HI Lennard, excellent question. I fear at the moment we can't give you the ultimate answer, but we can give you a couple of bits and pieces to consider:
1) We intend to join these two projects into one. But currently there is no roadmap how this should look like. We plan mid of April to start a discussion how we could join them and expect to give you a better answer then after this has happened. 2) None of the two is released so far. Because of this and because of 1) it is possible that any of the existing interfaces will be changed. 3) opencmis supports both AtomPub and web service today transparent to a user of the client API. There are two levels of the client API a provider interface (see provider packages) and a more convenient object-oriented client API (client packages). Currently the provider package is more mature than the client API, but the client API evolves quickly. See here (http://cwiki.apache.org/CMIS/opencmis-client-api.html) and here (http://cwiki.apache.org/CMIS/opencmis-provider-api.html) 4) There is a comparison sheet in the wiki between the two code streams: http://cwiki.apache.org/CMIS/chemistry-and-opencmis-comparison.html Let us know if you have more questions and it would be great to get feedback if you make your choice what your thoughts are. Jens -----Original Message----- From: Lennard Fuller [mailto:[email protected]] Sent: Mittwoch, 24. März 2010 01:32 To: [email protected] Subject: Am writing client code against the existing chemistry api... should I switch to the OpenCMIS? I'm currently writing a cmis client, and opted to take advantage of the chemistry api as it has greatly eased development. Since then I've noticed the addition of OpenCMIS and seen a good deal of activity around it. It may be far too early to tell, but is there any idea which client side api the eventual Chemistry will most resemble? The application I am working on is small, and refactoring to accommodate changes in the higher level api seem preferable to working directly with CMIS. 'Too early to tell' is a perfectly acceptable answer. Any suggestions/recommendations are appreciated, Lennard Fuller
