Gang & whoever is finding this interesting... I have initiated some initial discussions to see what kind of support there would be to "Help Depot to Help Us", i.e. transfer of Avalon Repository into the Depot codebase, and refactoring it into a more generic package than currently is the case.
For those of you who doesn't know; Depot is a project currently in Incubator, with the mission to solve "Jar Hell" in a cross-ASF manner. Their target is to unite the Maven, Gump, Ant and others, who have their ways of solving the Jar problem that plagues us all. Many existing goals overlaps with Avalon Repository, and it will be our duty to introduce our concerns as new goals for Depot. The Depot cause is noble, but progess is somewhat slow. The project is still somewhat in the infant stages and need more of a community around it to succeed. I think Avalon is the ideal partner for the Depot community/project. We have very strong use-cases, working code for these use-cases, and a vision that far exceeds some of the members in the Depot community. By transferring Avalon Repository, we would be breaking a principal that I am generally against, that is being dependent on pre-release code, in this case in form of a project-in-creation. This is a catch-22, Depot won't graduate out of Incubator until it has gathered a strong community and general support outside the handful committers, and noone wants to commit to it until it does. Together with the Depot people (they see it slightly differently, but the basis is the same), I have formulated a path which can work. 1. Avalon Repository is copied to the Depot codebase. 2. Avalon committers who are interested, gains access to the Depot codebase. 3. Changes to Avalon Repository here continues, and changes are applied to the Depot codebase manually in parallel. 4. The Depot codebase is refactored; - Package names changed into the Depot name space. - Splitting the various concerns into using the existing and/or new APIs in other Depot subprojects. 5. Merlin Runtime is 'adapted' to use the Depot version of Repository. 6. Avalon Repository cease to exist. 7. Likelihood of Depot graduating out of Incubator is much higher. The steps 1-4 is so that in the event of things doesn't work out, we have lost some effort, but don't need another effort to restore the damage. I also want to point out that Merlin itself will remain compatible, but I will not issue a guarantee that any 3rd party code using the Avalon Repository directly will remain compatible. I happen to believe there ain't many other projects/products that depends on Avalon Repository directly, so no point making compatibility a requirement. And where it becomes an issue, we would need to provide wrappers in the old package names, blah blah blah... not really a difficult issue. Furthermore, I think there are great 'political value' in this effort as well. It again shows that Avalon can proceed in positive spirit and in this case co-operate with other projects. Nicola Ken Barrozi, one of the lead Depot committers (also former Avalon strongman), though generally very negative to Avalon (or has been[1]), is very positive to the cross-pollination and seems to be looking forward to this. The following people have already shown support for the initiative, and indicated that they are interested to help out in this effort; * Niclas Hedhman * Alex Karasulu * Stephen McConnell * Timothy Bennett * Aaron Farr If anyone else, has such desire the door will remain open for any Avalon committer to join in to the Depot project. If anyone has some objections, now is the time to speak up. WDYAT??? Cheers Niclas P.S. This is not a vote. I want to see comments, the pros/cons and so on... [1] http://jroller.com/comments/nicolaken?anchor=avalon_must_die -- +------//-------------------+ / http://www.bali.ac / / http://niclas.hedhman.org / +------//-------------------+ --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
