Core plone development - panel discussionΒΆ

Tags: plone, ploneconf2007

Panel members: limi, jonstahl, optilude, zopepaul, wiggy.

What do we want?

Paul would like to have clarity whether Plone is a project or a framework; should we be a real humane CMS or a platform for further work by consultants? Martin thinks it is a bit hard to tell a good story about Plone nowadays as plone itself changes and so many layers below us change even harder. Limi's vision is to make Plone the Macintosh of the CMSs: easy to use, consistent. He wants to make it easy to get early successes with Plone, for instance by allowing plone to spit out static HTML so that it is usable with regular web hosting.

How can we know?

Jon emphasizes that many non-profits try had to be a listening organization. It is something that is hard. Hard, not because we're bad at it but just because it is hard. For organizing the 2006 conference he put a lot of effort in getting data out of previous organizers and out of the potential conference goers as he hadn't gone to any previous plone conference.

Influencing Plone

Limi's recommendation for people that want to give feedback: write a log. Just a one-line description in a text file every time you see a problem yourself or see someone have a problem. Joel Burton's whole "humane CMS" talk has just such a list as the basis. He keeps a huge text file and adds a line every time someone has a problem. Jon got a lot of initial influence by writing down, on his blog and in a non-confrontational way, what he ran into with plone.

Some comments from the room

Vincenzo: there's still some infrastructure problems in Plone. For instance, the versioning product lay around for a long time before it was added. Also linguaplone needs/needed to be pushed forward desperately. Shouldn't there be some foundation activity there? Pooling of resources/money? Limi added that linguaplone was a bootstrapping problem. Once the initial release was there, it became easier to get additional funding. Doing it as the foundation is hard: look at getpaid which is very transparent, but still you can get discussion about "whether we got our money's worth".

Limi also mentioned oxfam, that funded kupu, linguaplone and versioning. They got the right people to build it and invested a lot of money, but not a ridiculous amount. Paul adds that it can be hard to steer new development: on the Plone ship, everybody has a steering wheel. Goldegg was an "illuminating" experience. Godefroid added that we now have a lot of zope3 inside plone, for which goldegg was essential.

There is a gap between adding bugs and suggesting features. PLIPs are mostly administrative constructs. The gap is mostly "filled" by the plone developer mailinglist, where new developments are hashed out (and where perhaps the consensus boils up to add a PLIP).

Someone else said it was hard for them as a integrator company to make good choices with all the new functionality that's available. Archgenxml? Archetypes or formlib? Products/ or lib/python? Eggs? Various zope versions? How to make a good choice? Martin says it is impossible to make a top-down decision on this and enforce it on everybody. Limi says that it is perfectly OK to be conservative as an integrator. He himself uses archgenxml as it allows him to create products at google without being a good programmer. Archgenxml internally is a mess as so many people worked on it. But it does exactly what it promises to do and does it well. "Don't base your company on it, but do try to use it".

A good thing about goldegg was that the situation now is really different from when goldegg started. We as Plone developers got a lot more respect from the people lower down in the stack (zope, cmf) and we can now commit to cmf and zope and they cooperate with us. (Andreas Jung got a special round of big applause).

Supporting releases

More or less officially, the concept of separate infrastructure releases and UI releases has been left behind. Wichert mentions there's still a question of intermediate (like 3.1) releases. Do we want to do that? Martin would like it, but we have to make sure it is at most a 10-minute migration task. When we have both 2.5, 3.0 and 3.1, we really have to think about encouraging people to move from 2.5. Limi says we originally wanted to support older releases for one year. We're currently effectively doing more. One year is pretty good for an open source project. Vincenzo injects that the problem is more with the add-on products. Martin says we ought to have a discussion about this afterwards and not make any fixed-sounding announcements here.

Technorati tag: logo

About me

My name is Reinout van Rees and I work a lot with Python (programming language) and Django (website framework). I live in The Netherlands and I'm happily married to Annie van Rees-Kooiman.

Weblog feeds

Most of my website content is in my weblog. You can keep up to date by subscribing to the automatic feeds (for instance with Google reader):