On Tue, 2004-04-13 at 19:55, robert burrell donkin wrote: > hi simon > > the stuff in the sandbox is very old. > > the branch contains refactoring work, a rewriting of the way that the > reading works (plus some other simplifications). one of the problems > with the original betwixt code was that it was complex and monolithic. > i've been trying to break it up so that it's easier to understand and > easier to add variations in behaviour. > > the refactoring is (i think) nearly finish. what i need to do now is to > go through and remove duplicate and redundant methods (artifacts of the > refactoring), simplify over-complex code, re-assess the new interfaces > to make sure that they are right and document. > > development of new features has continued on the main branch. > > so, i'm not sure how best to answer your question. > > if you're looking for good ideas then there quite a gap between any of > the current implementation and how i think betwixt needs to work. > > if you're thinking about getting a bit more involve with betwixt then > it'd probably be best for me to explain the general plan and then point > out where the latest bits of code are to be found. (betwixt needs some > developer documentation.) > > this is probably a better time (for anyone who fancies it) to get > involved with betwixt than it's been for a long while. the new > refactored code seems to be easier for new developers to understand and > safely patch than the old, i think. (brian's certainly submitted a > number of good ones recently, thanks for that.) i hope sometime soon to > merge the new branch with the old and more eyes would be greatly > appreciated. i want to catch as much as possible to ensure good > backwards compatibility in the future. i've also been working with > jochen on adding support for betwixt-start-from-java to jaxme, a JAXB > implementation. if there are enough volunteers, i'm pretty confident > that big performance improvements (which is one area of weakness) could > be made relatively easier. > > - robert
Thanks Robert. I'm currently having a look at all the commons projects just to see what's going on in commons land. I particularly want to be aware of how other projects use Digester, as I hope to be involved in Digester for quite a while. I generally feel that people who intend to use a library are the best ones to work on it, and I don't currently have a need for object->xml mapping. But once I get into the code, who knows... Given the state described above, I will have a look at the stuff on the latest branch. Thanks for the info. Regards, Simon --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
