I wrote two articles about my companies semantic web implementation: part 1 and part 2. I was really great for me personally to bring this knowledge out in the open, because I really felt we made a contribution to the field here and it would be a shame if it would stay confined within the walls of a company. The first article has had only a thousand or so pageviews, which isn't much. But for me that isn't important. I am confident that it will be found one day by a ambitious developer, who will use this knowledge to build an open source system based upon it (among others).

Just before I started working for Procurios I had a job interview at Backbase (the "Ajax company"). I had some good conversations with Gerbert Kaandorp, the CTO. He might have hired me, despite my lack of CSS knowledge at the time (not knowing simple stuff like the difference between "display:block" and "display:inline"), if it wasn't for a project manager who decided that I really wasn't up to speed for the job (when Gerbert asked me if I knew about design patterns, I told him that I used them when they were useful (which was ok by him); the project manager asked me which patterns I had actually used lately. I said "observer", and "singleton", which didn't impress him at all. But that was not my fault: the projects I worked on just didn't require any other patterns. And you certainly don't want to use any design patterns just to impress your next employer...). Anyways, Gerbert asked me at one point if I had heard of the semantic web and if I had done something to support it. I said I hadn't. And up to that point I had not considered persuing semantic web stuff actively. My attitude had been more like "I don't really understand this RDF stuff, I will wait until it gets bigger". Gerbert made me change this.

There is more stuff that you come across and can't use immediately, but at Procurios I found a good use for semantic web technology. In my own spare time I had written a SPARQL interpreter in PHP. It isn't released because better ones are available. At a later time, I had a bunch of requirements for the rewrite of the document management system, which lead me to design a complex and powerful new system. However, it was not until Pieter Schiks pointed me to the fact that the Dutch governement was starting to use the Dublin Core specification (related to semantic web) more and more that the solution snapped into place. Currently I really believe that relational databases will be phased out at some point. But until then, some hard struggles will have to be fought. It really is a historic era for knowledge modelling.