Re: Bug#138541: ITP: debian-sanitize (was Re: inappropriate racist and other offensive material)
On Mon, Mar 18, 2002 at 11:53:19AM +1200, Nick Phillips wrote:
> On Sun, Mar 17, 2002 at 03:11:06PM -0500, mdanish@andrew.cmu.edu wrote:
>
> > Except that it's not granular enough: I can already pick out an easy example
> > of Common Lisp programmers who hate emacs, love vim. And functional
> > programmers in Haskell and SML who are the same way.
>
> OK, so they form a constituency of "Functional Programmers who Use Vim".
> Or go and join the existing vi group and persuade them to drop their
> objections to functional programming.
>
> Hey, that's politics.
>
> > What one needs is a system that can be agreed upon objectively;
>
> No, what one needs is a system that allows everyone who's got nothing
> better to do than sit around passing moral judgements on other peoples'
> efforts sit in their corner and make their judgements, but without
> affecting everybody else.
The point of the system I proposed is to provide an _objective_ framework
for people to base their own moral decisions on. That means to reduce
the system to simple, easily exibited terms. Saying something is
"unsuitable for minors" is a subjective, debatable categorization.
Saying that something has pictures of naked women in it is something
that can be demonstrated with your favorite image viewing program.
Whether you choose to elide pictures of naked women from your system
or not is a decision left to the individual user. The main point is
that it can be done on an _individual_ basis within the provided
framework, and that the framework is based on _objective_ criteria.
>
> There is already such a system. Anybody who wants to provide a cuss-free
> Debian is free to set up aan archive which contains packages such as
> bitchx-clean (bitchx minus the "bad" bits) which replaces, conflicts
> with & provides bitchx. Or whatever.
So every user needs to setup their own little archive for their own
personal moral decisions?
>
> The issue seems to be whether we should make it easier for people to
> do this. One way we could do this would be to provide a package called
> debian-censor (just so as to be clear what they're doing ;) ) which
> asks the admin for a URL for a "ratings file". Where the admin gets
> that ratings file from is up to them - they can get it from the
> Christian Coalition, the FSF, the Tourette's Syndrome Fan Club or
> wherever. It could even be possible to merge two such lists.
Certainly a possible implementation. But I think you have a different
idea in mind judging from below:
>
> In fact, this could become quite a powerful way to select packages
> in the first place - the GNOME people could provide a list that specifically
> includes all the stuff that they like, the Christian Coalition could
> provide a file that includes the bible-kjv-text but rejects unconditionally
> anything that mentions sex, the Emacs people could provide a file that
> unconditionally rejects vi and derivatives, the vi people could
> unconditionally reject emacs and only mildly object to functional
> programming, so that the strongly-rated inclusion of Common LISP by the
> functional programmers could override it....
This is a more bothersome system than I proposed because it requires
every single group to rate every single package. My proposal is to
pick criteria that cannot be disputed (such as the existence of "fuck"
in the program's output) and let people choose which criteria are
suitable for their system. A separate "ratings" file, even multiple,
could be certainly be maintained; its content would show what categories
contained what programs.
Sidenote: Common Lisp isn't a functional programming language and there
are certain people, to remain unnamed, out there who would assert this
matter a great deal more, to say the least. (Which is why I was careful
to mention both Haskell and SML).
It's a good example of why this project must be customizable on an individual
basis, though, if it wasn't already apparent.
>
> ...in a sense, a beefed-up version of tasksel.
>
> The file format should be simple enough that any fuckwit with too much
> time on their hands can produce something "useful", and there we go.
>
> Everybody happy.
Wanna make everybody happy? Then don't simplify the system to the point
where you cannot implement a meta-protocol. A meta-protocol would allow
changes to be made to the behavior of the system in a clear and consistent
manner, giving even more flexibility with no loss of power.
Consider that the vast majority of Common Lisp Object System implementations
are implemented using techniques outlined in the "Art of the Meta Object
Protocol"[1], and see how it has resulted in one of the most flexible
object systems available.
[1] http://www.elwoodcorp.com/alu/mop/
--
; Matthew Danish <mdanish@andrew.cmu.edu>
; OpenPGP public key: C24B6010 on keyring.debian.org
; Signed or encrypted mail welcome.
; "There is no dark side of the moon really; matter of fact, it's all dark."
Reply to: