With our knosos project we have a sprint evaluation every month. It seams a good time to blog about my development and the plans of next month. Last month our pilot project started and I have been adding some small features to groups (like delete of groups and subgroups) as well as a expansion of the subscribe module. I had to hack the subscription module to make it work with our architecture, so again it will be functionalities hard to make a sepreate module for. I had to do two hacks, first add a "group" subscription next to the other types and second make sure only the content you have access to can be seen. For example we made a type called "private message" what is a cck-content with specific node access settings (user level so user audience is required). To make sure only the content the user has access to would be past I had to change the queries and use a "db_rewrite_query" (the hook is called in node acess module), but the tricky part was the queries where related to uid, not nid, so I had to find a way around it. It works fine now, but I think architectural speaking its a mess.
For the next moth I have actually to tasks, I realy need to get my basic modules in drupal head (node access, groups, groups-contrib) I have been it now for some time. The other part is to make taxonomy work with our system-theory approach of groups/users . Just like with node access and with groups/users the taxonomy would be context-dependent, well I'll implement it and see what it brings. The reason of all the delay to get the modules in the head is because the focus is on research for changing a CMS to a system management tool. I do would like to have other people use it and try it out, but it will stay experimental for probably another year. We can do very powerful stuff with the node access at the moment, but it is asking a lot from the users. To make the module workable a lot of other functions will be needed to reduce the time spend on the actions to post any thing (like the access profile proposal).