Subject: Re: Should I release ugly source?
From: John Gilmore <gnu@toad.com>
Date: Fri, 02 Apr 1999 11:28:37 -0800

> Does anyone have any past experience with "releasing ugly source" that
> they would care to share?

Well, I've never released any ugly code :-).  But the, uh, hard to follow
and very prone to bugs, yeah, that's the ticket, parts of GDB:

	*  did the job for ten thousand users, failed for a few
	*  got better in (almost) every release

Now a perspective from wearing my software user's hat.  Van Jacobson's
approach was to release binaries of interesting multimedia software
(like "vat" and "sd", used for multicasting).  He would promise the
source code "real soon now when we clean it up, because it's too
ugly".  But invariably the binary-without-source situation persisted
for years; in some cases source has never come out.  I found it
extremely frustrating.  Others did as well -- one result was that "sd"
was rewritten from scratch by someone else, and now everyone uses "sdr".
The experience did not improve my opinion of Van, who I think does
excellent research otherwise.

See, there was nothing you could do with his software except "be a
user".  No way to collaborate, no way to help, no way to fix bugs or
solve problems.  He never "reduced the transaction costs of
cooperation" by releasing the source.  He didn't "play well with
others".  In my world that counts much more against someone than
writing ugly code.

As someone said many years before computers, "I'm writing you a long
letter because I don't have time to write you a short one."  There
isn't always time to clean up the code after you learn how you
shouldn't have written it.  Send us the letter anyway, Dean!

	John

PS:  Even if you don't have time to clean it up, perhaps someone
who wants to write a clean patch can learn from your mistakes.