Section 2.3.5 says this: > All packages should use virtual package names where appropriate, and > arrange to create new ones if necessary. They should not use virtual > package names (except privately, amongst a cooperating group of > packages) unless they have been agreed upon and appear in the list > of virtual package names.
In reality, most (all?) virtual packages seem to fall into the category "except privately, amongst a cooperating group of packages". As such, this paragraph doesn't seem to reflect reality very well. I could just propose a rewrite which made adding a package to the list the special case instead... but I think this merits rethinking anyway. What exactly is the list of virtual package names supposed to achieve, and why should it constrain those which are used? [I'm skipping the justifications for these various options, it should be obvious; all of these should have an addendum of "and rewrite 2.3.5 accordingly"] Option #1: Ditch the whole thing. Leave it to maintainers to sort it out, and replace this paragraph with guidelines about what virtual packages are for. Option #2: As #1, but build a list of virtual packages during the policy build process. This is based on the notion that the list is useful documentation, but I don't think this is a very good idea; the list would probably be out of date more often than not. Option #3: Ditch the file, and provide a tool in the debian-policy package which finds the current list. Option #4: Define what is meant by "privately amoung packages" more accurately, and rewrite the paragraph in terms of that. Other ideas? I'm hovering between #1 and #4. [Default option: bicker about it and get nothing useful done] -- .''`. ** Debian GNU/Linux ** | Andrew Suffield : :' : http://www.debian.org/ | Dept. of Computing, `. `' | Imperial College, `- -><- | London, UK
