I was very happy to stumble upon the XML-RPC interface which is available for Advogato, as I generally hate to use a web browser for things that I feel can be done better programatically (like the functionality on this site). I therefore began working on a Java program to interface with this site, using the XML-RPC interface.

The Apache XML-RPC libraries for Java seem quite nice so far, and I tested out all the methods in a hard coded way already to see how well they work. Due to the speed of Advogato, and the type of XML-RPC calls that were implemented, I think I will need to do a great deal of caching to make the program work well, and even with that in place there is a lot of site functionality that isn't available over XML-RPC

Since this site is certainly not add-driven, I would like to propose / suggest some additions / modifications to the XML-RPC interface of Advogato to allow for more robust / pleasant to look at clients to be written against it's functionality. Then again, feel free to blow me off as well :)

I suggest the following additions to the Advogato XML-RPC interface =>
1) As with the diary, methods to get / post articles.
2) Bulk methods, to get a number (X) entries / articles, rather than the one at a time current implementation.
3) A way to interact with the certification system.
4) A way to access the "recent posts" functionality, which I must say is most excellent.

And now relating to the topic of this diary entry, the XML-RPC implementation here got me into "research mode", and I wanted to see what other sites / services on the net had these (or SOAP) interfaces to their functionality (other than Google, which I had heard about weeks ago, but they have that "search limit", which it probably necessary from a resource standpoint, but rather bogus from a programmatic standpoint.

I looked through the IBM UDDI directory, and also another SOAP / XML-RPC service registration site, but over 50% of the stuff in there was complete garbage, and of the stuff that was there, I didn't find anything at all that was interesting enough to spend my time implementing a client that worked against it.

</ramble>