Path: utzoo!utgpu!news-server.csri.toronto.edu!clyde.concordia.ca!thunder.mcrcim.mcgill.edu!mouse
From: mouse@thunder.mcrcim.mcgill.edu (der Mouse)
Newsgroups: comp.sys.next
Subject: Re: Why no source for News? (was Re: Where are we going?)
Message-ID: <1990Nov18.105149.273@thunder.mcrcim.mcgill.edu>
Date: 18 Nov 90 10:51:49 GMT
References: <MailManager.658170664.1575.lane@ssrg-next-1.stanford.edu> <11162@milton.u.washington.edu>
Organization: McGill Research Centre for Intelligent Machines
Lines: 65

In article <11162@milton.u.washington.edu>, wjs@milton.u.washington.edu (William Jon Shipley) writes:
> Christopher Lane writes:
>> @begin(pyrotechnics)
> [...]
>> Eric's 'stand on my shoulders and reach even farther' comment is
>> right on the mark.  Why aren't the archives loaded with well
>> written, useful object classes (with source) that others can use to
>> build even more powerful programs?

Well, I could suggest a few reasons, but let's stick to the point....

>> Why is there a 'binaries' directory at all on the NeXT archive?  If
>> 'News' came with source, maybe others would have fixed the bugs,
>> returned them to the author and we'd have a version that didn't blow
>> up routinely.  The archive maintainers shouldn't even accept 'binary
>> only' donations (IMHO).

> Well, I imagine you'll get lots of commentary about the rest of your
> post, but I felt, as the author of News, gee, I was in a unique
> position to respond to this paragraph.

> I guess it hurts a little bit to be flamed for a piece of software I
> spent quite a bit of time writing, and then just gave away.

I don't feel you were being flamed for the software per se as much as
for your choice to make it binary-only.  It certainly would never have
found its way onto my machine, if I had one capable of running it.  I
have a hard enough time accepting binary-only software from the
hardware vendor; no way am I going to stand for binary-only in a case
like that where I don't have to.

> When I implement [...description of the way Save Article will work...]

> This is the behavior I want for the program that has my name on it.

Well, I suppose it's your program.  But it seems a little odd to *want*
to be known as an author of completely unconfigurable software.  (With
source, at least it can be configured by hacking on the source, even if
there is no run-time configuration.)  *I*'d be almost ashamed to
release such a thing - and in fact I haven't: I have never released a
binary-only program.  (Before a few who know me dump on me about mcsh:
I am restricted to giving out binary copies by our license with AT&T, I
don't consider it distribution software (I don't plug it, for example),
and if someone doesn't want a copy because I can't give out source, I
will entirely understand.  If I could separate out my changes from the
licensed code, I would do so, and distribute the changes.)

> The strong feeling I get is no one wants to be tested before doing
> volunteer work; gift horses themselves resent being looked in the
> mouth.

As is only natural.  (You seem to be a bit of an example of this
yourself....)

> News is *NOT* a hack.  It is a completely new, totally
> object-oriented piece of software.  [...]

If we can't get at the source, it might as well be spaghetti assembly
for all the difference it makes.

					der Mouse

			old: mcgill-vision!mouse
			new: mouse@larry.mcrcim.mcgill.edu