A longer while ago I mused into what direction my online photo collection should grow. I have been under the impression that faceted classification plus a nifty facet map visualization engine plus a nifty geo- and mapping engine could solve my problem. Tests in XFML, XUL, Graphviz and other mechanism produced some interesting results but also seeded doubts if a system of the envisioned magnitude is feasible at all. Then, on-the-fly categorization systems, called tagging, surfaced in collaborative too,s like del.ici.ous and flickr. As they seem to work astonishingly well, its time to figure out what of them could make it into my own system.
Addings tags to content items (photos, press excertps, web links) and providing defined metadata queries for structured access to the collection is what one would do both with a simple tagging system and with XFML (what else is it than tagging using a controlled vocabulary?). Given that building a tag query system is way easier than building a XFML query engine (of which none really exists in open source world so far), software complexity could be reduced significantly. Then again, a RDF-based system would be the real thing. Future compatible, and ready for a nifty XUL interface - see Dan Brickley's RDF for self-describing images, An Indexing and Querying System for Online Images and related papers.
Comes in SVG, the holy grail of open source fetish. Unfortunately, I am under the impression that SVG can be really useful for building interactive maps from a XML datasource, but who has an SVG plugin anyway? Automagically building graphs for facet/topic based navigation with Graphviz is interesting, too. Oh so many are the opportunities - until I see land, enjoy the occasional manual addition of a hand build collection of hand edited photos to my collection and to the One for the day section of this blog. If you don't know what XFML is, stay tuned. I've prepared a longer intro article on it before, which for some reason didn't find it into the blog so far, but will so now.
