RDF::RDF

Hoi, well I talked to the original authors of the RDF API I ported to PEAR again. They have decided that they rather stick to what they have and not participate in the PEAR version at all. The main problem is that they are fairly far ahead with their development, they have a nice userbase and a nice developerbase. They have an established API and structure and see little benefit in improving it. There are some imho severe limitations I solved as part of pearifying the code but appearently they havent been bitten by them enough: - there is no real error handling to speak of .. any error condition effectively means that RAP dies - some of the code is needlessly mixed (like there is RDQL code scattered inside classes that dont actually deal with RDQL) - improved directory structure to make code nagivation easier - the code is formatted quite horribly and doesnt follow a single CS The problem is that they are currently adding features and fixing bugs at a very fast rate. Especially the bug fixing part is fairly hard to keep up with, as their code is a bit hard to merge the fixes etc. All in all its also a duplicate effort. They are considering to reconsider PEAR once they are ready to move to PHP5, which will probably happen sometime in the first half of 2005. Now the question is what are we supposed todo? We have two packages that need an RDF API (XML_FOAF and XML_Annotea) and we may get others. We can either have external dependencies on the original RAP code or we can have our own fork (which is essentially what my RDF::RDF has now ended up being). The situation seems to be very sumoptimal eitherway. The fork is pretty much done and I guess we have the ressources to atleast merge all the bug fixes but we dont have the ressources to merge over all the new features and supported API's they keep adding. Opinions? regards, Lukas