Newsgroups: comp.lang.lisp
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!news.mathworks.com!europa.eng.gtefsd.com!howland.reston.ans.net!pipex!uknet!festival!edcogsci!jeff
From: jeff@aiai.ed.ac.uk (Jeff Dalton)
Subject: Re: Why do people like C? (Was: Comparison: Beta - Lisp)
Message-ID: <CywvC8.528@cogsci.ed.ac.uk>
Sender: usenet@cogsci.ed.ac.uk (C News Software)
Nntp-Posting-Host: bute-alter.aiai.ed.ac.uk
Organization: AIAI, University of Edinburgh, Scotland
References: <BUFF.94Oct28135705@pravda.world> <CynKz5.13s@cogsci.ed.ac.uk> <Pine.A32.3.91.941102150406.37864H-100000@swim5.eng.sematech.org>
Date: Mon, 7 Nov 1994 18:57:44 GMT
Lines: 111

In article <Pine.A32.3.91.941102150406.37864H-100000@swim5.eng.sematech.org> "William D. Gooch" <goochb@swim5.eng.sematech.org> writes:
>On Wed, 2 Nov 1994, Jeff Dalton wrote:
>
>> .... Our Symbolics once sat idle
>> for a couple of weeks becuase of an error it could recover
>> from if you just pressed the "proceed" key!  None of our
>> regular Symbolics users knew what to do about it!
>
>It's very hard to believe that any "regular Symbolics user" wouldn't know 
>what to do with a proceedable error.  Even a non-user could simply look 
>at the options provided by the debugger and decide to press <Resume> or 
><Abort>.  Unlike the Eunuchs I've seen, the Symbolics keyboard tells you 
>directly what keystroke corresponds to these actions.

It wasn't a question of figuring out which key to press.  An error
they'd never seen before occurred (while it was rebooting, if I 
recall correctly) and they didn't know what they should do.  Was
it ok to resume or not?

My point is that the way a Symbolics seemed to an experienced,
knowledgable user could be very different from how it appeared
to someone who didn't know so much about the machines and
their software.  Moreover, someone could still be in the latter
category even though they used the machines a fair amount.

Here's another story.  The windows on the Symbolics didn't scroll.
Instead, they just wrapped around when output reached the botton
and started writing over things from the top.  Some people may
prefer that behavior.  The point is that no one knew how to change
it, or even if it could be changed.

I'm not saying anyone who loved the Symbolics was wrong.  They could
be great machines, for some people.  I think I could have learned to
use the Symbolics fairly easily, if I'd had more time to use it.
Knowing Lisp well seemed to be a big advantage.  But for various
reasons many people were put off.

It also matters whether there are other people around who
know the ins and outs of the machines.  Xerox D-machines were
used here fairly heavily (in a slightly different part of the
University), but there were some people who knew a lot about
them and enough machines so that a larger user community could
develop and learn from one another.

>> .... 
>> For instance, my perception of Unix machines seems to be very
>> different from yours.  I don't think I've ever found myself back at
>> the Unix prompt wondering where everything went and cursing 'cause I
>> had to reload.
>
>Come now.  You mean to say you've *never* had an unexpected core dump? 

No, that's not what I meant.  I explained my remark, like here:

>> I have occasionally -- but vary rarely -- found myself back at the
>> Unix prompt instead of in the Lisp debugger.  But I don't curse
>> because I have to reload.  Indeed, I prefer to start over in a
>> clean image fairly often.

So yes I have, very rarely, found myself back at the Unix prompt
instead of in the Lisp debugger.  I suppose that's a random core
dump.  But it's not a serious problem for me, since it hardly ever
happens, much less happens at a point where I'm left wondering 
where everything went and cursing because I have to reload.

Note too that C programmers are usually much worse off.  They
have to compile their programs a special way ("-g") and remember
to run them under a debugger.  It's also harder to look at data
since it doesn't contain type information.  Now, many people
are obviously able to live with that; so it shouldn't seem odd
that someone can live with a Unix Lisp, or even think it's
reasonably good.

>Reloading isn't the issue, debugging is.  Core dumps are very hard to 
>debug from, and they don't give you much flexibility wrt finding out what 
>got you into the problem.  They also offer no means of recovering from it 
>so you can save the work you were doing before the crash happened.

Why would you ever have to debug from a core dump?  I haven't
done that since I was programming in assembler ~15 years ago.
If Lisp machine users still think in such terms, I call that
a point against those machines.

>"Start[ing] over in a clean image" is another completely distinct issue.  

Then I don't know what you had in mind when complaining about

  staring at the unix prompt wondering where everything went and why,
  and cursing 'cause you've got to reload everything to get back to
  where you were -

One way or another, that's simply not a significant problem for me.

>> I take your point that they're reliable, but I'm not sure bad
>> management was the only problem.  
>
>It wasn't, but it was the main one IMO.  Bad management and abysmal 
>marketing (when there was any) top my list.

Were people wrong to regard them as expensive machines that could
handle only one user at a time and had a limited range of software?

Here we had some VAXes as time-sharing machines (and used Franz
Lisp on them).  Unix workstations were a natural progression.
They were often run with too heavy a load, too little memory,
and with some people using them from terminals.  This situation
gradually got better.  But I'm pretty sure it cost much less than
buying enough symbolics machines.

-- jeff

