Hey, I'm working on a RESTful EntityStore for Qi4j, and it is going reasonably well. So far I've been using Apache Abdera (Atom server/client) to do it, and it works, but I'm considering throwing it out and doing all of it by hand for a couple of reasons: 1) Abdera brings in a ton of dependencies 2) Making the EntityStore look like an Atom feed is cool, but ultimately perhaps not that useful 3) There's a lot of overhead when I really want to just parse the headers manually for cache-optimizations, and then post binary data back to the client 4) Abdera insists on using URL's which are lower-case, which makes it really difficult to include QualifiedIdentifiers on the path.
So, I'm going to try and do all of it on my own instead, and see how that works. But so far it seems like the general idea is correct, which is that by using HTTP I get the transport and all that for free, and doing caching the HTTP way is really neat (using last modified checks). With clients that can store state locally it is going to be superfast, since most requests will say "not changed since last check". /Rickard _______________________________________________ qi4j-dev mailing list [email protected] http://lists.ops4j.org/mailman/listinfo/qi4j-dev
