Palaver's present and future
New features are memcache support for pgsql backend, and a memcache backend. The memcache backend is a more robust version of the in memory storage. All data is stored in memcached. This allows for palaver restarts without losing all of your data. It is not not as persistent as the other storage mechanisms but if you want a quick and fast MUC then it works, and is a bit better than the in memory dict storage.
I have also been adding MUC support to Wokkel and would like to use that protocol implementation for palaver.
The next release will be dependent on wokkel. So, palaver will then become just the storage mechanism code. Eventually, the wokkel code may make it to twisted words. I think this staging method will allow for quick use and testing. So, by the time it reaches twisted words it will have been used and tested well, beyond the required unit tests.
I am excited about the future of XMPP and Twisted words. The flexibility of twisted is gonna allow for a lot of cool components and servers!