Is the semantic web a realistic goal?
12th May 2008 by David North
The vision of the semantic web was that of Tim Berners-Lee’s - to ensure it was a universal medium for data, information and knowledge.
A noble idea, I’m just not sure this is going to be a realistic goal due to human nature’s tendency to try and exploit technologies for their own ends.
Certainly if this goal is achievable now is the right time for it to take off. There is a buzz online with regard to various technologies that can be used to achieve it (including microformats) with many proponents to semantic web cause.
The issue is that all these methods rely on correct and honest use online. In it’s simplest form the semantic web attempted to use meta tags in years go by. These became a well known method for black hat search engine optimisers to try and shoe horn inappropriate keywords into pages to try and fool the search engines. The result? Search engines generally ignore meta tags because of this easy abuse.
So is this also going to be the case for the new drive for a universal information medium? Not currently maybe. As it stands the uptake isn’t large enough so there doesn’t seem to be any advantage in trying to provide inaccurate data. But just wait until the tipping point is reached I’m sure that search engines will try to lean on this extra layer to improve results. Some people will be waiting to exploit it.
The only thing we can hope is that the lessons constantly being learnt from the constant battle against the black hats can be applied against them again to ensure this new semantic vision isn’t muddied. I also feel that social media is going to become even more of a watchdog for content quality - those millions of eyes put to good use. The advantages of the semantic web can’t be ignored - machines understanding human content opens so many doors.
As usual though the main issue here is getting people to use the technologies. From a business perspective people aren’t going to adopt this without it providing some kind of return on investment. Implementation is simple in most cases so I can perhaps see people dipping their toes in the water and this gradual adoption may start the tide. I for one will be looking at the hCalendar, hCard and XFN formats as a starting point.
I think the key will always be in the hands of the human. Why not allow a simple thumbs up, thumbs down and cross system (thumbs up = you like the result, thumbs down = the result wasn’t any good, cross (or x) = the result was spam, offensive or irrelevant) that relies on a specific number of votes before action is taken? The more thumbs up from independent sources, the more trusted that metadata provider becomes, and if you hit the X that result is dropped and/or the provider is penalised.
The more results we get back from users, the more intelligent the systems can be. Unfortunately they require intelligent human input but have to be savvy enough to be able to wheedle out fraudulent votes.
I still don’t know why I can’t just click a link within Google to stop experts-exchange.com from being listed, at least for my own profile.