Xref: utzoo comp.unix.sysv386:8872 comp.os.cpm:5310
Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!swrinde!cs.utexas.edu!execu!sequoia!balkan!wrangler!ssbn!bill
From: bill@ssbn.WLK.COM (Bill Kennedy)
Newsgroups: comp.unix.sysv386,comp.os.cpm
Subject: Re: PL/I compiler
Message-ID: <2080@ssbn.WLK.COM>
Date: 11 Jun 91 00:30:58 GMT
References: <1991Jun4.073924.2610@weyrich.UUCP> <60LR3ND@math.fu-berlin.de> <1991Jun6.144734.244@batcomputer.tn.cornell.edu> <1991Jun10.124814.12965@ticipa.pac.sc.ti.com>
Reply-To: bill@ssbn.WLK.COM (Bill Kennedy)
Followup-To: comp.unix.sysv386
Distribution: usa
Organization: W.L. Kennedy Jr. and Associates, Pipe Creek, TX
Lines: 41

I have been reading this thread with considerable interest because I have
used each of the Digital Research PL/I implementations with varying degrees
of success.  DRI has a couple of grand of my shekels, bought 'em all, and I
share some of Jim's opinions but not all of them.

stinson@tifsim.pac.sc.ti.com (Jim Stinson) writes:
>
>Digital Research's PL/I compiler was very screwed up. After buying the compiler
>for $600 - $500, and sending in the registration form, they set a list of about

This could very well be the case with the DOS version, it was a hastily
converted derivative of the CP/M-86 version which was a hastily translated
derivative of the CP/M-80 version.  I found PL/I-80 to be a very reliable
and robust compiler that was able to do a lot of nifty things on an 8 bit
machine.  Even as we phosphor I have an application running on an ancient
IMSAI in PL/I-80, I installed it in October 1984 and there has never been
so much as a whimper out of it.  It's heavily overlaid but when you've only
got 64K for the OS and everyone on earth, it's to PL/I-80's credit that it
works as well as it does and overlays as well as it does.

>50 or more bugs with it. Things like A = A * B where A and B were floating
>point numbers messed up the stack and you program would eventually dump. 
>X = 105331 where X is an integer, would assign 15331, that is, the comiler
>would drop 0's from integer constants from time to time. There were others,
>I don't remember them all, I just could not believe some one would sell a
>compiler for so much knowing it did not work and wouldn't give refunds. 
>
>    Jim Stinson  stinson@tifsim.csc.ti.com, @tifsim.pac.sc.ti.com

I share Jim's feelings about the later "releases".  DRI just grabbed a
can of paint and tried to paint the horse and call it a zebra.  What was
especially irritating was (after paying them for "Professional Programmer
Support") to report a lethal bug and getting a letter back saying that
the problem was indeed reproducible but the product was "mature and no
longer in need of support".  The one that comes to mind in the DOS world
was when you went to the second level overlay having PL/I get it's FCB's
crossed and display your Access Manager index files as screens and use
you Display Manager files as AM indicies...
-- 
Bill Kennedy  internet  bill@ssbn.WLK.COM or ssbn!bill@attmail.COM
              uucp      {att,cs.utexas.edu,pyramid!daver}!ssbn.wlk.com!bill