Sunday, May 13, 2007

User powered content and SemWeb

As it usually happens with a emergent long term concepts, people tend to go overboard and attempt to boil the ocean and lose touch with reality. Semantic web in its current direction appears to be laden with fantastic descriptions but offers little incentives for business to agree on one standard.

Obviously, all new technologies take time to mature and get adopted by businesses but then one successful adoption comes with hundred other candidates that died their natural death in the lab. Let's see what issues might prevent the concept of web applications using other web application and create a fantastic meaningful web.

Inherent problem with user powered content

How a user contributed content is described in terms of its metadata has a very significant role to play in SemWeb. Metadata is what structured search engines would use to locate relevant information for SemWeb. Unfortunately, there are inherent problems [Metacrap].

People are lazy
Let's face it, most of us are lazy. Effective SemWeb is only possible if a complicated set of metadata in terms of taxonomy or conceptual relationship to one another is described correctly. Creating and maintaining these schemas or adapting a pre-existing one is no trivial task. Solution is not to open this to people because there is no incentive for a user to take the pain of going beyond a couple of clicks.

People Lie
It's no wonder that your inbox is inundated with spam that reads the subject line as "Here is the information you requested". It is equally no wonder that your web search for a "enterprise application integration" might result in hundreds of press releases from companies which have nothing to do with EAI.

Lying exists in real world to compete for business as much as in the online world. How do we ensure that metadata is reliable? This is one thing where democracy does not really work.

People don't agree on terminology
You can't force people to describe something your way. I might like to call a porn, an erotic thriller and I can't stop you to call a nude beach, a naturalist shorefront. Your wish!
However, this creates a big problem for SemWeb, because computers cannot figure this out reliably.


No comments: