Glen, There's tons of stuff to do and so little time, can we please defer it as per others feelings?
thanks, dims On 4/27/06, Glen Daniels <[EMAIL PROTECTED]> wrote: > Eran Chinthaka wrote: > > I tried to do this with Sandesha2, Policy and Axis2 in the project, in > > addition to Axiom. I found that this will change 242 places in 83 files. > > Doing this kind of non-critical refactoring is not that good, at least > > in the current stage. > > a) This kind of refactor doesn't change functionality at all - therefore > it's a much less intrusive/worrysome change than some kind of behavioral > thing which might actually affect the operation of dependent code (this > is just a simple rename/recompile). What's the worry? > > b) It is much better to change this stuff before 1.0. Yes, we'll do > stuff like this after 1.0 but it'll be a bit more of a pain, so I'd > rather get it out of the way. > > c) 83 files across four projects really isn't that many (were you really > surprised by that number?), and the tools make this kind of thing easy. > > I would much rather do this change now. The only valid concern I can > see is worrying about other people's code which might already depend on > this API. If we're really concerned about that (I don't think it's a > big deal before 1.0), we could deprecate getName() and still change all > the Apache code to use getURI()/getNamespaceURI() - then get rid of it > in 1.1 or 1.2. > > I'll certainly live with it if everyone else feels we should leave it, I > just hope that this isn't a trend towards locking down these APIs > forever... I guarantee we didn't get everything right the first time > around and would like to be able to fix things. :) My apologies, by the > way, if I come across as cranky about this stuff - I don't intend to, I > just want things to be as good/easy/clean as they can be. > > --Glen > -- Davanum Srinivas : http://wso2.com/blogs/
