Wayne D's Blog

While driving to Ottawa over the Christmas break I took the time to get caught up on Buzz Out Loud. In one episode Molly Wood opined on how cool it would be when the semantic web was finally working and she could get her semantic web feed populated with funny cat pictures. I wanted to reference the particular episode but other than listening the six or so I had stored up all over again I will cross my fingers and hope I have got the reference close to the ballpark.

There was much talk this year about the semantic web being the "Next Big Thing". At its core the semantic web is about making web data easier for computers to parse and determine relationships which in the end is supposed to make it easier for you and I to find what we are looking for on the web. There are two reasons this technology is doomed to fail, firstly because people are lazy and secondly because some people are dishonest.

When I write my blog post there isn't much chance that I will the spend more time creating a semantic web description for the content it contains. I thought the whole point of computers was to make my life easier? Why should I have to write the thing twice? What about video and audio podcasts? After I edit them do I then have to write an rdf file in addition to the show notes so a computer can figure out what's inside?

The preponderance of link farms, keyword spam and tag abuse shows that if there is a way to game the system someone will try to use it. Google is in a running pitched battle to keep their search results relevant because some people are always trying to improve their result placing because they know if they aren't on the front page of results they are in trouble. The history of the web is full of people using metadata in a way that was not intended by the standards writers. The semantic web is just another form of metadata. Under carefully controlled circumstances metadata can be quite useful but in the wild west of the world wide web its pretty much useless. Imagine sending your semantic web retreiver engine out to find funny cat pictures and getting results littered with cat food ads as an example?

Cuil.com promised that it's Semantic web technology would out google Google and we know how that turned. I haven't yet seen a website advertising semantic web technology that has impressed me. I've written a couple of times about Twine.com and though I use it as a bookmark sharing tool the recommendations engine which is supposed to be powered by semantic web technologies has never recommended anything I am remotely interested in. This is not surprising because the descriptive metadata for the web resources that Cuil.com and Twine.com are using is nonexistent. They have to parse the web page and figure it out on their own. That begs the question, other than using semantic as a buzzword what do these two or any of the other purported semantic web driven website mean when they say semantic web? My suspicion is that it has become a hollow meaningless marketing term. The semantic web was last year's big thing, I wonder what it's going to be this year.