On Mon, Sep 29, 2003 at 10:01:19AM -0400, Jeremy Hankins wrote: > Burden of proof arguments are, at best, very trick to make -- I > suggest you not rely on it. Certainly I don't buy it in this case. > Unless you can actually point to someplace that says this is current > practice, I don't think you have a basis for saying that it is > actually a conscious practice at all.
Well... you would have to claim that the people who drafted and discussed the Social Contract, and the 90 who voted for it, were unaware of the GNU Manifesto and its presence in the emacs packages. This is an extraordinary claim. The document was, if anything, better known then than it is now (at least if you divide by the size of the free software community). I certainly knew about it long before I got involved with Debian. [practical problems] > > That doesn't seem too hard. Outdated info is one example already > suggested -- out-of-date references to charitable orgs, or addresses, > etc. Or what about a case where the snippet is written in the first > person, and the project changes hands? It would likely be desirable > to rewrite it referring to the original author. In such a case we can always drop it from the package. Rewriting it might be better, but our options are to drop it now or to keep it while it's relevant. Is there any advantage to dropping it _before_ it becomes outdated? Would anyone start to rely on the presence of a particular snippet, for example? > Let's split the question in two: > > * Should snippets be unmodifiable? Does it serve any purpose for the > community? There's first the question of whether modifiable snippets are an option. I'd like to have some more examples than the GNU Manifesto and the hypothetical one, actually. Is the emacs etc directory the only source of real snippets? Richard Braakman
