On Sat, Mar 16, 2002 at 04:52:19PM -0500, Jeff Licquia wrote: > On Sat, 2002-03-16 at 15:39, Steve Langasek wrote: > > I object to using any subjective method (such as popular vote) for > > determining which packages should be conflicted with in such a package. > OK; we can scratch the Conflicts: part unless someone else can give a > good reason to have it. The package can act like vrms, and provide a > report of some kind. Perhaps the package could also take configuration > (adding or removing packages for personal preference), and could > possibly produce a package via equivs or some such for those who want > it. > > In addition, if you use the simple criterion of having Debian developers > > indicate whether they find a given package offensive, I think the only > > two packages you'll get a majority of developers to say they find > > offensive are vi and emacs. > True. Unfortunately, when you're talking about something as subjective > as offense, there aren't many good classification systems that won't > themselves be offensive to someone. Democratic vote strikes me as one > of the few that's hard to challenge. Who, then, becomes the target audience for such a package? How can users tell if this package is something they want if all they know is that it measures how much stuff offends Debian Developers? Bad code offends Debian Developers; broken licenses offend Debian Developers. By and large, swear words do not offend Debian Developers. A democratic vote is difficult to challenge precisely because it's a BAD metric here. The only thing you're measuring is how much DD's like or dislike certain things. It neither guarantees that the excluded packages are offensive to a given user, nor gives a user the means to find out if he shares the values of the DDs who voted. > - Most developers, joking aside, are capable of distinguishing between > technical preference and moral repugnance. The more important assumption at issue here is that it's worthwhile to have a group of DDs decide on behalf of others what does or does not constitute an 'offensive' package, without being subjected to open review. Steve Langasek postmodern programmer
Attachment:
pgpZJLeiqDNIc.pgp
Description: PGP signature