Over the last couple of years I have been fortunate to attend Semantic technology and life Science conferences. I really enjoy chatting to people and finding out what they are doing and sharing experiences. I've been particularly interested by several discussions with people who have tried Semantic technology projects and failed. I never really understood why these projects failed, but many of these people put it down to various aspects of Semantic technologies such as immaturity, difficulty with this and that etc. It wasn't until I spoke to people that had completed a successful semantic project, after several failures, and merge that with some of my own experiences that I started to understand. One individual told me quite bluntly that it wasn't until individuals with semantic web experience were brought into a project that success stories started to happen. It wasn't through lack of anything, enthusiasm, brain power etc, but it seemed that existing technology teams just couldn't make semantic technology work for them. The result was often that semantic technology was thrown out and disparaged.
It happened again today!
Sometime ago a software company presented a project they were about to embark on. It looked great they had thought about a flexible system and were going to use RDF. Today we had the update and initial demo of the software. It didn't start well for me as the technology slides had no mention of RDF and semantic technologies. Instead they had spent ages creating a standard architectural service based - RDBMS system. It got worse when they spoke about how flexible their system was. I started to probe with a few questions, "How many concept types can you deal with?" ... "currently four, but we are looking to add more" they said... "but of course I can create my own concepts can't I?" ... "errr no, that would be through consulting to customise the system".... "but I can create my own attributes and specify how the concepts interact can't I?"... "err no, but we are looking to add that later"..... "OK so why is concept x not a subtype of concept y?" .... "That's because we brought that concept into the project a bit later and it was difficult to add it without a lot of work".
I started to suggest that if semantic technology had been used the concepts could have been created as ontologies which could change on the fly as could relationships and the attributes around them and we would not be restricted by having to recreating schemas in the relational back end to cope if RDF had been used. The explanation came back that when they looked into using semantic technology they had little success, blaming this and that and stating it was immature lacking x, y and z etc. It became clear that they had not understood the fundamentals of the semantic web and were trying to fit semantic technology into a predefined java, OO world and had not really understood that you have to think and design a bit differently. I must say my blood did boil a little and I was a bit rude, which I should not have done, it was unprofessional of me so I do apologise.
Don't get me wrong, I'm not having a go at anyone; this is just my view on things as I have seen and heard them. But, if it is true that we have to get individuals with semantic experience in on these projects for them to become a true success, then how are we going to do it. These companies don't seem to know they need them, saying ..."we don't need semantic technology people as the technology isn't up to much. If our guys can't make it work it's no good". There isn't a huge glut of semantic techies out there for them to choose from even if they did realise.
It's a tricky one.