Robert Burrell Donkin wrote: > On Tue, May 12, 2009 at 3:28 PM, Bernd Fondermann <[email protected]> > wrote: > > <snip> > >> "Contribution" covers everything from installing, running and reporting >> test results to bug reports, documentation patches, code patches. >> >> Today I contributed a patch to another Apache project I never >> contributed to before. The patch was utterly trivial: One word changed >> in a toString() method. It was created in less than 5 minutes. The patch >> was applied within less than 10 minutes! >> >> What I want to say by this is: It doesn't take much to make a difference >> on an open source project. The borders between user, contributor and >> committer are floating. > > +1 > >> I can understand why people are frustrated. I am, too. >> It's because something has changed within the last couple of month: >> Three committers who were heavily using James on a daily basis moved on. >> They were first hand testers of TRUNK. >> >> Instead, at the moment, TRUNK is largely untested. So nobody knows >> exactly how good it really is (or they don't report here). >> >> So if the community asks for a milestone release, we all must be aware >> of the fact that we then need to do actual testing on the released >> software. Otherwise, releasing a milestone serves no real purpose. So >> others need to jump in and test. > > +1 > > one of the conflicts is that james is both a final user application > and an extensible platform for developing mail applications > > for users who are interested just in a plain mail application then > avalon-phoenix shouldn't be a problem. it works well and has been well > tested in production. 2.3.x really is the right choice. > > the avalon-phoenix framework has major issues as a development > platform. most developers prefer spring. the 3.x version is the right > choice for developers looking to extend james. > > most of the developers who remain active are interested in developing > mail applications rather than creating and maintaining a basic user > email application. > > IMHO both codebases are just too big for new developers. i would > prefer lots of smaller, more accessible libraries from which the two > versions are build. > > if people aren't willing to contribute to testing 3.0 then one option > would be to think about 2.x as the stable, user application and 3.x as > the extensible developer version. it should be possible to extend > 2.x's lifespan by factoring out libraries from 3.x to replace code in > 2.x. avalon-pheonix support could then be dropped from 3.0.
+1 (insightful) Bernd --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
