On Sun, Dec 02, 2001 at 06:43:24PM -0800, Thomas Bushnell, BSG wrote:
> I understand what "software" means, and I guess it's quite sad that
> you don't. Oh well. I don't claim there *is* a rigid
> definition--it's *you* who are seemingly obsessed with the need to
> rigidly define everything in sight for fear that some pedant or
> sophist will accidentally or deliberately get it wrong.
This is a flagrant distortion, and fallacious reasoning to boot, so I
won't dignify it be replying further.
> > I submit that I'm not the only one. Sunnavind Fenderson and Scott Dier
> > don't appear to think the issue is better left swept under the rug.
> > Bernd Warken and Thomas Uwe Guettmueller don't seem to either, though it
> > is unclear to me that they agree with either one of us as to the
> > particulars.
>
> So why not a policy that calls for judgment?
Mine does. See how much invariant text there is. If there's more than
a certain amount, decide whether you still think the package should be
in the Debian GNU/Linux Distribution, and if so, raise the issue on
debian-legal so that the judgement of several people can applied.
I'm happy to add this language, or something similar, to my proposal if
it would alleviate some of your concerns.
> You move from "we need a policy" to "we need a policy that could be
> implemented by an automated system devoid of human judgment".
Incorrect, and I'm already on record indicating otherwise.
Message-ID: <[🔎] 20011202075756.GD26052@deadbeast.net>
"Whether certain language is legally binding or not is always a
determination that will have to be made by a human."
Why is it so difficult for you to consider my actual proposal as opposed
to the distorted straw man you keep presenting? To be honest, you
surprise me; I would not have expected such tactics from you.
> Before you propose any more policies, I expect *you* to rather more
> thoroughly investigate their impact.
I intend to. I'm sorry to offend you by asking people more familiar
with the GNU Emacs Manual to assist.
> The subject of the Emacs and GCC manuals have come up several times;
> you are the one proposing proposals, why should not you be the one to
> gather the information on whether these given manuals do or do not run
> afoul of your proposed standard?!
Again, I'm sorry to offend you by asking people more familiar
with the GNU Emacs Manual to assist. You are using the GCC manual as a
red herring; I'm already on record regarding the impact of my proposal
on it, in the very message containing my proposal.
Message-ID: <[🔎] 20011201175129.C22208@deadbeast.net>
"5) The only package already in Debian that I know of that may be
affected by this guideline is, unfortunately, the GNU Emacs Manual.
Other GNU manuals, such as those for gawk, gcc, make, texinfo, and
glibc, are not affected as far as I have been able to determine (I own
paper copies of these manuals and did check). Even the GNU Emacs Manual
itself may not be affected depending on the quantity of material within
it identified as Invariant."
Furthermore, I used the GCC Manual itself to help me arrive at my figure
of 32,768 bytes (again, Anthony Towns suggested 32,000 and I accepted
that):
Message-ID: <[🔎] 20011201175129.C22208@deadbeast.net>
The size of the GNU GPL and the "Funding Free Software" portion
of the gcc manual together in plain-text form is 20,410 bytes.
This is *without* regarding all adjacent white space characters
as one byte, and without excluding the portion of the GPL that
is actually a binding license. Once that is done (condensing
whitespace and omitting the "TERMS AND CONDITIONS" part of the
GNU GPL, which are already covered by clause 2 above), this
auxiliary material consumes only 7,928 bytes. Therefore, 32,768
bytes strikes me as a reasonable limit.
If you have to selectively ignore my emails in an effort to score
rhetorical points, then you are undermining your position.
> I would be ok with a proportional limit, provided that it was set high
> enough to cause no problems for things already in the archive.
So anything failing DFSG 3 that happens to be in main due to an
oversight should be grandfather by my proposal? That's what "no
problems" means, and would be grounds for rejecting my proposal outright
as an attempt to repeal DFSG 3.
No, the existence of packages with unmodifiable text already in main is
something that should inform our process, but cannot be determinative
because of the possibility that there is already something in main that
should not be. My proposal is a guideline, not a suicide pact.
I believe Debian should have a standard a priori the GNU Emacs Manual
(for example), and not reason backwards on the assumption that
everything that is in main must belong there. People find DFSG
violations in main regularly. The intent of my proposal is not to grant
categorical immunity to any class of these violations.
--
G. Branden Robinson | Software engineering: that part of
Debian GNU/Linux | computer science which is too
branden@debian.org | difficult for the computer
http://people.debian.org/~branden/ | scientist.
Attachment:
pgpU4qvb9UB1p.pgp
Description: PGP signature