Last week the 2009 Semantic Technology Conference was held in San Jose. People from various industries and organizational specialties came together to learn about the latest happenings around semantic solutions and how their firms might be able to benefit.
From knowledge base and support management to systems analysis, very large enterprises like Boeing, the government, Time Inc, Intuit, etc had people attending and/or presenting. There seemed to be a few packaged and well defined solutions in the medical/biology space but a lot of the sector seems to have somewhat of an art form angle to it that requires one to get very deep into the technologies/standards available to work towards a solution.
Ontology, triple stores, jena, SPARQL, terms of the trade. One thing the sector needs to continue to work on is the education of really specific use cases and solutions that enterprises can take advantage of. As someone who has been a developer in the web and mobile space for the past decade and been involved in 100 some-odd CRM implementations I initially found myself attending presentations and feeling like an un-educated buffoon trying to grasp why someone would adopt a semantic style approach to solving their problems instead of traditional means.
There is definitely a “kool-aid” type effect around the sector where it seemed sometimes that everyone at the conference viewed semantics as the gospel and 
only true way to tackle problems. The most common argument I heard from presenters and attendees was the argument against traditional relational database solutions for defining and solving problems where many-to-many relationships ruled. One presenter gave his argument why ontology based solutions are superior and proceeded to show an example application that handled a purchasing and shopping cart scenario one solution that is perfectly suited for relational databases.
I went up to the presenter afterwards and asked him why he used such an example to argue against relational solutions using Oracle/MySQL/etc. He answered that it wasn’t the best example and that there are many other uses. This was a theme that seemed to be continually repeated.
As the conference went on and I saw a couple cool demos at the booths and had some great conversations and I started to feel the kool-aid effect a bit myself. A bay area firm named Franz was showing off their AllegroGraph triple store and some interesting features they have built into the latest 3.2 version inspired by 
some of the research and projects they have been conducting with various government and public interest groups. Also on hand was Powerset under the Microsoft Bing brand which had a lot of interested people stopping by the Bing booth as well as Thomson Reuters and their fantastic Calais service.
Two awesome highlights from the event were two of the keynotes. One from a startup called Siri that launched at the AllThingsD Conference. These guys have a 
very slick solution for only being at an alpha stage of development. They are definitely on track for nailing the user experience side to a digital personal assistant. The second great keynote feature Angle contributor Nova Spivak who
did a one-on-one interview with Russell Foltz-Smith from WolframAlpha. Having only played around minimally with the service it was eye opening here Foltz-Smith talk about their up and coming API release as well as future paid pro type accounts and future offerings around targeted advertising and premise base solutions.
After sipping from the fire-hose of semantic info the conference conference provided my mind was racing through the weekend re-architecting past solutions into ways that might be better suited for ontology/graph stores…, a few more weeks and I might be a full convert from the relational db camp…, probably not but I am starting to see how data portability is one of the great problems this technology can solve. I look forward to seeing the advances the industry makes through 2010.
