The reason that I personally voted for the HP code base boils down to a few simple points.
1) The code maturity. The HP code base has been through some maturity growth so I expect that more of the design has been shaken out and core runtime suits a potentially broader audience. 2) The JAX-RS spec compliance IMO wasn't that far off and we can probably reach that fairly quickly. I did plug in the HP runtime into some of the IBM fvt tests and managed to get a feel for how far we are (subresouce locators, minor things mentioned in JAX-RS javadoc that isn't in the spec itself, etc. need some fixes but nothing that can't be done). 3) Feature set that suited the needs of more customers. Although I think we should (still) keep the requirements to the barest of minimums on the core runtime (i.e. spring kept modularized), the HP code base already had that work done. To elaborate a bit, I *personally today* do not have an interest in Spring support or some of the other features but I believe there is a need for that and I didn't see any major performance / architectural penalty for these features. There are things I do find interesting like the client. Beyond the bug fixes and the testcases being ported over, I do think we should consider caching the calculations for target method selection and choosing of Providers of repeated requests (some of that work in JIRA WINK-3). How we go about that, I will create a separate thread. I'm still looking at the best way to port the testcases over right now. On Wed, Jun 24, 2009 at 8:27 AM, Nicholas L Gallardo<[email protected]> wrote: > Kevan, > >> Note that there are some serious gaps in the above -- I haven't seen >> discussion of the two codebases on our dev list, discussion about why >> one codebase over another, a transcript of the IRC discussion, or a >> summary of the IRC discussion, etc. Hoping that we can fill in some of >> these gaps? > > Fair points. I take responsibility for not getting the summary and > transcript out in time. The latter is available and I'm working on the > summary right now. > > While I'm doing that, does someone else care to net out the discussion of > why the HP code over the IBM code? > > -Nick > > > > Nicholas Gallardo > WebSphere - REST & WebServices Development > [email protected] > Phone: 512-286-6258 > Building: 903 / 5G-016 > Kevan Miller <[email protected]> > > > Kevan Miller <[email protected]> > > 06/23/2009 09:38 PM > > Please respond to > [email protected] > > To > [email protected] > cc > > Subject > Re: [VOTE] Selecting Wink baseline > > On Jun 23, 2009, at 10:40 AM, Snitkovsky, Martin wrote: > >> Hi all, >> >> >> I would like to launch the vote for selecting the HP code as a Wink >> baseline. >> >> +1 to recommend HP codebase as a Wink baseline. > > I doubt that I would have any objection. And would be *extremely* hard- > pressed to go against any community consensus on the subject. > However... as a project mentor, the most important behavior that I'm > looking for from a new community is clear and open communication on > our dev list. And I'm not seeing it... > > I was traveling this morning and could not attend the IRC chat (which > is probably just as well). From my perspective, here's what I've seen: > > 1) A call for a community conference call, which was properly morphed > (IMO) into a community IRC chat > 2) Postiing of some documentation of the separate codebases on the > community wiki (I must confess that I haven't read this documentation) > 3) I assume that the IRC chat took place > 4) A call for a vote to select the HP code as the baseline for the > Wink implementation. > > Note that there are some serious gaps in the above -- I haven't seen > discussion of the two codebases on our dev list, discussion about why > one codebase over another, a transcript of the IRC discussion, or a > summary of the IRC discussion, etc. Hoping that we can fill in some of > these gaps? > > --kevan > > > > > > > > > -- - Bryant Luk
