Xref: utzoo comp.sys.amiga.tech:239 comp.sys.amiga:17461
Path: utzoo!mnetor!uunet!husc6!mailrus!nrl-cmf!ames!pasteur!agate!eris!mwm
From: mwm@eris (Mike (My watch has windows) Meyer)
Newsgroups: comp.sys.amiga.tech,comp.sys.amiga
Subject: Re: Assembly Source Level Collection (ASL)
Message-ID: <8554@agate.BERKELEY.EDU>
Date: 10 Apr 88 20:12:05 GMT
References: <2524@umd5.umd.edu>
Sender: usenet@agate.BERKELEY.EDU
Reply-To: mwm@eris.UUCP (Mike (My watch has windows) Meyer)
Organization: Missionaria Phonibalonica
Lines: 99
Keywords: assembly source level colloection ASL

In article <2524@umd5.umd.edu> spier@umd5.umd.edu (Lori Spier) writes:
<TITLE: Assembly Source Level Collection for the Amiga

Nice idea. But I've got some corrections, and a suggestion:

<I find that once you know the instructions WELL, you can
<be just as productive as a C programmer, with faster code.

Key word - "WELL". How many C-level and HLL langauges can you learn
well in the time it takes you to learn a non-trivial machine
instruction set well?

<I don't want to
<get on my soapbox, but for every program that you show me in C that is the
<same size and just as fast as the assembly language counterpart, I will show
<you five assembly programs which are just the opposite.

Ok. My favorite C progam is mg. The _most_ important thing it does is
run on my Amiga, my Sun, several VAXen and a Cray. When you've got an
editor written in assembler that does that, _then_ I'll switch :-).

The second most important thing is that new features added to one
version usually show up in all the others gratis. Make sure you get
that one, too.

<Well, I feel that the reason people are backing away from assembly language
<is that there really is not that many examples out in the PD and the ones
<which are out there, are not that easy to get a hold of or referenced for
<easy access.  My point is that if there was a central collection where people
<could get source code to aide them in programming in assembly language, more
<people would be inclined to do so.  My answer: create such a collection.

Not a bad idea. Writing things that aren't usefull on other machines
(e.g. - frags) in assembler doesn't cost anything in portability. On
the other hand, the PD C code for the Amiga is a little above average.
This makes it readable. Assembler code that is a little above average
is unreadable.

Since the goal is to provide examples to encourage people to write
assembler instead of C, you should definitely change:

<REQUIREMENTS
<------------
<
<The only requirement is that any source you donate is either yours, you
<created it, or it was placed in the public domain.  Before any code is
<placed on any of the disks, it will be checked out by myself and if it is
<necessary, I will contact the author.

To be:

The requirements are that the source code be either public domain or
publicly available, and that the code should provide a good example of
assembler programming.

Oh yeah - note the quiet change. Many people these days aren't putting
code in the public domain, as they lose all control of what others can
and can't do with the code. Instead, people copyright code, and then
state that "This code may be freely distributed under conditions X
except for Y" (X varies a lot, Y tends to prevent people from selling
the program for money). You might consider adding shareware to the
list, also. I wouldn't distribute it, but you might.

[I'm about to flame a bit. You may wish to skip from here on down.]

I don't agree that it's as easy to write C as to write in an assembler
you know well. I've been writing 68K assembler since '84, and I still
can't write it as fast as I can write C (Ok, I'll admit I've been
writing C since '79). The main reason isn't that the assembler is
harder to deal with - that's true for some machines, but not the 68K -
it's that I have to write significantly more english around the code
to bring the level of readability of assembler up to what I get when I
write C.

If you're on a brain-dead architechture that doesn't deal with stacks
well, or you're trying to get todays typical 10lb program to run on
one of yesterdays machines with 5lbs of memory, then you need to write
in assembler to get acceptable performance, or to get the thing to fit
at all. 

But with the Amiga, I've got a machine that spends most of it's cycles
waiting on me to type things at it, has an architechture that doesn't
punish you for using a stack, and has more memory than the first three
mainframes I worked with. Why work harder to have more of something I
already have in abundance?

Assembler has it's place, and I can and do write it when I need to.
Even on the Amiga, if I'm doing something CPU-intensive. But I'll
write the bulk of the code in C, and then spend the half the time I
saved by not writing it in assembler in the first place analyzing and
rewriting (maybe even in assembler) the 10% of the program that's
soaking up 90% of the time.

	<mike
--
But I'll survive, no you won't catch me,		Mike Meyer
I'll resist the urge that is tempting me,		ucbvax!mwm
I'll avert my eyes, keep you off my knee,		mwm@berkeley.edu
But it feels so good when you talk to me.		mwm@ucbjade.BITNET