And now for the political answers... ;)
On Wed, 2005-11-16 at 16:30 -0500, Kevin McCann wrote:
I don't mean to be a pessimist, but I don't think this will ever happen.
I hope this isn't the case, but realistically, things have to change before we're going to make significant progress. I hope that's more along the lines of a huge lotto winning rather than becoming unexpectedly unemployed ;). Seriously though, Kevin is essentially right that the development process needs to open up so that it can progress regardless of my availability.
- The Mailman project is not as open as it could be. There is too much control over who can contribute code, how they can do it, when they can do it. I understand wanting to maintain quality and stability, but effort and goodwill are going to waste.
I've been meaning to do this for many weeks now, but haven't had time to do it properly. So let me just blurt it out now and welcome Mark Sapiro to the development team. As you know Tokio Kikuchi has been doing a great job in getting the trunk back in shape and committing lots of fixes to the 2.1 branch. Mark has now joined us, as I'm sure those of you watching the checkins list have noticed. Welcome, Mark!
If there are people out there that think they have the background, skills, desire, and time to also help out, please let us know. I think we can reasonably grow the developer community and still maintain a high level of quality. We don't have a process in place to approve new developers, but I think in general we can model ourselves on the Python community. It's a meritocracy and membership requires an existing committer to vouch for you. I hope we don't need to get real formal here.
- Not everyone sees the need for a highly integratable MLM, despite the fact that people have been begging for it for half a decade. They beg for it on the Mailman list. They beg for it on the Sympa list. But MLM developers apparently do not work with organizations or people who see the need to integrate MLM's with other collaborative tools. MLM developers do not buy into the concept of making the three main data stores available in SQL (those being list config, member and message archives) so that one can easily get and update this data from within a Drupal, Mambo or Xaraya CMS. "Why on God's Green Earth would you want to do that?" they wonder, despite the fact that many, many people have been drooling over the thought of being able to do so for a long time now.
I disagree that MLM developers do not see the value in this. I definitely do. That I haven't been able to /do/ anything about that makes me sad, but I'm definitely in favor of Mailman becoming less of an island and more of a component in the larger universe of collaboration tools. SQL makes my brain hurt, but I think there are Pythonic ways to accomplish the goal of making the three data stores available to external SQL-speaking tools.
- Backward compatability is an issue and puts the kebosh on dramatic departures from exisitng MLM versions.
There's no question that backward compatibility is an issue. For example, I think it's going to be very difficult to migrate a Mailman 2 system to any future Mailman 3 database. But it can probably be done with a lot of heuristics and hints from the admins.
With a Mailman 3, radical changes would be needed, in my opinion, but are developers willing to have MM3 be a new, different, separate beast than MM2? I have a feeling that there is not enough wilingness to let MM3 be a fresh, new start.
It has to be. It also has to avoid second system syndrome failure. That's the challenge. No one's going to argue that the existing default data model is usable in Mailman 3.
Personally, I think a new MLM is needed, built from the ground up, and taking into account today's wants and needs. An MLM built for the 21st century. Completely open source and well-managed by people who have the time and the inclination to do it. New Project with New Ideas and Eager People. Either MM3 needs to start happening or a new MLM project needs to be created. If some people don't have time, fine, but then loosen the reins. Let it happen. And if it comes down to money, well, some people may be surprised at the amount of funding that is available for these kinds of projects. But the projects need to be able to move forward, unencumbered by control and competing commitments. Otherwise nobody will fund it.
I agree. I hope MM3 can be that vehicle because I think we've solved a lot of the problems any MLM is going to have. I think we have a pretty good architecture for the backend delivery machinery. I'd save a lot of that (with some updating and streamlining, but most of that's already in the MM3 code base -- yes, there /is/ an MM3 code base!).
Again, I'm not trying to throw out criticisms, just stating realities.
Do developers here have any comments? Is there interest to move forward with MM3 now, one way or another? Is there interest among other parties to start a new project? Who wants to see a highly integratable MLM, and are you willing to contribute in any of these areas: design, coding, project management, documentation, funding?
Great questions. Can we move this to the mailman3-dev list though? I'll probably be taking some time off during the holidays and I'm planning to spend some time discussing my thoughts on it.
-Barry