"Joseph Kesselman/CAM/Lotus" <[EMAIL PROTECTED]> writes: > Re DOMNode versus DOM_Node naming conventions: I understand the desire to > not break existing code, and I understand the desire to give the new > interfaces the most "standard" name possible, but I'm just a bit nervous > about this particular solution. I think that's going to be a bit > error-prone. We're going to spend a lot of time explaining the difference > between "DOM Node" (in the abstract), "DOMNode" (new interface), and "DOM > Underscore Node" (old implementation)... and if we forget to explicitly say > "Underscore" during a phone call, or if someone who doesn't understand the > distinction tries to mix the two or simply leaves the underscore out as a > typo, it's going to be a nuisance to discuss. > > I have to admit that I don't yet have a better alternative to offer.
I have to agree here. I don't have any (paying) clients to support, so that doesn't bother me, and luckily the perl API could care less whether there are C++ object handles or vanilla pointers, so that hasn't broken things for me either. But the naming is a bitch. Currently I only support the new IDOM implementation. If you change the names of the classes, I can cover that up, but all the Xerces-C documentation will be different. I believe that there is a simple solution: Bump the Major Version number to 2.0 This describes an architectural change of large proportion, that in this case is *not* backwards compatible for the DOM. If people want to use 2.0 they will have to port their DOM applications. If they want to continue using the old DOM, they should stick with 1.7.0. jas. --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
