GSoC 2008: Distributed Search for eXist

Distributed Search for eXist (GSoC 2008) Hi everybody, I would like to introduce myself and give an overview for one of two accepted Google Summer of Code projects for eXist this year (http://code.google.com/soc/2008/exist/about.html). I'm postgraduate student, attending Master programme of Computer and Information Science at Faculty of Electrical Engineering and Computer Science in Maribor, Slovenia. My interests and courses are in multimedia (and) databases. My diploma (degree) work last year would not be possible without eXist, using two other open source tools for practical "semantic" application. I have 12 years of working experience in network security, system engineering, web mastering and application development. Several open source projects were translated to my native language by small team or me only (most known are Drupal and Claroline) and now a chance to participate in open source code development is both a challenge and a (hopefully long) commitment. Distributed search project abstract (to start discussion) is here: http://code.google.com/soc/2008/exist/appinfo.html?csaid=B4609CD935485A3 Shortly: eXist's XQuery expression engine can fetch XML documents from local database instance (server) and from remote URLs but it would be good to "spawn" search to several - local or remote - database instances at the same time and merge already parsed retrieved results to final result tree before sending it to client. So eXist would not retrieve XML document from remote URL first and then parse&merge etc. In this context let me state two things more. Distributed databases are databases which have parts of the data located in several database instances. Imagine adding additional server for better reliability, scalability, upgrading etc. Distributed search facility must find all results from distributed database (having same schema) but we should also think on different databases. The other thing is that W3C hasn't recommended yet how this "execute" XQuery expression should be done. Future XQuery 1.1 proposals mention only referencing nodes which will be done practically during this project (so for example we can use reference and get further data from its database instance where it resides) while most academic papers declare new, non-standard XQuery language construct. Since parsing URLs in XQuery expression can distinguish remote and local URLs such construct is really not needed so the aim here is to add proper code logic to efficiently retrieve and combine query results. Thanks for reading, more when I start working with my mentor (Leif-Jöran Olsson) and other eXist developers, provide requested forms to Google and - continue to read great community feedback (now nearly one year for me) in this list! Sergej M. Rinc