====
TODO
====

refactoring
-----------

- customized memcache lib (does this also belong to ultra memcache?)

   - the main reason why we need a custom concept is the support for update
     expiring values in memcache before they expire. We should use
     a metadata container which is able to store the expire data and the
     real data. This whould allow to set the data periodicaly back to 
     memcache before the data expires. (a meta data container whould contain
     the real data and an additional timestamp, others?)

   - other reasons why we need to merge the python memcache lib and our
     implementation into one implementation is that we do several things twice.

     a customized library whould allow to:
     
     - convert data only once after wrapping in a metadata container
     
     - use the "arbitrary 16-bit unsigned integer" concept as key for
       converters. This concept is hidden in the python-memcached library
       see: http://code.sixapart.com/svn/memcached/trunk/server/doc/protocol.txt

     - implement custom get, set methods with explicit converter usage. This
       allows to bypass the hole key and data conversion

   This means the new library whould be a mix of high level API and offer some
   base concept and more with a flexibler API
