Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: $Revision: 1.6.2.16 $; site pbear.UUCP
Path: utzoo!linus!philabs!cmcl2!seismo!harvard!think!pbear!peterb
From: peterb@pbear.UUCP
Newsgroups: net.works
Subject: Re: Re: Not again!?  Assembler vs High-L
Message-ID: <3500001@pbear.UUCP>
Date: Wed, 24-Apr-85 14:24:00 EDT
Article-I.D.: pbear.3500001
Posted: Wed Apr 24 14:24:00 1985
Date-Received: Thu, 2-May-85 06:06:25 EDT
References: <5461@utzoo.UUCP>
Lines: 35
Nf-ID: #R:utzoo:-546100:pbear:3500001:000:1449
Nf-From: pbear!peterb    Apr 24 15:24:00 1985



Laura,

don't knock the pdp-8 so much. I HAVE written a quicksort in pdp-8
assembler, and it was simply ordering 12 bit keys field on a 48 bit record.
I squashed that thing down to one page of code. (The stacks took up an
entire field.)

The 8 was (and is) a good machine. It introduced some nifty concepts and had
some real problems as well. But to think that multiuser OS's ran on that
beast is amazing. To realize that the kernel was only 8k (including buffers)
is amazing!!!

Have you ever seen a full ANSI fortran compiler fit in under 50kb ? On the
8, fort II fit in 8k and fort IV in 12.

People hated the thing since their literals squashed up against the end of
their code, but the best part of it was that it forced the mind to take the
extra steps (now thought of as superfolous) to optimize the code and above
that, optimize the algorithm to complete the task.

It was not uncommon for people to spend days trying to squash programs down
to size and to squash an extra word out of a page.  I bet you that given a
simple task, I can come up with an extremely compact pdp-8 version of the
code that is tighter than those produced for multi register machines.

My experience on that machine has become invaluable now.

Don't knock it till you try it (or if you have, just think that this was
loved by scientists and lab techs around the world less than 20 years ago).

					Peter Barada
					ima!pbear!peterb
					ihnp4!inmet!pbear!peterb