDP-10 market penetration of any
>DEC layered product.
>
>A product like FORTRAN-10/20 couldn't have been developed for nearly the
>same cost in people, time and money without some implementation language
>like BLISS-10.  What BLISS substitutes were already available on the -10
>in the early 70's?

But you've got your years so screwed up.  That project started when
the KI was getting announced.

DEC hadn't developed a compiler that could be used for compiler
implementations.  So they did the usual thing that was done back
then and acquired work done at CMU ala G. Bell.  Norma Abel
(I think) could talk about why they picked that approach to 
the implementation lanaguage.  Sara was there, too, but I don't 
know if she was in that part of the planning stages.  I was a
mere Tape Prep type then, so I usually got the info at the
other end of the development cycle.  I do remember that the
work was getting done on 3-5.

/BAH

Subtract a hundred and four for e-mail.


Article 5482 of alt.sys.pdp10:
Path: news3.best.com!news1.best.com!144.212.100.101.MISMATCH!newsfeed.mathworks.com!feeder.qis.net!feed1.news.rcn.net!rcn!d2
From: jmfbahciv@aol.com
Newsgroups: alt.sys.pdp10
Subject: Re: BLISS-10
Date: Mon, 19 Jul 99 09:58:55 GMT
Organization: UltraNet Communications, Inc.
Lines: 48
Message-ID: <7mv3pv$mia$1@autumn.news.rcn.net>
References: <7kg08v$iht$1@news.latrobe.edu.au> <37706667.3038342@netnews.worldnet.att.net> <Pine.NXT.4.20.9906222324400.20271-100000@Tomobiki-Cho.CAC.Washington.EDU> <3790AE07.5087C480@MA.UltraNet.Com> <7msaq4$mtr$2@autumn.news.rcn.net> <3791E161.2D1D82FE@MA.UltraNet.Com> <7mv1bu$52h$3@autumn.news.rcn.net>
X-Trace: 5IhIGNyWHvLbGSbd0t1XZAfxnFdcPN2c/rkYPl1eAjI=
X-Complaints-To: abuse@rcn.com
NNTP-Posting-Date: 19 Jul 1999 11:56:47 GMT
X-Newsreader:  News Xpress Version 1.0 Beta #4
Xref: news3.best.com alt.sys.pdp10:5482

In article <7mv1bu$52h$3@autumn.news.rcn.net>, jmfbahciv@aol.com wrote:
>In article <3791E161.2D1D82FE@MA.UltraNet.Com>,
>   "Alan H. Martin" <AMartin@MA.UltraNet.Com> wrote:
>>jmfbahciv@aol.com wrote:
>>> 
>>> In article <3790AE07.5087C480@MA.UltraNet.Com>,
>>>    "Alan H. Martin" <AMartin@MA.UltraNet.Com> wrote:
>>> >
>>> >....  (Note the irony: without BLISS-10 to implement FORTRAN-10, I
>>> >wonder whether there would ever have been an optimizing replacement 
for
>>> >F40.  Without an optimizing FORTRAN, there would have been a lot less
>>> >demand for KL's).
>>> 
>>> Why do you think that?
>>
>>DEC invested money in replacing F40 with FORTRAN-10/20 because the
>>customers said it was necessary (just like those TU7x's).  The last I
>>heard, FORTRAN-10/20 had the highest PDP-10 market penetration of any
>>DEC layered product.
>>
>>A product like FORTRAN-10/20 couldn't have been developed for nearly the
>>same cost in people, time and money without some implementation language
>>like BLISS-10.  What BLISS substitutes were already available on the -10
>>in the early 70's?
>
>But you've got your years so screwed up.  That project started when
>the KI was getting announced.
<snip>

And we're definitely talking about very different projects.
F40 was FortranII.  Fortran-10/20 was Fortran IV.  I don't
remmeber when the optimizer was done nor who did it.  I got
a feeling it was that guy :-) but I can't remember his name.
Somebody who disappeared fairly quickly?

The Fortran IV project had problems getting it tested.  So
they offered a reward for every bug found.  I won $1.00.  And
that was before and after the time DEC bought the RCA building
and moved to Marlboro.  I don't remember the optimiser being
a part of that test bed.

/BAH


/BAH

Subtract a hundred and four for e-mail.


Article 5484 of alt.sys.pdp10:
Path: news3.best.com!news2.best.com!feed1.news.rcn.net!rcn!not-for-mail
From: "Alan H. Martin" <AMartin@MA.UltraNet.Com>
Newsgroups: alt.sys.pdp10
Subject: Re: BLISS-10
Date: Mon, 19 Jul 1999 11:19:49 -0400
Lines: 38
Message-ID: <37934215.3F11A0C4@MA.UltraNet.Com>
References: <7kg08v$iht$1@news.latrobe.edu.au> <37706667.3038342@netnews.worldnet.att.net> <Pine.NXT.4.20.9906222324400.20271-100000@Tomobiki-Cho.CAC.Washington.EDU> <3790AE07.5087C480@MA.UltraNet.Com> <7msaq4$mtr$2@autumn.news.rcn.net> <3791E161.2D1D82FE@MA.UltraNet.Com> <7mv1bu$52h$3@autumn.news.rcn.net> <7mv3pv$mia$1@autumn.news.rcn.net>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
X-Trace: OMOIUWZCQiP4ilzsc7Vc1CDTU4nAvF5Op2cFBBb5Xls=
X-Complaints-To: abuse@rcn.com
NNTP-Posting-Date: 19 Jul 1999 15:20:28 GMT
X-Accept-Language:  en,en-US,en-GB,es
X-Mailer:  Mozilla 4.6 [en] (Win95; U)
Xref: news3.best.com alt.sys.pdp10:5484

jmfbahciv@aol.com wrote:
> 
> In article <7mv1bu$52h$3@autumn.news.rcn.net>, jmfbahciv@aol.com wrote:
> >In article <3791E161.2D1D82FE@MA.UltraNet.Com>,
> >   "Alan H. Martin" <AMartin@MA.UltraNet.Com> wrote:
> >>A product like FORTRAN-10/20 couldn't have been developed for nearly the
> >>same cost in people, time and money without some implementation language
> >>like BLISS-10.  What BLISS substitutes were already available on the -10
> >>in the early 70's?
> >
> >But you've got your years so screwed up.  That project started when
> >the KI was getting announced.

It's not as if some substitute language bloomed and died before
BLISS-10.


> And we're definitely talking about very different projects.
> F40 was FortranII.  Fortran-10/20 was Fortran IV.  ...

And how many million-dollar systems do you think would have been bought
with nothing more than a non-optimizing FORTRAN-II?

A big part of the -10 market needed a product (F10) whose timely
delivery depended upon BLISS-10.


>... I don't
> remmeber when the optimizer was done nor who did it.  I got
> a feeling it was that guy :-) but I can't remember his name.
> Somebody who disappeared fairly quickly?

Frank did most of the parser; Norma did (at least) most of the global
optimizer; Sara did (at least) most of the local optimizer and code
generation.
				/AHM
-- 
Alan Howard Martin				AMartin@MA.UltraNet.Com


Article 5491 of alt.sys.pdp10:
Path: news3.best.com!news1.best.com!144.212.100.101.MISMATCH!newsfeed.mathworks.com!europa.netcrusader.net!207.172.3.37!feed1.news.rcn.net!rcn!d13
From: jmfbahciv@aol.com
Newsgroups: alt.sys.pdp10
Subject: Re: BLISS-10
Date: Thu, 22 Jul 99 08:45:15 GMT
Organization: UltraNet Communications, Inc.
Lines: 54
Message-ID: <7n6skn$2l$2@autumn.news.rcn.net>
References: <7kg08v$iht$1@news.latrobe.edu.au> <37706667.3038342@netnews.worldnet.att.net> <Pine.NXT.4.20.9906222324400.20271-100000@Tomobiki-Cho.CAC.Washington.EDU> <3790AE07.5087C480@MA.UltraNet.Com> <7msaq4$mtr$2@autumn.news.rcn.net> <3791E161.2D1D82FE@MA.UltraNet.Com> <7mv1bu$52h$3@autumn.news.rcn.net> <7mv3pv$mia$1@autumn.news.rcn.net> <37934215.3F11A0C4@MA.UltraNet.Com>
X-Trace: ZB2r1OFYwjAl+pm+RZaX/4Uz5B+RWAx1JqiAFIi2whw=
X-Complaints-To: abuse@rcn.com
NNTP-Posting-Date: 22 Jul 1999 10:43:35 GMT
X-Newsreader:  News Xpress Version 1.0 Beta #4
Xref: news3.best.com alt.sys.pdp10:5491

In article <37934215.3F11A0C4@MA.UltraNet.Com>,
   "Alan H. Martin" <AMartin@MA.UltraNet.Com> wrote:
>jmfbahciv@aol.com wrote:
>> 
>> In article <7mv1bu$52h$3@autumn.news.rcn.net>, jmfbahciv@aol.com wrote:
>> >In article <3791E161.2D1D82FE@MA.UltraNet.Com>,
>> >   "Alan H. Martin" <AMartin@MA.UltraNet.Com> wrote:
>> >>A product like FORTRAN-10/20 couldn't have been developed for nearly 
the
>> >>same cost in people, time and money without some implementation 
language
>> >>like BLISS-10.  What BLISS substitutes were already available on the 
-10
>> >>in the early 70's?
>> >
>> >But you've got your years so screwed up.  That project started when
>> >the KI was getting announced.
>
>It's not as if some substitute language bloomed and died before
>BLISS-10.
>
>
>> And we're definitely talking about very different projects.
>> F40 was FortranII.  Fortran-10/20 was Fortran IV.  ...
>
>And how many million-dollar systems do you think would have been bought
>with nothing more than a non-optimizing FORTRAN-II?

I think, in the KL era you mentioned, most of them.  And stop
talking about F-II.  That was gone by KI days.

>
>A big part of the -10 market needed a product (F10) whose timely
>delivery depended upon BLISS-10.

Where did you get all this info?  

>
>
>>... I don't
>> remmeber when the optimizer was done nor who did it.  I got
>> a feeling it was that guy :-) but I can't remember his name.
>> Somebody who disappeared fairly quickly?
>
>Frank did most of the parser; Norma did (at least) most of the global
>optimizer; Sara did (at least) most of the local optimizer and code
>generation.

That's right; Frank.  Somedays I wish I was gifted with the
ability to remember names :-).

/BAH

Subtract a hundred and four for e-mail.


Article 5492 of alt.sys.pdp10:
Path: news3.best.com!news2.best.com!feed1.news.rcn.net!rcn!d13
From: jmfbahciv@aol.com
Newsgroups: alt.sys.pdp10
Subject: Re: TU10/20/7x (was: SA10 front panel (was Re: Multiple stupid TOPS-10 questions...))
Date: Thu, 22 Jul 99 08:47:40 GMT
Organization: UltraNet Communications, Inc.
Lines: 43
Message-ID: <7n6sp8$2l$3@autumn.news.rcn.net>
References: <Pine.LNX.4.10.9905171202310.15552-100000@bony.umtec.com> <7mgjsk$9fr$1@shell3.ba.best.com> <7mhtf7$ns2$3@autumn.news.rcn.net> <V2cj3.16$C4.2784@news2.voicenet.com> <7mrj5n$hrj$1@shell3.ba.best.com> <iXHk3.320$O2.15699@ord-read.news.verio.net> <7n4a1k$d0u$1@autumn.news.rcn.net> <R4pl3.27$XM5.798@ord-read.news.verio.net>
X-Trace: Xkl5WKVtmfYIvgtYA6morZj/yzh+fOcdZcXUnQLCP4c=
X-Complaints-To: abuse@rcn.com
NNTP-Posting-Date: 22 Jul 1999 10:46:00 GMT
X-Newsreader:  News Xpress Version 1.0 Beta #4
Xref: news3.best.com alt.sys.pdp10:5492

In article <R4pl3.27$XM5.798@ord-read.news.verio.net>,
   jeverett@wwa.DEFEAT.UCE.BOTS.com (John Everett) wrote:
>In article <7n4a1k$d0u$1@autumn.news.rcn.net>, jmfbahciv@aol.com says...
>>
>>In article <iXHk3.320$O2.15699@ord-read.news.verio.net>,
>>   jeverett@wwa.DEFEAT.UCE.BOTS.com (John Everett) wrote:
>>>When I was young, foolish, and a bit impulsive, (and also responsible 
for 
>>MTA 
>>>support in TOPS-10) I wrote a somewhat intemperate memo addressed to 
>>>practically anyone who was anyone in 36-bit land. 
>>>
>>Jeez, it certainly took them long enough.  Jim had a chuckle
>>because, one day, DEC was talking about a deal with STC in
>>one end of the building (MRO) and talking about suing them
>>in the other end of the building.
>
>It should be said the the above referenced memo was written while we were 
>still in Maynard.

Which was pre-1975.  But, if you wrote the memo, it was earlier than
that.  Wasn't it?  I don't remember seeing you when the move to
Marlboro happened.

>
>>When the KL was first powered up, it blew the Marlboro
>>plant's electric grid.  Is this true?  
>
>I don't recall that this happened. I'm also trying to recall where the KL 
was 
>first powered up, Maynard or Marlboro. I know a lot of the design and 
>engineering took place on 5-5 in Maynard, but can't recall where the first 
>prototype was built and tested. Anyone? RCC, if you're still lurking 
perhaps 
>you remember.
>

Or Jud; or Alan.  But I haven't anything about him for years.

/BAH


Subtract a hundred and four for e-mail.


Article 5498 of alt.sys.pdp10:
Path: news3.best.com!news2.best.com!news.maxwell.syr.edu!news.idt.net!nntp.farm.idt.net!news
From: "Chris Ward" <ccw@idt.net>
Newsgroups: alt.sys.pdp10
Subject: Re: BLISS-10/F40/F10
Date: Sat, 24 Jul 1999 00:32:35 -0400
Organization: IDT (Best News In The World)
Lines: 50
Message-ID: <7nbfl1$p8b@nnrp1.farm.idt.net>
References: <7kg08v$iht$1@news.latrobe.edu.au> <37706667.3038342@netnews.worldnet.att.net> <Pine.NXT.4.20.9906222324400.20271-100000@Tomobiki-Cho.CAC.Washington.EDU> <3790AE07.5087C480@MA.UltraNet.Com> <7msaq4$mtr$2@autumn.news.rcn.net> <3791E161.2D1D82FE@MA.UltraNet.Com> <7mv1bu$52h$3@autumn.news.rcn.net> <7mv3pv$mia$1@autumn.news.rcn.net> <37934215.3F11A0C4@MA.UltraNet.Com>
NNTP-Posting-Host: ppp-32.ts-1-bay.hob.idt.net
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 5.00.2314.1300
X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2314.1300
Xref: news3.best.com alt.sys.pdp10:5498


Alan H. Martin <AMartin@MA.UltraNet.Com> wrote in message
news:37934215.3F11A0C4@MA.UltraNet.Com...
> jmfbahciv@aol.com wrote:
> >
> > In article <7mv1bu$52h$3@autumn.news.rcn.net>, jmfbahciv@aol.com
wrote:
> > >In article <3791E161.2D1D82FE@MA.UltraNet.Com>,
> > >   "Alan H. Martin" <AMartin@MA.UltraNet.Com> wrote:
> > >>A product like FORTRAN-10/20 couldn't have been developed for
nearly the
> > >>same cost in people, time and money without some
implementation language
> > >>like BLISS-10.  What BLISS substitutes were already available
on the -10
> > >>in the early 70's?
> > >
> > >But you've got your years so screwed up.  That project started
when
> > >the KI was getting announced.
>
> It's not as if some substitute language bloomed and died before
> BLISS-10.
>
>
> > And we're definitely talking about very different projects.
> > F40 was FortranII.  Fortran-10/20 was Fortran IV.  ...
>
> And how many million-dollar systems do you think would have been
bought
> with nothing more than a non-optimizing FORTRAN-II?
>
At least when I worked with it, the sematics of the F40 language was
almost identical to the F10 language, at least when we were
implementing Stewart's XRAY76 programs on the 10.  F40 implemented
subroutines via JSR's (I think), but whatever it did, it left the
return address in the first address of the subroutine, which was
actually the "classical" Fortran (II or IV) manner.  F10 implemented
subroutine calls via PUSHJs and POPJs (but I do not think it was
recursive).  I also remember the compiler generating optimizer
errors - the same type which plauge compilers today.  So we turned
the optimizer off.  Or tested the results VERY thuroughly.

But F40 was indeed Fortran IV, and not Fortran II.  Fortran II had a
quite different look to the PRINT statement (I think leaving out the
device number).  And it had the full range of conditionals, and more
complete than the IBM 1130 FORTRAN I started out with (only
arithmetic IFs), and that was called FORTRAN IV.




Article 2537 of alt.sys.pdp10:
Path: nntp1.ba.best.com!news1.best.com!newsfeed.mathworks.com!wn3feed!worldnet.att.net!207.207.0.27!nntp2.aus1.giganews.com!nntp3.aus1.giganews.com!news1.giganews.com.POSTED!not-for-mail
From: Timothy Stark <sword7@grace.speakeasy.org>
Subject: Re: BLT 17,17 bug had been fixed but..
Newsgroups: alt.sys.pdp10
References: <6TiY5.46657$%j3.530564@news6.giganews.com> <90teal$10b$1@bob.news.rcn.net> <J9AY5.115963$DG3.2253745@news2.giganews.com> <9105q9$8d3$1@bob.news.rcn.net> <gCeZ5.121937$DG3.2418389@news2.giganews.com> <9154bi$8am$1@bob.news.rcn.net> <UpwZ5.77193$IP1.2582629@news1.giganews.com> <917jrq$2s6$2@bob.news.rcn.net> <Z7YZ5.59759$%j3.763939@news6.giganews.com> <91a8c1$fj$1@bob.news.rcn.net> <HD4_5.84262$IP1.2834867@news1.giganews.com> <874s06ex9i.fsf@k9.prep.synonet.com>
User-Agent: tin/1.4.2-20000205 ("Possession") (UNIX) (Linux/2.2.17 (i686))
Lines: 28
Message-ID: <3rc_5.85329$IP1.2884825@news1.giganews.com>
NNTP-Posting-Date: Thu, 14 Dec 2000 17:09:51 CST
Organization: Giganews.Com - Premium News Outsourcing
X-Trace: sv2-wP74oLoWgiANa58r0IWH5dWl5riKBhoSbVcLcRkMEZ2tuyMoixgHbgaDMfiTxuVY1URS+YXUogeBshB!N0e2Qe3YjxbgPcCtBo7JfWUP0NM=
X-Complaints-To: abuse@GigaNews.Com
X-Abuse-Info: Please be sure to forward a copy of ALL headers
X-Abuse-Info: Otherwise we will be unable to process your complaint properly
Date: Thu, 14 Dec 2000 23:09:52 GMT
Xref: nntp1.ba.best.com alt.sys.pdp10:2537

Paul Repacholi <prep@prep.synonet.com> wrote:
> Timothy Stark <sword7@grace.speakeasy.org> writes:

>> jmfbahciv@aol.com wrote:
>> > Neither of these BLISSes were shipped with sources so you'll only
>> > find the EXEs (and in the case of -36, RELs and librarys).
>> 
>> I am surprised.  However, DEC recently released BLISS-32/64 with
>> source codes to public but they are for OpenVMS only.  I have them.

> What, with SOURCE??? Are you sure of that? Where?

Yes, BLISS-11 sources.

Its directory is http://www.openvms.compaq.com/freeware/BLISS/

> The BLISS-11 sources are out there now, so that is MOST of the CMU
> BLISS-10. Oh for a complete set of BLISS compilers, including
> the memory compaction stuff...

-- Tim Stark

-- 
Timothy Stark	<><	Inet: sword7@speakeasy.org
--------------------------------------------------------------------------
"For God so loved the world, that he gave his only begotten Son, that 
whosoever believeth in him should not perish, but have everlasting life.
Amen." -- John 3:16 (King James Version Bible)


Article 2671 of alt.sys.pdp10:
Path: nntp1.ba.best.com!news2.best.com!news.maxwell.syr.edu!feed2.news.rcn.net!feed1.news.rcn.net!rcn!not-for-mail
From: "Alan H. Martin" <AMartin@MA.UltraNet.Com>
Newsgroups: alt.sys.pdp10
Subject: Re: BLT 17,17 bug had been fixed but..
Date: Tue, 19 Dec 2000 01:24:13 -0500
Lines: 89
Message-ID: <3A3EFF0D.A94ECC02@MA.UltraNet.Com>
References: <6TiY5.46657$%j3.530564@news6.giganews.com> <90teal$10b$1@bob.news.rcn.net> <J9AY5.115963$DG3.2253745@news2.giganews.com> <9105q9$8d3$1@bob.news.rcn.net> <gCeZ5.121937$DG3.2418389@news2.giganews.com> <9154bi$8am$1@bob.news.rcn.net> <UpwZ5.77193$IP1.2582629@news1.giganews.com> <917jrq$2s6$2@bob.news.rcn.net> <Z7YZ5.59759$%j3.763939@news6.giganews.com> <91a8c1$fj$1@bob.news.rcn.net> <HD4_5.84262$IP1.2834867@news1.giganews.com> <874s06ex9i.fsf@k9.prep.synonet.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
X-Trace: UmFuZG9tSVZA4AQx6q5Bbo2O6k04fZSE8aaU38aoPOchUEOQhEn0X/9Luz1wVdxW
X-Complaints-To: abuse@rcn.com
NNTP-Posting-Date: 19 Dec 2000 06:24:52 GMT
X-Mailer:  Mozilla 4.75 [en] (Win95; U)
X-Accept-Language:  en,en-US,en-GB,es
Xref: nntp1.ba.best.com alt.sys.pdp10:2671

Paul Repacholi wrote:
> 
> Timothy Stark <sword7@grace.speakeasy.org> writes:
> 
> >...  DEC recently released BLISS-32/64 with
> > source codes to public but they are for OpenVMS only.  I have them.
> 
> What, with SOURCE??? Are you sure of that? Where?

I'd like to see those VMSINSTAL freeware savesets of BLISS-32/64
expanded someplace...


> The BLISS-11 sources are out there now, so that is MOST of the CMU
> BLISS-10.  ...

BLISS-10 is in:

http://pdp-10.trailing-edge.com/pdp-10/BB-D868B-BM_TOPS20_V3A_2020_DIST.html

BLISS-11 is in:

http://pdp-10.trailing-edge.com/www/lib10/0325/index.html


Here is a cheat sheet called MODULE.TXT I wrote up in the early 80's by
whiffreading the BLISS-10 sources:

.SUP	.BLI	Name	Purpose

AD0	LOADDR	LOADDR	Predicates and helpers for I/X/Y field generation
AD2	H2ADDR	ADDRESS	I/X/Y field and byte pointer code generation
AR2N	AR2N	ARITH	Unary/binary operator code generation (CTCE/DELAY/CODE)
AS3	H3ASSY	H3ASSY	Makes listing palatable to MACRO-10
	BEGIN	BEGIN	Common parameter file (require file)
	BL10ER	BL10ER	INDEX output - indices into BLIS10.ERR (require file)
CC3	H3CCL	H3CCL	COMPIL-class command routines
CN1N	CN1N	CONTROL	Manipulates code linked lists, generates control code
CN2N	CN2N	H2CNTR	Generates prolog/epilog code
CN3	CN3	H3CNTR	Generates stack, timer code, version numbers
DE0N	DE0N	LODECL	Helpers for declaration scope, forwards
DE1N	DE1N	H1DECL	Declaration processing
DE3N	DE3N	H3DECL	Common code for routine/funct/struct decls, machops
DR0N	DR0N	LODRIV	I/O, error, listing utility routines
DR3N	DR3N	DRIVER	Main driver code, listing file routines
GT0	LOGTRE	LOGTRE	Graph table helper routines
GT1N	GT1N	H1GTRE	Main graph table routines
GT2	H2GTRE	H2GTRE	Special case graph table routines
	INDEX	INDEX	Utility program to convert .ERR file into .BLI file
LD3N	LD3N	H3LDIN	REL generator, macro listing routines
LOIO.MAC	LOIO	Low level I/O, including command scanner
LS0	LOLSTP	LOLSTP	Low level linked list processing routines
LS3N	LS3N	H3LSTP	Predicates and helpers for code list processing
LX0N	LX0N	LOLEXA	Lexical analyzer
LX1N	LX1N	H1LEXA	Scope related lexical analysis helpers
LX3	H3LEXA	H3LEXA	Purges GENSYMS list at the end of a routine
MA0N	MA0N	LOMACR	Macro expansion processing
MA1N	MA1N	H1MACR	Macro definition processing
NOONCE.MAC	LOONCE	Once only initialization code
PE3	H3PEEP	H3PEEP	Peephole optimizer
RG2	H2REGI	H2REGI	Register allocator
RG3N	RG3N	H3REGI	Register data structure initialization
RQ1N	RQ1N	RQ1N	Require file code, and LIBRARY declarations too
	SIX12	SIX12	Debugger
SN1N	SN1N	H1SYNT	Parser routines (including top level semantics)
XR0N	XR0N	LOXREF	Cross reference data management routines
XR3N	XR3N	H3XREF	Cross reference listing routines

Compare that to the description of BLISS-11 in Wulf, Johnson, Weinstock,
Hobbs, and Geschke's _The Design of an Optimizing Compiler_:

http://www.amazon.com/exec/obidos/ASIN/0444001581
http://www.amazon.com/exec/obidos/ASIN/0444001646

Some stuff is similar; compare the routine EXPRESSION in:

http://pdp-10.trailing-edge.com/pdp-10/BB-D868B-BM_TOPS20_V3A_2020_DIST/3A-SOURCES/SN1N.BLI
with:
http://pdp-10.trailing-edge.com/www/lib10/0325/SYNTAX.BLI

However, I believe that a lot of the stuff in BLISS-11 (syntax-directed
translation-based(?) optimizations) just doesn't exist in BLISS-10.

I think BLISS-11 should be a much better hint at Common BLISS's
internals.  Another reference would be _The GEM Optimizing Compiler
System_, http://www.digital.com/DTJ808/DTJ808SC.TXT
				/AHM
-- 
Alan Howard Martin				AMartin@MA.UltraNet.Com


Article 2758 of alt.sys.pdp10:
Newsgroups: alt.sys.pdp10
Path: nntp1.ba.best.com!news1.best.com!newsfeed.mathworks.com!newsfeed.cwix.com!news.umass.edu!world!weiner
From: weiner@world.std.com (Sam Weiner)
Subject: Re: BLT 17,17 bug had been fixed but..
Message-ID: <G5wII5.n45@world.std.com>
Date: Thu, 21 Dec 2000 04:44:28 GMT
References: <6TiY5.46657$%j3.530564@news6.giganews.com> <91a8c1$fj$1@bob.news.rcn.net> <HD4_5.84262$IP1.2834867@news1.giganews.com> <874s06ex9i.fsf@k9.prep.synonet.com>
Organization: The World Public Access UNIX, Brookline, MA
Lines: 40
Xref: nntp1.ba.best.com alt.sys.pdp10:2758

In article <874s06ex9i.fsf@k9.prep.synonet.com>,
Paul Repacholi  <prep@prep.synonet.com> wrote:
>Timothy Stark <sword7@grace.speakeasy.org> writes:
>
>> jmfbahciv@aol.com wrote:
>> > Neither of these BLISSes were shipped with sources so you'll only
>> > find the EXEs (and in the case of -36, RELs and librarys).
>> 
>> I am surprised.  However, DEC recently released BLISS-32/64 with
>> source codes to public but they are for OpenVMS only.  I have them.
>
>What, with SOURCE??? Are you sure of that? Where?

As others have pointed out, on the VMS freeware CD which is
also available on the net, sorry I don't have the URL handy.

>The BLISS-11 sources are out there now, so that is MOST of the CMU
>BLISS-10. Oh for a complete set of BLISS compilers, including
>the memory compaction stuff...

The sequence is BLISS-10, BLISS-11 which ran on TOPS-10/20 and
has been in the DECUS library for a while (get the book by Wulf,
good quick read,) Common BLISS which is much more like BLISS-11
than BLISS-10.  Among other things, the BLISS-10 macro (preprocessor)
facility was primitive next to BLISS-11 and Common BLISS.  I wonder
if BLISS-36 conditioned code is still in the sources.

Even before it was included on the DEC distributions, you could
get BLISS-10 from CMU if you knew the right people.  That is where
the copy I first used at Rapidata came from.  BLISS-36 was another
matter, it was like pulling teeth.  Same with RMS-10 even after it
was being used by COBOL.

When is one of the multitude of -10 emulators going to be released
so we can use this stuff instead of just talking about it?

Sam





Article 2542 of alt.sys.pdp10:
Path: nntp1.ba.best.com!news2.best.com!feed2.news.rcn.net!feed1.news.rcn.net!rcn!207-172-245-54
From: jmfbahciv@aol.com
Newsgroups: alt.sys.pdp10
Subject: Re: BLT 17,17 bug had been fixed but..
Date: Fri, 15 Dec 00 09:49:53 GMT
Organization: UltraNet Communications, Inc.
Lines: 31
Message-ID: <91ctd7$aj0$1@bob.news.rcn.net>
References: <6TiY5.46657$%j3.530564@news6.giganews.com> <90teal$10b$1@bob.news.rcn.net> <J9AY5.115963$DG3.2253745@news2.giganews.com> <9105q9$8d3$1@bob.news.rcn.net> <gCeZ5.121937$DG3.2418389@news2.giganews.com> <9154bi$8am$1@bob.news.rcn.net> <UpwZ5.77193$IP1.2582629@news1.giganews.com> <917jrq$2s6$2@bob.news.rcn.net> <Z7YZ5.59759$%j3.763939@news6.giganews.com> <91a8c1$fj$1@bob.news.rcn.net> <HD4_5.84262$IP1.2834867@news1.giganews.com>
X-Trace: UmFuZG9tSVafjF8WWaDffJwAgeNHMQyqiZiboHaj/yZP5Yo8iFCoOD2DPZO+iEgF
X-Complaints-To: abuse@rcn.com
NNTP-Posting-Date: 15 Dec 2000 10:56:39 GMT
X-Newsreader:  News Xpress Version 1.0 Beta #4
Xref: nntp1.ba.best.com alt.sys.pdp10:2542

In article <HD4_5.84262$IP1.2834867@news1.giganews.com>,
   Timothy Stark <sword7@grace.speakeasy.org> wrote:
>jmfbahciv@aol.com wrote:
>> Neither of these BLISSes were shipped with sources so you'll only
>> find the EXEs (and in the case of -36, RELs and librarys).
>
>I am surprised.  

We had a hell of a time, convincing management to ship
even the EXE.

However, DEC recently released BLISS-32/64 with
>source codes to public but they are for OpenVMS only.  I have them.

Good Grief!  Somebody is thinking with their brain?  That's
really good news.

>
>> You should have it.  Do a DIRECT BL????.* on DEC:
>> <snip>
>
>Hmm. I will check them again.  When I installed TOPS-20 v7.04, I found 
>out missing BLISS compiler in SYS: but I had seen them in CUSP tapes.
>Ok, they already had.
Well, try to be careful.  I shipped BLISS-10, BLISS-36, and
BLISS-16 so that the customers had a snowball's chance in
Massachusetts to do their own work.

/BAH

Subtract a hundred and four for e-mail.


Article 2545 of alt.sys.pdp10:
Path: nntp1.ba.best.com!news1.best.com!newsfeed.mathworks.com!sunqbc.risq.qc.ca!cpk-news-hub1.bbnplanet.com!news.gtei.net!feed2.news.rcn.net!rcn!nntp.giganews.com!nntp3.aus1.giganews.com!news2.giganews.com.POSTED!not-for-mail
From: Timothy Stark <sword7@grace.speakeasy.org>
Subject: Re: BLT 17,17 bug had been fixed but..
Newsgroups: alt.sys.pdp10
References: <6TiY5.46657$%j3.530564@news6.giganews.com> <90teal$10b$1@bob.news.rcn.net> <J9AY5.115963$DG3.2253745@news2.giganews.com> <9105q9$8d3$1@bob.news.rcn.net> <gCeZ5.121937$DG3.2418389@news2.giganews.com> <9154bi$8am$1@bob.news.rcn.net> <UpwZ5.77193$IP1.2582629@news1.giganews.com> <917jrq$2s6$2@bob.news.rcn.net> <Z7YZ5.59759$%j3.763939@news6.giganews.com> <91a8c1$fj$1@bob.news.rcn.net> <HD4_5.84262$IP1.2834867@news1.giganews.com> <91ctd7$aj0$1@bob.news.rcn.net>
User-Agent: tin/1.4.2-20000205 ("Possession") (UNIX) (Linux/2.2.17 (i686))
Lines: 35
Message-ID: <vsp_5.133896$DG3.2631970@news2.giganews.com>
NNTP-Posting-Date: Fri, 15 Dec 2000 07:58:51 CST
Organization: Giganews.Com - Premium News Outsourcing
X-Trace: sv2-sVoouhlE2TI6QIthYC9+g0wUvDOyAxKWTba7nqvXTRfSW9sg/uoWd1jvaXPDijdLBirAlwjUY7upnXD!Az7koeGkK05GlwcOG+0i9x5cfjI=
X-Complaints-To: abuse@GigaNews.Com
X-Abuse-Info: Please be sure to forward a copy of ALL headers
X-Abuse-Info: Otherwise we will be unable to process your complaint properly
Date: Fri, 15 Dec 2000 13:58:51 GMT
Xref: nntp1.ba.best.com alt.sys.pdp10:2545

jmfbahciv@aol.com wrote:
>>> You should have it.  Do a DIRECT BL????.* on DEC:
>>> <snip>
>>
>>Hmm. I will check them again.  When I installed TOPS-20 v7.04, I found 
>>out missing BLISS compiler in SYS: but I had seen them in CUSP tapes.
>>Ok, they already had.
> Well, try to be careful.  I shipped BLISS-10, BLISS-36, and
> BLISS-16 so that the customers had a snowball's chance in
> Massachusetts to do their own work.

Barb:

Hmmm.  I looked into CUSP/DIST tapes for TOPS-10 and TOPS-20.
For TOPS-10 v7.03/v7.04, I did not see any BLISS compiler in them but
just *.REQ, etc..  RMS-10 package comes with sources in written by
BLISS language. 

When I debugged ACTDAE... Yes, my debug log file showed weird code
looks like very different comparsion with code written by hands.
It was generated by a BLISS compiler.  It used stack area as local
variables.

For TOPS-20 v4.1, I found BLISS compiler but it is for TOPS-20 only.

Ok, thank you again for info.

-- Tim Stark

-- 
Timothy Stark	<><	Inet: sword7@speakeasy.org
--------------------------------------------------------------------------
"For God so loved the world, that he gave his only begotten Son, that 
whosoever believeth in him should not perish, but have everlasting life.
Amen." -- John 3:16 (King James Version Bible)


Article 2665 of alt.sys.pdp10:
Path: nntp1.ba.best.com!news1.best.com!newsfeed.mathworks.com!europa.netcrusader.net!207.172.3.44!feed2.news.rcn.net!feed1.news.rcn.net!rcn!not-for-mail
From: "Alan H. Martin" <AMartin@MA.UltraNet.Com>
Newsgroups: alt.sys.pdp10
Subject: Re: BLT 17,17 bug had been fixed but..
Date: Mon, 18 Dec 2000 23:54:20 -0500
Lines: 49
Message-ID: <3A3EE9FC.A73D5CAF@MA.UltraNet.Com>
References: <6TiY5.46657$%j3.530564@news6.giganews.com> <90teal$10b$1@bob.news.rcn.net> <J9AY5.115963$DG3.2253745@news2.giganews.com> <9105q9$8d3$1@bob.news.rcn.net> <gCeZ5.121937$DG3.2418389@news2.giganews.com> <9154bi$8am$1@bob.news.rcn.net> <UpwZ5.77193$IP1.2582629@news1.giganews.com> <917jrq$2s6$2@bob.news.rcn.net> <Z7YZ5.59759$%j3.763939@news6.giganews.com> <91a8c1$fj$1@bob.news.rcn.net>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
X-Trace: UmFuZG9tSVbZtp2djL1GtpmbpO0KxEnmzp9U/TCQTOfRW5ZNgZTkznHcqGCFHS3k
X-Complaints-To: abuse@rcn.com
NNTP-Posting-Date: 19 Dec 2000 04:54:58 GMT
To: jmfbahciv@aol.com
X-Mailer:  Mozilla 4.75 [en] (Win95; U)
X-Accept-Language:  en,en-US,en-GB,es
Xref: nntp1.ba.best.com alt.sys.pdp10:2665

jmfbahciv@aol.com wrote:
> 
> In article <Z7YZ5.59759$%j3.763939@news6.giganews.com>,
>    Timothy Stark <sword7@grace.speakeasy.org> wrote:
> 
> >I am looking for BLISS-10 compiler for TOPS-10/20 operating system.

(I'm glad you subsequently got copies).


> BLISS-10 is on the CUSP tape; filename is probably BLIS10.EXE
> among others ...

I don't see a BLISS-10 on any of the online TOPS-10 cusp tapes:
http://pdp-10.trailing-edge.com/pdp-10/CUSPMAR86BINSRC_1OF2_BB-X128B-SB.html
http://pdp-10.trailing-edge.com/pdp-10/CUSPJUL86UPD_BB-JF24A-BB.html
http://pdp-10.trailing-edge.com/pdp-10/CUST_SUP_CUSP_BB-X130C-SB.html
http://pdp-10.trailing-edge.com/pdp-10/CUSTSUPCUSPMAR86_BB-X130B-SB.html
	-or-
http://pdp-10.trailing-edge.com/pdp-10/CUST_SUP_CUSP_BB-X130C-SB.html

I wonder if at least the binary was on the Fortran layered product tape,
perhaps among other places.


>... (there were a number of GETSEGs).

Nope.  I've seen BLISS-36 run with multiple hisegs (named BLSSG1 through
BLSSG4, I think); perhaps only on TOPS-10, since
http://pdp-10.trailing-edge.com/pdp-10/BB-J939F-BM.HTML
only builds a single-hiseg BLISS-36.

However, BLISS-10 does not have multiple hisegs.


> Neither of these BLISSes were shipped with sources ...

The place Timothy Stark subsequently found BLISS-10 (``TOPS-20 V3A 2020
DISTRIBUTION TAPE 9 TRACK 1600 BPI MAGTAPE DRIVE# 273 S/N 2 SYS,
BB-D868B-BM''
http://pdp-10.trailing-edge.com/pdp-10/BB-D868B-BM_TOPS20_V3A_2020_DIST.html )

was the latest version, V7E(227), and it contains all the sources.

I still don't recall how TOPS-10 sites got BLISS-10 sources, but we
certainly had them in the field.
				/AHM
-- 
Alan Howard Martin				AMartin@MA.UltraNet.Com


Article 2685 of alt.sys.pdp10:
Path: nntp1.ba.best.com!news1.best.com!newsfeed.mathworks.com!europa.netcrusader.net!207.172.3.44!feed2.news.rcn.net!feed1.news.rcn.net!rcn!207-172-245-95
From: jmfbahciv@aol.com
Newsgroups: alt.sys.pdp10
Subject: Re: BLT 17,17 bug had been fixed but..
Date: Tue, 19 Dec 00 13:04:03 GMT
Organization: UltraNet Communications, Inc.
Lines: 65
Message-ID: <91nqad$43p$2@autumn.news.rcn.net>
References: <6TiY5.46657$%j3.530564@news6.giganews.com> <90teal$10b$1@bob.news.rcn.net> <J9AY5.115963$DG3.2253745@news2.giganews.com> <9105q9$8d3$1@bob.news.rcn.net> <gCeZ5.121937$DG3.2418389@news2.giganews.com> <9154bi$8am$1@bob.news.rcn.net> <UpwZ5.77193$IP1.2582629@news1.giganews.com> <917jrq$2s6$2@bob.news.rcn.net> <Z7YZ5.59759$%j3.763939@news6.giganews.com> <91a8c1$fj$1@bob.news.rcn.net> <3A3EE9FC.A73D5CAF@MA.UltraNet.Com>
X-Trace: UmFuZG9tSVZFBp8NudFrzvLryfLXj0mV3aOj5NqdF7j8TbVkidf3oGvG10662O5H
X-Complaints-To: abuse@rcn.com
NNTP-Posting-Date: 19 Dec 2000 14:11:25 GMT
X-Newsreader:  News Xpress Version 1.0 Beta #4
Xref: nntp1.ba.best.com alt.sys.pdp10:2685

In article <3A3EE9FC.A73D5CAF@MA.UltraNet.Com>,
   "Alan H. Martin" <AMartin@MA.UltraNet.Com> wrote:
>jmfbahciv@aol.com wrote:
>> 
>> In article <Z7YZ5.59759$%j3.763939@news6.giganews.com>,
>>    Timothy Stark <sword7@grace.speakeasy.org> wrote:
>> 
>> >I am looking for BLISS-10 compiler for TOPS-10/20 operating system.
>
>(I'm glad you subsequently got copies).
>
>
>> BLISS-10 is on the CUSP tape; filename is probably BLIS10.EXE
>> among others ...
>
>I don't see a BLISS-10 on any of the online TOPS-10 cusp tapes:
>http://pdp-10.trailing-edge.com/pdp-10/CUSPMAR86BINSRC_1OF2_BB-X128B-SB.ht
ml
>http://pdp-10.trailing-edge.com/pdp-10/CUSPJUL86UPD_BB-JF24A-BB.html
>http://pdp-10.trailing-edge.com/pdp-10/CUST_SUP_CUSP_BB-X130C-SB.html
>http://pdp-10.trailing-edge.com/pdp-10/CUSTSUPCUSPMAR86_BB-X130B-SB.html
>	-or-
>http://pdp-10.trailing-edge.com/pdp-10/CUST_SUP_CUSP_BB-X130C-SB.html
>
>I wonder if at least the binary was on the Fortran layered product tape,
>perhaps among other places.

Yes, it was there, too.  Sara had a very bad habit of shipping
code that belonged exclusively on the CUSP tape.
>
>
>>... (there were a number of GETSEGs).
>
>Nope.  I've seen BLISS-36 run with multiple hisegs (named BLSSG1 through
>BLSSG4, I think); perhaps only on TOPS-10, since
>http://pdp-10.trailing-edge.com/pdp-10/BB-J939F-BM.HTML
>only builds a single-hiseg BLISS-36.
>
>However, BLISS-10 does not have multiple hisegs.
>
>
>> Neither of these BLISSes were shipped with sources ...
>
>The place Timothy Stark subsequently found BLISS-10 (``TOPS-20 V3A 2020
>DISTRIBUTION TAPE 9 TRACK 1600 BPI MAGTAPE DRIVE# 273 S/N 2 SYS,
>BB-D868B-BM''
>http://pdp-10.trailing-edge.com/pdp-10/BB-D868B-BM_TOPS20_V3A_2020_DIST.ht
ml )
>
>was the latest version, V7E(227), and it contains all the sources.
>
>I still don't recall how TOPS-10 sites got BLISS-10 sources, but we
>certainly had them in the field.

Any site that got BLISS-10 sources got them under the table.
What BLISS did we use to build SYSERR (I think that was the
name of another monster we shipped).  I think there are two
more products that we had to use bliss10.

/BAH


/BAH

Subtract a hundred and four for e-mail.


Article 2708 of alt.sys.pdp10:
Path: nntp1.ba.best.com!news1.best.com!nntp.primenet.com!nntp.gblx.net!cpk-news-hub1.bbnplanet.com!news.gtei.net!feed2.news.rcn.net!feed1.news.rcn.net!rcn!not-for-mail
From: "Alan H. Martin" <AMartin@MA.UltraNet.Com>
Newsgroups: alt.sys.pdp10
Subject: Re: BLT 17,17 bug had been fixed but..
Date: Tue, 19 Dec 2000 20:44:56 -0500
Lines: 28
Message-ID: <3A400F18.1A7A4CB@MA.UltraNet.Com>
References: <6TiY5.46657$%j3.530564@news6.giganews.com> <90teal$10b$1@bob.news.rcn.net> <J9AY5.115963$DG3.2253745@news2.giganews.com> <9105q9$8d3$1@bob.news.rcn.net> <gCeZ5.121937$DG3.2418389@news2.giganews.com> <9154bi$8am$1@bob.news.rcn.net> <UpwZ5.77193$IP1.2582629@news1.giganews.com> <917jrq$2s6$2@bob.news.rcn.net> <Z7YZ5.59759$%j3.763939@news6.giganews.com> <91a8c1$fj$1@bob.news.rcn.net> <3A3EE9FC.A73D5CAF@MA.UltraNet.Com> <91nqad$43p$2@autumn.news.rcn.net>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
X-Trace: UmFuZG9tSVbImJifbQ+Na36bIjIzsxq9wlivh7EPS8bXsQ1N2Y/TCF38aXqQWOel
X-Complaints-To: abuse@rcn.com
NNTP-Posting-Date: 20 Dec 2000 01:45:37 GMT
X-Mailer:  Mozilla 4.75 [en] (Win95; U)
X-Accept-Language:  en,en-US,en-GB,es
Xref: nntp1.ba.best.com alt.sys.pdp10:2708

jmfbahciv@aol.com wrote:
> 
> In article <3A3EE9FC.A73D5CAF@MA.UltraNet.Com>,
>    "Alan H. Martin" <AMartin@MA.UltraNet.Com> wrote:
...
> >I still don't recall how TOPS-10 sites got BLISS-10 sources, but we
> >certainly had them in the field.
> 
> Any site that got BLISS-10 sources got them under the table.

Sure, if ``TOPS-20 V3A 2020 DISTRIBUTION TAPE 9 TRACK 1600 BPI MAGTAPE
DRIVE# 273 S/N 2 SYS, BB-D868B-BM'' is ``under the table''.


> What BLISS did we use to build SYSERR (I think that was the
> name of another monster we shipped).  ...

The program name is familiar, but I can't find it on any Trailing
Edge tapes.


>...  I think there are two
> more products that we had to use bliss10.

APL would be one of them.
				/AHM
-- 
Alan Howard Martin				AMartin@MA.UltraNet.Com


Article 2724 of alt.sys.pdp10:
Path: nntp1.ba.best.com!news1.best.com!nntp.primenet.com!nntp.gblx.net!europa.netcrusader.net!207.172.3.44!feed2.news.rcn.net!feed1.news.rcn.net!rcn!207-172-216-122
From: jmfbahciv@aol.com
Newsgroups: alt.sys.pdp10
Subject: Re: BLT 17,17 bug had been fixed but..
Date: Wed, 20 Dec 00 11:25:57 GMT
Organization: UltraNet Communications, Inc.
Lines: 42
Message-ID: <91q8uo$jmn$3@bob.news.rcn.net>
References: <9154bi$8am$1@bob.news.rcn.net> <UpwZ5.77193$IP1.2582629@news1.giganews.com> <917jrq$2s6$2@bob.news.rcn.net> <Z7YZ5.59759$%j3.763939@news6.giganews.com> <91a8c1$fj$1@bob.news.rcn.net> <3A3EE9FC.A73D5CAF@MA.UltraNet.Com> <91nqad$43p$2@autumn.news.rcn.net> <3A400F18.1A7A4CB@MA.UltraNet.Com> <91q4r8$9h2$1@bob.news.rcn.net>
X-Trace: UmFuZG9tSVbc0q7FknmMCYbC1pta+t+6QSL+n9VuJ65c+vL0SflMYVGjF+q/tsoC
X-Complaints-To: abuse@rcn.com
NNTP-Posting-Date: 20 Dec 2000 12:33:28 GMT
X-Newsreader:  News Xpress Version 1.0 Beta #4
Xref: nntp1.ba.best.com alt.sys.pdp10:2724

In article <91q4r8$9h2$1@bob.news.rcn.net>, jmfbahciv@aol.com wrote:
>In article <3A400F18.1A7A4CB@MA.UltraNet.Com>,
>   "Alan H. Martin" <AMartin@MA.UltraNet.Com> wrote:
>>jmfbahciv@aol.com wrote:
>>> 
>>> In article <3A3EE9FC.A73D5CAF@MA.UltraNet.Com>,
>>>    "Alan H. Martin" <AMartin@MA.UltraNet.Com> wrote:
>>....
>>> >I still don't recall how TOPS-10 sites got BLISS-10 sources, but we
>>> >certainly had them in the field.
>>> 
>>> Any site that got BLISS-10 sources got them under the table.
>>
>>Sure, if ``TOPS-20 V3A 2020 DISTRIBUTION TAPE 9 TRACK 1600 BPI MAGTAPE
>>DRIVE# 273 S/N 2 SYS, BB-D868B-BM'' is ``under the table''.
>
>Those weren't BLISS-10 sources.  And I see that it's a -20 tape,
>not a -10 tape.

<snip>

I should also point out that, just because there existed a
tape in SDC, does not mean that the tape was available for
customer purchase.  I never reviewed an SPD for BLISS-10.
It was not a product made available for customers.  

Eric did make a source tape of BLISS-10 after cleaning 
up a Big Mess.  I advised him to make that source tape
and have Magee submit it to SDC.  That was our way
of ensuring archiving.  (Unfortunately, SDC apparently
changed their business plan.)  But there wasn't any
way for a customer to order that tape since we did NOT
make an SPD for it.

I wish people had been taught how we did our business.  It
wasn't just writing programs and automagically the customers
got it.

/BAH


Subtract a hundred and four for e-mail.


Article 2526 of alt.sys.pdp10:
Path: nntp1.ba.best.com!news1.best.com!news2.best.com!news.maxwell.syr.edu!news-out.uswest.net!uunet!chi.uu.net!arb.uu.net!ffx.uu.net!spool1.news.uu.net!spool0.news.uu.net!reader3.news.uu.net!not-for-mail
Message-ID: <3A38668C.205FC9B5@trailing-edge.com>
Date: Thu, 14 Dec 2000 06:19:56 -0400
From: Tim Shoppa <shoppa@trailing-edge.com>
Organization: Trailing Edge Technology
X-Mailer: Mozilla 3.03Gold (X11; I; OpenVMS V7.2     AlphaServer 1200 5/533 4MB)
MIME-Version: 1.0
Newsgroups: alt.sys.pdp10
Subject: Re: BLT 17,17 bug had been fixed but..
References: <6TiY5.46657$%j3.530564@news6.giganews.com> <90teal$10b$1@bob.news.rcn.net> <J9AY5.115963$DG3.2253745@news2.giganews.com> <9105q9$8d3$1@bob.news.rcn.net> <gCeZ5.121937$DG3.2418389@news2.giganews.com> <9154bi$8am$1@bob.news.rcn.net> <UpwZ5.77193$IP1.2582629@news1.giganews.com> <917jrq$2s6$2@bob.news.rcn.net> <Z7YZ5.59759$%j3.763939@news6.giganews.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Lines: 23
NNTP-Posting-Host: 63.73.218.130
X-Trace: reader3.news.uu.net 976792796 23752 63.73.218.130
Xref: nntp1.ba.best.com alt.sys.pdp10:2526

Timothy Stark wrote:
> I am looking for BLISS-10 compiler for TOPS-10/20 operating system.
> Does anyone have it?

I think at least a couple versions are in the PDP-10 Archives already,
although I cannot promise that they are complete.  In particular,
at

  http://pdp-10.trailing-edge.com/

on tape BB-D868E-BM, TOPS20 V41 2020 DIST 1OF2, we have

PS:<4-1-DOCUMENTATION>BLIS10.DOC.1                   152 25-JAN-1982
PS:<NEW-SUBSYS>BLIS10.EXE.1                          760 15-APR-1982
PS:<NEW-SUBSYS>BLIS10.HLP.1                           16 12-MAR-1978
PS:<LANGUAGE-SOURCES>BLIS10.ERR.1                     56 15-APR-1982
PS:<LANGUAGE-SOURCES>BLIS20.CMD.1                     16 25-MAR-1982
PS:<LANGUAGE-SOURCES>BLIS20.CTL.1                     24  3-FEB-1983
PS:<LANGUAGE-SOURCES>BLIS20.LNK.1                      8 27-FEB-1978

Does this help any?

Tim.


Article 2543 of alt.sys.pdp10:
Path: nntp1.ba.best.com!news1.best.com!newsfeed.mathworks.com!europa.netcrusader.net!207.172.3.44!feed2.news.rcn.net!feed1.news.rcn.net!rcn!207-172-245-54
From: jmfbahciv@aol.com
Newsgroups: alt.sys.pdp10
Subject: Re: BLT 17,17 bug had been fixed but..
Date: Fri, 15 Dec 00 09:52:38 GMT
Organization: UltraNet Communications, Inc.
Lines: 25
Message-ID: <91ctic$aj0$2@bob.news.rcn.net>
References: <6TiY5.46657$%j3.530564@news6.giganews.com> <90teal$10b$1@bob.news.rcn.net> <J9AY5.115963$DG3.2253745@news2.giganews.com> <9105q9$8d3$1@bob.news.rcn.net> <gCeZ5.121937$DG3.2418389@news2.giganews.com> <9154bi$8am$1@bob.news.rcn.net> <UpwZ5.77193$IP1.2582629@news1.giganews.com> <917jrq$2s6$2@bob.news.rcn.net> <Z7YZ5.59759$%j3.763939@news6.giganews.com> <3A38668C.205FC9B5@trailing-edge.com> <0y4_5.62357$%j3.782275@news6.giganews.com>
X-Trace: UmFuZG9tSVbZ5DhOFnGlxyedvlDUZM+TOP6jQ0S6dPyaSVNmF4AIqSAKwqG5ljkW
X-Complaints-To: abuse@rcn.com
NNTP-Posting-Date: 15 Dec 2000 10:59:24 GMT
X-Newsreader:  News Xpress Version 1.0 Beta #4
Xref: nntp1.ba.best.com alt.sys.pdp10:2543

In article <0y4_5.62357$%j3.782275@news6.giganews.com>,
   Timothy Stark <sword7@grace.speakeasy.org> wrote:
>Tim Shoppa <shoppa@trailing-edge.com> wrote:
>> on tape BB-D868E-BM, TOPS20 V41 2020 DIST 1OF2, we have
>
>> PS:<4-1-DOCUMENTATION>BLIS10.DOC.1                   152 25-JAN-1982
>> PS:<NEW-SUBSYS>BLIS10.EXE.1                          760 15-APR-1982
>> PS:<NEW-SUBSYS>BLIS10.HLP.1                           16 12-MAR-1978
>> PS:<LANGUAGE-SOURCES>BLIS10.ERR.1                     56 15-APR-1982
>> PS:<LANGUAGE-SOURCES>BLIS20.CMD.1                     16 25-MAR-1982
>> PS:<LANGUAGE-SOURCES>BLIS20.CTL.1                     24  3-FEB-1983
>> PS:<LANGUAGE-SOURCES>BLIS20.LNK.1                      8 27-FEB-1978
>
>Well, thank you for info.  I forget that TOPS-20 v4.1 has them.

Careful.  What got shipped on the -20 tapes may not be the
code that built the stuff on the -10.  Double check that
the edit level.  The edit level is very important and 
had a strict standard of assignation within LCG (Large
Computer Group)....at least, it did while some old farts
were still around.

/BAH

Subtract a hundred and four for e-mail.


Article 2544 of alt.sys.pdp10:
Path: nntp1.ba.best.com!news1.best.com!newsfeed.mathworks.com!newsfeed.direct.ca!look.ca!nntp2.aus1.giganews.com!nntp3.aus1.giganews.com!news1.giganews.com.POSTED!barn.net1plus.com!192.168.1.12
From: R.J.S. <not@available.right.now>
Newsgroups: alt.sys.pdp10
Subject: Re: BLT 17,17 bug had been fixed but..
Organization: Very Organized
Message-ID: <hlvj3to326rjblokbugpa7mqogg66sdvfv@4ax.com>
References: <6TiY5.46657$%j3.530564@news6.giganews.com> <90teal$10b$1@bob.news.rcn.net> <J9AY5.115963$DG3.2253745@news2.giganews.com> <9105q9$8d3$1@bob.news.rcn.net> <gCeZ5.121937$DG3.2418389@news2.giganews.com> <9154bi$8am$1@bob.news.rcn.net> <UpwZ5.77193$IP1.2582629@news1.giganews.com> <917jrq$2s6$2@bob.news.rcn.net> <Z7YZ5.59759$%j3.763939@news6.giganews.com> <3A38668C.205FC9B5@trailing-edge.com>
X-Newsreader: Forte Agent 1.8/32.548
X-No-Archive: yes
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
X-Original-NNTP-Posting-Host: 192.168.1.12
X-Original-Trace: 15 Dec 2000 06:27:52 -0500, 192.168.1.12
Lines: 18
NNTP-Posting-Date: Fri, 15 Dec 2000 05:28:07 CST
X-Trace: sv2-pDTtVG2oo4Wt9x3GS6zhHlWHGz8bGZTbQjzU7bjptyUslh2ZCpmKhfVlo7K7Jk2+//YWBGfRZxpC9iT!PevW7uyCxQ5L0agwUOn8mKsS8P0WIQ==
X-Complaints-To: abuse@GigaNews.Com
X-Abuse-Info: Please be sure to forward a copy of ALL headers
X-Abuse-Info: Otherwise we will be unable to process your complaint properly
Date: Fri, 15 Dec 2000 11:28:08 GMT
Xref: nntp1.ba.best.com alt.sys.pdp10:2544

On Thu, 14 Dec 2000 06:19:56 -0400, Tim Shoppa
<shoppa@trailing-edge.com> wrote:

>Timothy Stark wrote:
>> I am looking for BLISS-10 compiler for TOPS-10/20 operating system.
>> Does anyone have it?

For quite some time, it was very common to hear someone wandering the
halls of MR1 say "Bliss is ignorance!"  

I tried to use it for some system software, but the code it generated
was about 3 times the size of well written Macro-10, and it took me
longer to figure out all the compiler bugs and work around them.

The assembly language optimizer in Bliss only worked on a very small
window of instructions, and while it may have been great for the VAX,
it did nothing for RISC style machines.  For the -10, it really
sucked.


Article 2546 of alt.sys.pdp10:
Path: nntp1.ba.best.com!news1.best.com!newsfeed.mathworks.com!nntp.abs.net!cpk-news-hub1.bbnplanet.com!news.gtei.net!feed2.news.rcn.net!feed1.news.rcn.net!rcn!207-172-245-54
From: jmfbahciv@aol.com
Newsgroups: alt.sys.pdp10
Subject: Re: BLT 17,17 bug had been fixed but..
Date: Fri, 15 Dec 00 13:29:42 GMT
Organization: UltraNet Communications, Inc.
Lines: 37
Message-ID: <91da9b$n8j$7@bob.news.rcn.net>
References: <6TiY5.46657$%j3.530564@news6.giganews.com> <90teal$10b$1@bob.news.rcn.net> <J9AY5.115963$DG3.2253745@news2.giganews.com> <9105q9$8d3$1@bob.news.rcn.net> <gCeZ5.121937$DG3.2418389@news2.giganews.com> <9154bi$8am$1@bob.news.rcn.net> <UpwZ5.77193$IP1.2582629@news1.giganews.com> <917jrq$2s6$2@bob.news.rcn.net> <Z7YZ5.59759$%j3.763939@news6.giganews.com> <3A38668C.205FC9B5@trailing-edge.com> <hlvj3to326rjblokbugpa7mqogg66sdvfv@4ax.com>
X-Trace: UmFuZG9tSVZVGJJmSPaxHw2Tw1AJ9yx6TJ7Q+iDiqnQMlCxT7vXWYXSllO9WsZSz
X-Complaints-To: abuse@rcn.com
NNTP-Posting-Date: 15 Dec 2000 14:36:27 GMT
X-Newsreader:  News Xpress Version 1.0 Beta #4
Xref: nntp1.ba.best.com alt.sys.pdp10:2546

In article <hlvj3to326rjblokbugpa7mqogg66sdvfv@4ax.com>,
   R.J.S. <not@available.right.now> wrote:
>On Thu, 14 Dec 2000 06:19:56 -0400, Tim Shoppa
><shoppa@trailing-edge.com> wrote:
>
>>Timothy Stark wrote:
>>> I am looking for BLISS-10 compiler for TOPS-10/20 operating system.
>>> Does anyone have it?
>
>For quite some time, it was very common to hear someone wandering the
>halls of MR1 say "Bliss is ignorance!"  
>
>I tried to use it for some system software, but the code it generated
>was about 3 times the size of well written Macro-10, and it took me
>longer to figure out all the compiler bugs and work around them.
>
>The assembly language optimizer in Bliss only worked on a very small
>window of instructions, and while it may have been great for the VAX,
>it did nothing for RISC style machines.  For the -10, it really
>sucked.

And that's why I blew up like I did.  It appears that we designed
our cold start procedure into a corner with no way out.  Tim's
problem is an example of depending on a HLL for all system 
software; it's a pure CATCH-22 problem with the only solution
has the assumption that a functional system exists before
a coldstart installation.

The current design has painted itself to ultimately rely
on BLISS-10 code generation working reproducibly well.

BLISS-10 could generally reproduce the same machine code
(I am assuming no changes).  BLISS-36 could not.

/BAH

Subtract a hundred and four for e-mail.


Article 2672 of alt.sys.pdp10:
Path: nntp1.ba.best.com!news1.best.com!newsfeed.mathworks.com!feeder.qis.net!feed2.news.rcn.net!feed1.news.rcn.net!rcn!not-for-mail
From: "Alan H. Martin" <AMartin@MA.UltraNet.Com>
Newsgroups: alt.sys.pdp10
Subject: Re: BLT 17,17 bug had been fixed but..
Date: Tue, 19 Dec 2000 01:39:48 -0500
Lines: 29
Message-ID: <3A3F02B4.7F338F2D@MA.UltraNet.Com>
References: <6TiY5.46657$%j3.530564@news6.giganews.com> <90teal$10b$1@bob.news.rcn.net> <J9AY5.115963$DG3.2253745@news2.giganews.com> <9105q9$8d3$1@bob.news.rcn.net> <gCeZ5.121937$DG3.2418389@news2.giganews.com> <9154bi$8am$1@bob.news.rcn.net> <UpwZ5.77193$IP1.2582629@news1.giganews.com> <917jrq$2s6$2@bob.news.rcn.net> <Z7YZ5.59759$%j3.763939@news6.giganews.com> <3A38668C.205FC9B5@trailing-edge.com> <hlvj3to326rjblokbugpa7mqogg66sdvfv@4ax.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
X-Trace: UmFuZG9tSVYE8gDlhi1wN2TH6N+0RRMP6F1ZSFO86wacFmxUjK9/X4A7ND2pjI4U
X-Complaints-To: abuse@rcn.com
NNTP-Posting-Date: 19 Dec 2000 06:40:27 GMT
X-Mailer:  Mozilla 4.75 [en] (Win95; U)
X-Accept-Language:  en,en-US,en-GB,es
Xref: nntp1.ba.best.com alt.sys.pdp10:2672

"R.J.S." wrote:
> 
> >Timothy Stark wrote:
> >> I am looking for BLISS-10 compiler for TOPS-10/20 operating system.
...
> I tried to use it for some system software, but the code it generated
> was about 3 times the size of well written Macro-10, and it took me
> longer to figure out all the compiler bugs and work around them.
> 
> The assembly language optimizer in Bliss only worked on a very small
> window of instructions, ...

Maybe 10 instructions.  (Fortran-10 buffered 50).


>... and while it may have been great for the VAX,
> it did nothing for RISC style machines.  For the -10, it really
> sucked.

BLISS-11, BLISS-32, BLISS-16, BLISS-64 *and* BLISS-36 have global
optimizers.  I've seen BLISS-36 notice that a complex boolean was still
sitting in a register pages of code away from where it had been computed.

BLISS-10 is far simpler - perhaps has little more than that peephole
optimizer.  BLISS-10 compares to its successors like the PDP-11 C
compiler compares to gcc.
				/AHM
-- 
Alan Howard Martin				AMartin@MA.UltraNet.Com


Article 2687 of alt.sys.pdp10:
Path: nntp1.ba.best.com!news1.best.com!newsfeed.mathworks.com!europa.netcrusader.net!207.172.3.44!feed2.news.rcn.net!feed1.news.rcn.net!rcn!207-172-245-95
From: jmfbahciv@aol.com
Newsgroups: alt.sys.pdp10
Subject: Re: BLT 17,17 bug had been fixed but..
Date: Tue, 19 Dec 00 13:13:49 GMT
Organization: UltraNet Communications, Inc.
Lines: 49
Message-ID: <91nqsm$43p$4@autumn.news.rcn.net>
References: <6TiY5.46657$%j3.530564@news6.giganews.com> <90teal$10b$1@bob.news.rcn.net> <J9AY5.115963$DG3.2253745@news2.giganews.com> <9105q9$8d3$1@bob.news.rcn.net> <gCeZ5.121937$DG3.2418389@news2.giganews.com> <9154bi$8am$1@bob.news.rcn.net> <UpwZ5.77193$IP1.2582629@news1.giganews.com> <917jrq$2s6$2@bob.news.rcn.net> <Z7YZ5.59759$%j3.763939@news6.giganews.com> <3A38668C.205FC9B5@trailing-edge.com> <hlvj3to326rjblokbugpa7mqogg66sdvfv@4ax.com> <3A3F02B4.7F338F2D@MA.UltraNet.Com>
X-Trace: UmFuZG9tSVZ5uzNKsQYDGfIlN3P4XSBtpFkCQXIKsJbBDaTYDgC1Vce53ON9HjIj
X-Complaints-To: abuse@rcn.com
NNTP-Posting-Date: 19 Dec 2000 14:21:10 GMT
X-Newsreader:  News Xpress Version 1.0 Beta #4
Xref: nntp1.ba.best.com alt.sys.pdp10:2687

In article <3A3F02B4.7F338F2D@MA.UltraNet.Com>,
   "Alan H. Martin" <AMartin@MA.UltraNet.Com> wrote:
>"R.J.S." wrote:
>> 
>> >Timothy Stark wrote:
>> >> I am looking for BLISS-10 compiler for TOPS-10/20 operating system.
>....
>> I tried to use it for some system software, but the code it generated
>> was about 3 times the size of well written Macro-10, and it took me
>> longer to figure out all the compiler bugs and work around them.
>> 
>> The assembly language optimizer in Bliss only worked on a very small
>> window of instructions, ...
>
>Maybe 10 instructions.  (Fortran-10 buffered 50).
>
>
>>... and while it may have been great for the VAX,
>> it did nothing for RISC style machines.  For the -10, it really
>> sucked.
>
>BLISS-11, BLISS-32, BLISS-16, BLISS-64 *and* BLISS-36 have global
>optimizers.  I've seen BLISS-36 notice that a complex boolean was still
>sitting in a register pages of code away from where it had been computed.
>
>BLISS-10 is far simpler - perhaps has little more than that peephole
>optimizer.  BLISS-10 compares to its successors like the PDP-11 C
>compiler compares to gcc.

It's worse than that.  The BLISS compilers before DECnet
are different than the BLISS compilers after DECnet.
That BLISS-11 that you guys are talking about is a before-DECnet
compiler. 

Comparing before-DECnet BLISSes with after-DECnet-BLISSes
is more like comparing FORTRAN II with FORTRAN IV.  

And it even gets more complicated than that because BLISS
development continued a CMU _without us getting any of the
updates_.  So BLISS-10 and BLISS-11 compiler development
is really some flavor of compiler that had been frozen
in the early 70s (I'm guessing) or so.

The person we need for all of this BLISS-10 and RMS stuff
is Eric Macgrath.  Somebody go kick him.

/BAH

Subtract a hundred and four for e-mail.


Article: 15345 of alt.sys.pdp10
Path: iad-read.news.verio.net!dfw-artgen.news.verio.net!dfw-peer.news.verio.net!news.verio.net!crtntx1-snh1.gtei.net!nycmny1-snh1.gtei.net!news.gtei.net!news-out.visi.com!hermes.visi.com!news.maxwell.syr.edu!feeder.qis.net!sn-xit-02!supernews.com!postnews1.google.com!not-for-mail
From: jeremybarker@email.com (Jeremy Barker)
Newsgroups: comp.os.vms,alt.sys.pdp10
Subject: Re: BLISS pros and cons, was: Re: historical evidence of what went wrong  at DEC
Date: 22 Jan 2002 07:47:42 -0800
Organization: http://groups.google.com/
Lines: 31
Message-ID: <5b86b9ee.0201220747.3f527e97@posting.google.com>
References: <a0l3hm$1ih4$1@citadel.in.taronga.com> <a11g1s$1s3$2@bob.news.rcn.net> <0H2Z7.2548$cD4.5034@www.newsranger.com> <V53Z7.267$5Y4.5736@news.cpqcorp.net> <Pine.NXT.4.50.0201031728530.20229-100000@Tomobiki-Cho.CAC.Washington.EDU> <11jZ7.280$5Y4.6342@news.cpqcorp.net> <3C35C22E.F44C6D8E@aaa.com> <3C35C48C.30803@compaq.com> <qhitahj434.fsf@ruckus.brouhaha.com>
NNTP-Posting-Host: 213.122.15.133
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 8bit
X-Trace: posting.google.com 1011714463 13867 127.0.0.1 (22 Jan 2002 15:47:43 GMT)
X-Complaints-To: groups-abuse@google.com
NNTP-Posting-Date: 22 Jan 2002 15:47:43 GMT
Xref: dfw-artgen.news.verio.net comp.os.vms:124448 alt.sys.pdp10:15345

Eric Smith <eric-no-spam-for-me@brouhaha.com> wrote in message news:<qhitahj434.fsf@ruckus.brouhaha.com>...
> John Reagan <john.reagan@compaq.com> writes:
> > While Compaq never sold BLISS for NT, we certainly continue to use it
> > for the Fortran products (both on Alpha NT and IA32 NT) since the GEM
> > component of the compiler is written mostly in BLISS.
> 
> Would it be that hard to do an automatic translation of BLISS to C?
> I'm not a BLISS expert, but it seems to me that the semantics are
> simple enough.  I'd expect that the hardest part would be trying to
> identify which variables are used as pointers vs. numeric types.  For
> variables that are used as both, the generated C code would need to
> include casts.
> 
> As far as I know, standard C does not specify any pair of integer and
> pointer types such that casting back and forth is guaranteed to work.
> But in practice on any given implementation there usually are suitable
> types.  So a BLISS to C translator should allow the specification of
> the C types to be used.

It's probably easier to translate C to BLISS than the reverse. 
Especially given that nothing in C can match the BLISS macro facility.

It's certainly the case that various people have looked at the
possibility of a BLISS to C translator but the killer is the macro
facility built into the BLISS compiler front-end.  I also know that
someone looked at the possibility of making the BLISS compiler
generate C source code rather than object files (the aim was to widen
the portability of programs written in BLISS) but this idea was
abandoned for a variety of reasons.

jb


Article: 15581 of alt.sys.pdp10
Path: iad-read.news.verio.net!dfw-artgen.news.verio.net!dfw-peer.news.verio.net!news.verio.net!crtntx1-snh1.gtei.net!cpk-news-hub1.bbnplanet.com!news.gtei.net!news.tele.dk!small.news.tele.dk!212.74.64.35!colt.net!newspeer.clara.net!news.clara.net!peernews!peer.cwci.net!news5-gui.server.ntli.net!ntli.net!news6-win.server.ntlworld.com.POSTED!not-for-mail
Message-ID: <3C558E84.83E2B32A@iee.org>
From: "a.carlini" <arcarlini@iee.org>
Reply-To: arcarlini@iee.org
X-Mailer: Mozilla 4.7 [en] (Win98; I)
X-Accept-Language: en
MIME-Version: 1.0
Newsgroups: comp.os.vms,alt.sys.pdp10
Subject: Re: BLISS pros and cons, was: Re: historical evidence of what went 
 wrong  at DEC
References: <a0l3hm$1ih4$1@citadel.in.taronga.com> <3C35C48C.30803@compaq.com> <qhitahj434.fsf@ruckus.brouhaha.com> <5b86b9ee.0201220747.3f527e97@posting.google.com> <a329d2$1vaj$1@citadel.in.taronga.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Lines: 20
Date: Mon, 28 Jan 2002 17:46:44 +0000
NNTP-Posting-Host: 62.254.135.163
X-Complaints-To: abuse@ntlworld.com
X-Trace: news6-win.server.ntlworld.com 1012240154 62.254.135.163 (Mon, 28 Jan 2002 17:49:14 GMT)
NNTP-Posting-Date: Mon, 28 Jan 2002 17:49:14 GMT
Organization: ntl Cablemodem News Service
Xref: dfw-artgen.news.verio.net comp.os.vms:125311 alt.sys.pdp10:15581

Peter da Silva wrote:
> 
> In article <5b86b9ee.0201220747.3f527e97@posting.google.com>,
> Jeremy Barker <jeremybarker@email.com> wrote:
> >It's probably easier to translate C to BLISS than the reverse.
> >Especially given that nothing in C can match the BLISS macro facility.
> 
> That's what m4 was invented for. It works just fine as a C preprocessor, too.

And m4 was used as one for the DECnis 
FDDI line card too IIRC.

So not *everything* was done in BLISS :-)

Antonio

-- 

---------------
Antonio Carlini             arcarlini@iee.org


Article: 15348 of alt.sys.pdp10
Path: iad-read.news.verio.net!dfw-artgen.news.verio.net!dfw-peer.news.verio.net!news.verio.net!crtntx1-snh1.gtei.net!nycmny1-snh1.gtei.net!cpk-news-hub1.bbnplanet.com!news.gtei.net!news.maxwell.syr.edu!feeder.qis.net!sn-xit-02!supernews.com!postnews1.google.com!not-for-mail
From: jeremybarker@email.com (Jeremy Barker)
Newsgroups: comp.os.vms,alt.sys.pdp10
Subject: Re: BLISS pros and cons, was: Re: historical evidence of what went wrong at DEC
Date: 22 Jan 2002 08:13:39 -0800
Organization: http://groups.google.com/
Lines: 51
Message-ID: <5b86b9ee.0201220813.4e642eea@posting.google.com>
References: <a0l3hm$1ih4$1@citadel.in.taronga.com> <Pine.NXT.4.50.0201031728530.20229-100000@Tomobiki-Cho.CAC.Washington.EDU> <11jZ7.280$5Y4.6342@news.cpqcorp.net> <Pine.NXT.4.50.0201040717520.21933-100000@Tomobiki-Cho.CAC.Washington.EDU> <a14pr1$1vou$1@citadel.in.taronga.com> <Pine.NXT.4.50.0201041017210.22025-100000@Tomobiki-Cho.CAC.Washington.EDU>
NNTP-Posting-Host: 213.122.162.3
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 8bit
X-Trace: posting.google.com 1011716020 14678 127.0.0.1 (22 Jan 2002 16:13:40 GMT)
X-Complaints-To: groups-abuse@google.com
NNTP-Posting-Date: 22 Jan 2002 16:13:40 GMT
Xref: dfw-artgen.news.verio.net comp.os.vms:124454 alt.sys.pdp10:15348

Mark Crispin <mrc@CAC.Washington.EDU> wrote in message news:<Pine.NXT.4.50.0201041017210.22025-100000@Tomobiki-Cho.CAC.Washington.EDU>...
> On 4 Jan 2002, Peter da Silva wrote:
> > That is, at the time VMS development was started "C" was not even on the
> > radar. Even if it might have made more sense to use C than BLISS (and
> > I can't argue otherwise, really) I don't think there's any reason to expect
> > the people making the decision to have even *heard* of C.
> 
> I first heard about VAX and VMS in the mid-70s.  By that time, they should
> have known what Bell Labs was doing with all those PDP-11s, even if at
> that time UNIX was mostly internal.
> 
> BCPL was widely available by that time.

VAX and VMS were conceived around 1975.  I think it's stretching it a
bit to say that BCPL was "widely" available then.  I should say that I
learned how to write properly structured programs using BCPL.
 
> > A stripped-down Algol or PL/1 might have been better solutions.
> 
> Agreed.  Univac used Algol, and PL/1 was the system language on Multics.
> Both of these were already old hat by that time.

Maybe so.  As I understand the situation, BLISS was favoured by DEC
because of the DEC-CMU connections and because of the research that
had been done at CMU on producing highly optimised code (the TOPS-10
hosted BLISS-16 compiler).  My view (which others will likely disagree
with) is that it was only around the mid-1970s timeframe that DEC
became seriously interested in compilers.  It had previously done
pretty much everything in assembly language of one sort or another and
compilers were just something that customers needed so had to be
provided.  Some compilers were bought in (BASIC-PLUS on RSTS for
example).  You have to remember that DEC was basically a hardware
company that used software (which was often given away) to leverage
hardware sales.  That model only really changed after the PDP-11
appeared and VAX was the first hardware designed along with software.
 
> > Actually, I've found that it's easier to write portable code in Fortran
> > than in most languages, though that may partly have been due to the kinds
> > of things you actually wrote in Fortran.
> 
> Definitely true.  DEC adhered to Fortran II style I/O for a long time, but
> for the most part Fortran IV is Fortran IV no matter where you go.

I remember that the computing service at Cambridge University produced
a guide to FORTRAN portability which showed what language features
were portable and to what degree.  Although the core of the language
was highly portable there were lots of features that IBM (especially
in the S/370 H compiler) and others added that were distinctly
non-portable.
 
jb


Article: 15347 of alt.sys.pdp10
Path: iad-read.news.verio.net!dfw-artgen.news.verio.net!dfw-peer.news.verio.net!iad-peer.news.verio.net!news.verio.net!news.maxwell.syr.edu!newsfeed1.uni2.dk!news.get2net.dk!not-for-mail
From: "Dr. Dweeb" <Dweeb@NoSpam.com>
Newsgroups: comp.os.vms,alt.sys.pdp10
References: <a0l3hm$1ih4$1@citadel.in.taronga.com> <a11g1s$1s3$2@bob.news.rcn.net> <0H2Z7.2548$cD4.5034@www.newsranger.com> <V53Z7.267$5Y4.5736@news.cpqcorp.net> <5b86b9ee.0201220733.97fe903@posting.google.com>
Subject: Re: BLISS pros and cons, was: Re: historical evidence of what went wrong at DEC
Lines: 7
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2600.0000
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000
Message-ID: <UZf38.111$EU.3174@news.get2net.dk>
Date: Tue, 22 Jan 2002 16:57:57 +0100
NNTP-Posting-Host: 129.142.252.2
X-Complaints-To: abuse@uni2.dk
X-Trace: news.get2net.dk 1011715060 129.142.252.2 (Tue, 22 Jan 2002 16:57:40 MET)
NNTP-Posting-Date: Tue, 22 Jan 2002 16:57:40 MET
Organization: UNI2 Internet Kunde
Xref: dfw-artgen.news.verio.net comp.os.vms:124451 alt.sys.pdp10:15347

CompaQ killed the BLISS/NT intel and Alpha compilers, but they existed for
quite some time.  OracleRdb/NT used them extensively.

Copies may exist, or maybe CompaQ got soft and put them in the public
domain.




Article: 15349 of alt.sys.pdp10
Path: iad-read.news.verio.net!dfw-artgen.news.verio.net!dfw-peer.news.verio.net!news.verio.net!crtntx1-snh1.gtei.net!washdc3-snh1.gtei.net!nycmny1-snh1.gtei.net!cpk-news-hub1.bbnplanet.com!news.gtei.net!nntp.abs.net!feeder.qis.net!sn-xit-02!supernews.com!postnews1.google.com!not-for-mail
From: jeremybarker@email.com (Jeremy Barker)
Newsgroups: comp.os.vms,alt.sys.pdp10
Subject: Re: BLISS pros and cons, was: Re: historical evidence of what went wrong at DEC
Date: 22 Jan 2002 08:26:26 -0800
Organization: http://groups.google.com/
Lines: 24
Message-ID: <5b86b9ee.0201220826.19a064@posting.google.com>
References: <a0l3hm$1ih4$1@citadel.in.taronga.com> <a11g1s$1s3$2@bob.news.rcn.net> <0H2Z7.2548$cD4.5034@www.newsranger.com> <V53Z7.267$5Y4.5736@news.cpqcorp.net> <3c3e24d1.37592805@news.process.com>
NNTP-Posting-Host: 213.122.37.220
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 8bit
X-Trace: posting.google.com 1011716787 15056 127.0.0.1 (22 Jan 2002 16:26:27 GMT)
X-Complaints-To: groups-abuse@google.com
NNTP-Posting-Date: 22 Jan 2002 16:26:27 GMT
Xref: dfw-artgen.news.verio.net comp.os.vms:124459 alt.sys.pdp10:15349

goathunter@goatley.com (Hunter Goatley) wrote in message news:<3c3e24d1.37592805@news.process.com>...
> On Thu, 3 Jan 2002 15:20:09 -0500, "Fred Kleinsorge"
> <kleinsorge@star.zko.dec.com> wrote:
> >
> >Just religion.  Larry will tell you that Ada is the best language.  UNIX
> >hacks will tell you C.  The kiddies will tell you JAVA.  I'm sure someone
> >still loves RPG-II.
> >
> But we all know that objectively, BLISS is best.
> 
> >The early VMS developers jumped on it as a way to write much of the kernel
> >code for VMS - as opposed to writing it in Macro-32.  The language itself
> >never caught on.
> 
> It never helped that DEC priced the compiler so high---second only to Ada in
> the '80s.  If they'd given the compiler away, or at least made it as affordable
> as C, PLI, and others, it might have had a chance, IMO.

I always thought that the exorbitant pricing of the BLISS compiler was
a mistake.  I think there was a lot of pressure internally to not sell
it (it being seen as a competitive advantage) so the high price was
probably intended as a deterrant.

jb


Article: 15459 of alt.sys.pdp10
Newsgroups: comp.os.vms,alt.sys.pdp10
Path: iad-read.news.verio.net!dfw-artgen.news.verio.net!dfw-peer.news.verio.net!news.verio.net!crtntx1-snh1.gtei.net!nycmny1-snh1.gtei.net!news.gtei.net!news-out.visi.com!hermes.visi.com!uunet!ash.uu.net!world!not-for-mail
From: weiner@TheWorld.com (Sam Weiner)
Subject: Re: BLISS pros and cons, was: Re: historical evidence of what went wrong at DEC
Sender: news@world.std.com (Mr Usenet Himself)
Message-ID: <GqH7EF.6xn@world.std.com>
Date: Fri, 25 Jan 2002 04:06:14 GMT
References: <a0l3hm$1ih4$1@citadel.in.taronga.com> <V53Z7.267$5Y4.5736@news.cpqcorp.net> <3c3e24d1.37592805@news.process.com> <5b86b9ee.0201220826.19a064@posting.google.com>
Nntp-Posting-Host: shell01.theworld.com
Organization: The World Public Access UNIX, Brookline, MA
X-Newsreader: trn 4.0-test72 (19 April 1999)
Lines: 40
Xref: dfw-artgen.news.verio.net comp.os.vms:124933 alt.sys.pdp10:15459

In article <5b86b9ee.0201220826.19a064@posting.google.com>,
Jeremy Barker <jeremybarker@email.com> wrote:
>goathunter@goatley.com (Hunter Goatley) wrote in message
>news:<3c3e24d1.37592805@news.process.com>...
>> On Thu, 3 Jan 2002 15:20:09 -0500, "Fred Kleinsorge"
>> <kleinsorge@star.zko.dec.com> wrote:
>> >
>> >Just religion.  Larry will tell you that Ada is the best language.  UNIX
>> >hacks will tell you C.  The kiddies will tell you JAVA.  I'm sure someone
>> >still loves RPG-II.
>> >
>> But we all know that objectively, BLISS is best.
>> 
>> >The early VMS developers jumped on it as a way to write much of the kernel
>> >code for VMS - as opposed to writing it in Macro-32.  The language itself
>> >never caught on.
>> 
>> It never helped that DEC priced the compiler so high---second only to Ada in
>> the '80s.  If they'd given the compiler away, or at least made it as
>affordable
>> as C, PLI, and others, it might have had a chance, IMO.
>
>I always thought that the exorbitant pricing of the BLISS compiler was
>a mistake.  I think there was a lot of pressure internally to not sell
>it (it being seen as a competitive advantage) so the high price was
>probably intended as a deterrant.

We started with BLIS10 straight from CMU in mid 70s but
migrated to BLISS-36 once it was finally available.  I
still miss the macro facility.  And unlike C, it doesn't
pretend to be all things to all people.

DEC had the same policy with other things, such as making
the RMS API available outside FORTRAN and COBOL.  It would
have saved us time, money, and much pain.

Sam





Article: 21444 of alt.sys.pdp10
Path: iad-read.news.verio.net!dfw-artgen!sjc-peer.news.verio.net!news.verio.net!pln-e!extra.newsguy.com!lotsanews.com!nntp.abs.net!feed2.news.rcn.net!feed1.news.rcn.net!rcn!208-59-182-144
From: jmfbahciv@aol.com
Newsgroups: alt.sys.pdp10
Subject: Re: BCPL
Date: Mon, 22 Sep 03 13:30:07 GMT
Organization: UltraNet Communications, Inc.
Lines: 49
Message-ID: <bkn1m8$d3r$3@bob.news.rcn.net>
References: <3f60bac1$0$12650$afc38c87@news.easynet.co.uk> <bk4keg$p0u$1@news.louisville.edu> <bked3l$d1@library2.airnews.net> <bkes57$fom$4@bob.news.rcn.net> <bkmt09$v3b@library1.airnews.net>
X-Trace: UmFuZG9tSVagPv0eMOkbuz/ft0aLZv/YsP+N2k5ZTV3rDO+k7nlNGyP7Q/55HFom
X-Complaints-To: abuse@rcn.com
NNTP-Posting-Date: 22 Sep 2003 14:41:12 GMT
X-Newsreader:  News Xpress Version 1.0 Beta #4
Xref: dfw-artgen alt.sys.pdp10:21444

In article <bkmt09$v3b@library1.airnews.net>,
   huw.davies@kerberos.davies.net.au wrote:
>jmfbahciv@aol.com wrote:
>> In article <bked3l$d1@library2.airnews.net>,
>>    huw.davies@kerberos.davies.net.au wrote:
>
>> 'ey, Huw.
>
>Hi
>
>> You can do all the BLISS you want.  Just don't ask me to
>> package it.  If I have to package it, I suggest you come
>> prepared:  suit of armor, asbestos shorts, blarney, and
>> a control file that will build the product _without_
>> warnings or informational messages are requirements.
>
>Indeed. Using BLISS was sometimes a matter of faith.

<mutter..wink>  I would not have used the word faith.

> .. As I've
>said before, I was using BLISS-10 not BLISS-36

BLISS-36 is a whole 'nuther story about how not to business.

> .. and still have the
>source code to the compiler handy when things became "interesting"

One of the problems with BLISS-10 is that we (LCG) never
managed it well.  It was always in that limbo land of sorta
supported and not sorta supported and nobody in charge of that
language was willing to treat like a real product.  I don't know
how much having CMU fingers in the pie fucked it up.  If I ever
heard that CMU was "involved" in something, I nixed it or I gave
up on it.  I've always traced these problems directly back to 
Bell's <ahem>thinking.

I had some problems with the lanagage itself but they were minor.
But it was my first lesson in what happens when HLLs are used
as a language to deliver computing service.  The number of files 
one had to keep hanging around to do anything was horrible.
This back when disks weren't cheap and a max capacity of 80,000
blocks.

/BAH



Subtract a hundred and four for e-mail.