Michael G Schwern wrote: >> We do not want to support any random subclass of >> version.pm - we only want to support true version.pm objects. > > This, too, strikes me as odd. Why not?
Because the PAUSE indexer runs everything in a safe compartment, with a very restrictive methodology, and every additional class of version objects would have to be added manually. As I said, I would rather have implemented version objects as a first class type, rather than as a class. I'm just trying to keep the abstraction as tight as possible. That's not to say that version objects are a dead end once it hits 1.0. I already have version::Limit (allows a module author to strictly define what future releases are compatible with what earlier API's). At some point, I will release version::Math (to make it easier to create scripts to increment $VERSION when releasing modules to CPAN). But those things don't need to be supported during the release/build/install cycle. Only version.pm objects meet that particular need, because of all of the work I took to make sure it DTRT under all circumstances (e.g. it works "like a number" when it needs to and works "like a version" when it needs to). John -- John Peacock Director of Information Research and Technology Rowman & Littlefield Publishing Group 4501 Forbes Blvd Suite H Lanham, MD 20706 301-459-3366 x.5010 fax 301-429-5747