Date:	Mon, 6 May 1996 09:45:22 +0200 (IST)
From:	Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
To:	DJ Delorie <dj AT delorie DOT com>
Cc:	djgpp-workers AT delorie DOT com
Subject:	Re: [rms AT gnu DOT ai DOT mit DOT edu: Re: [david AT thermoteknix DOT co DOT uk: RE: gcc/as/ld, COFF and embedding]]
In-Reply-To:	<199605060120.VAA08949@delorie.com>
Message-Id:	<Pine.SUN.3.91.960506093435.13129L-100000@is>
Mime-Version:	1.0

On Sun, 5 May 1996, DJ Delorie wrote:

> Background: someone installed djgpp, had problems, sent e-mail to rms.
> RMS thinks that the bug reporting mechanism isn't obvious enough.
[snip]
> Perhaps it would be good for each zip file to come with a README which
> tells people where to find the FAQ and the bug reporting procedures.

My experience (both before and after the FAQ was written) is that no
amount of readme files, no matter how close and handy they are, will
ensure that nobody ever bothers RMS with bug reports that should have been
directed to DJGPP.  As long as there is no way to physically prevent
people from sending e-mail without passing a readme/faq reading exam,
somebody somewhere will see the name ``GCC'' and write to FSF. 

So I think that adding a readme to the packages is a good idea only if it
doesn't complicate the distribution and packaging process (a minor gripe
would be that I'll have to cope with unzip programs asking me whether to
overwrite the existing copy; that annoyed me quite a bit in v1.x with all
those readme.dj lying around).  It cannot stop people from writing to
wrong addresses.  It cannot even help us make our conscience clean,
because there's always ways to make readme's more clear and unequivocal. 

- Raw text -