----- Original Message ---- > From: Tom Hobbs <[EMAIL PROTECTED]> > To: [email protected] > Sent: Monday, November 10, 2008 9:05:29 AM > Subject: RE: River-261, namespace change > > Hi all, > > It's my feeling that the namespace change should be deferred until after > AR2. Not being sure, but working on the assumption that, AR2 is near to > being ready it makes sense to tie up the last few bits and get it > released. > > Although the namespace change, at a superficial level, feels like a > simple change; given Jim's example from the IiopExporter, there are > probably more than a few places where we might get tripped up. Or maybe > I'm just projecting my own fallibility... :-) > > I would suggest releasing AR2 without it, but then releasing AR2.5 > quickly afterwards. This can be considered a rebranding exercise which > we could tie back to some of Niclas' points in his "Future of > Jini/River" post. > > So this would include things like; > > - River-261, namespace change
I assume this will make for a lot of refactoring of applications making heavy use of Jini. Would this not be appropriate for after the next release and then to go to something like 3.0 when the namespace change occurs? Maybe an initial 3.0 where the name change and some other patches can be applied, then 3.5 can be a feature enhancement release with higher goals? Seems from a 2.x to a 2.x such large changes are not expected, but between major releases it is common, and then folks have a fresh new Apache line and version starting witht he 3.x series. A clean break if you will. > - Move to Maven (which I'm personally not in favour of, but if it aids > adoption...) I don't see how this is going to aid adoption. Tomcat went through the same discussion, and decided to stick with Ant as there seemingly was no real benefit in going Maven for the build system. Now, releasing libraries into the main Maven repository would be good, but as far as moving the build that way, what purpose is that actually going to serve besides busy work? > - Reconciliation of the "million documents, a couple of sites, dozens of > related projects" > - Make the build/test cycle much easier and complete for newly > downloaded sources Seems these and the name change would be a good move forward. To me personally, it would make sense with this and a name change to call it River 3.0 and be done with AR...move out of incubator? That would show a good community too which I saw someone mention. I don't know if Jini can ever be dropped though...such a legacy even for people who have not used it. Apache River (Jini)? > - Invention of some intelligent-defaults to make the black-box install > more simple for the general case > - Invention of some "lofty goals" for where the project is going and > what it should do rather than a series of non-show stopping bug fixes > 3.5? Maybe some patch releases in between? > We should be pushing for graduation harder so the technology starts to > get known as "Apache River", rather than "The project formally known as > Jini". But again, if AR2 is almost ready to go, then I don't think we > should delay it for this. > Formally known or Apache River I don't think it matters as much as looking alive and that the project is moving forward, but they do matter none the less, and especially getting rid of the AR1,2,etc as that makes people wonder...what the heck is that. Like I said above though I think the name Jini needs to be mentioned for a while until folks see River and Jini as synonymous. The things mentioned before would certainly make the community/project look alive and vibrant if they can happen. It seems a good patch release for now with any patches which can be put in if there is going to be a push for that. If there isn't a push for a patch release, i.e. folks get their patches in etc and move forward, I would give an opinion it would be best to move forward with the namespace change after a 2.5 release with the patches which have been applied now or can be applied now, and work towards a 3.0 release with the documentation, build, etc changes plus patches. Nothing worse than folks waiting on some patches then get them and realize they have to also do a bunch of refactoring to get some bug fixes. Probably put in the release notes that the upcoming release will be one to move forward and refactoring will be expected. Then while moving towards that 3.0 push the community can grow and it is a good time for new folks like myself and others to get involved more when documentation, build, etc are being flushed out for the masses. > I know I've not been involved much or for long, but this is my two pence > worth. > Same here. Wade
