The meeting agenda has been available for a while now. To complement it, here is my list of talking points:
MgdSchema is the new object abstraction back-end for Midgard proposed by Alexander and being developed by Piotras. With MgdSchema the Midgard content types would be defined using XML schemas instead of hardcoded into the Midgard libraries. While MidCOM datamanager already allows site developers to freely define their own content types, their storage using Midgard's parameters API causes performance issues.
The new model would solve the performance problem by translating schemas into database tables when Midgard is loaded. It would also make Midgard object types accessible from other programming languages by using the libdba and GObject technologies from the GNOME project. Alexander was able to create Python, Ruby and .Net/Mono bindings to his proof of concept during the Poznan meeting.
mRFC 0008 was passed this week, we also have to talk about implementing the new Midgard release process. I want to have a list of TODO items assigned to people by the time we leave Linköping. So far I see the following items:
- Set up the Midgard release calendar to midgard-project.org
- Create a translators' area on the site
- Consolidate and update translation instructions for Aegir, MidCOM and Spider
- Start gathering a list of translators with their assigned languages
- Create a report showing completion percentages for the different translations
- Schedule a "translation test drive" to try to get a number of languages to an acceptable 60-70% level
- Work out how to migrate 1.6 installations for the new filesystem-based MidCOM in 1.7
Since Piotras is attending, we will also take a look at solving the Midgard Core bug affecting the net.nemein.registrations component. The new mgd_debug_* functions in Midgard 1.6 should make that easier.
Updated 2004-11-12 from Linköping: Due to AnyKey's designer Robert Renling being linked with the GNOME Journal community, another important focus point for the gathering will be discussion on GNOME integration. As MgdSchema will essentially bring GNOME technologies to the server it would make sense to get the two communities communicate. The Open Source desktop severely needs server and groupware integration, and Midgard can help there.