This could be a giant leap forward for the web... toward a useful giant global graph.
Common Tag, which was released yesterday, is a logical extension of Linked Data. In a nutshell, it offers an accessible and open standard to incorporate semantic tags in web content. It is supported by a range of up-and-coming semantic players, the most notable among them being Yahoo SearchMonkey.
That, in itself, is a solid step forward. Its simple set of standards adds the necessary muscle to the RDFa skeleton. Common Tag ties tagged concepts to URIs, defined addresses centralized in a handful of trusted repositories, yet even advocates straightforward cut-and-paste to tag HTML. Simple and powerful.
This is no technological breakthrough, all the pieces were already present before, but just like coins without a banking system, RDFa and Linked Data needed an easy-to-understand and easy-to-communicate mechanism to add semantics to content. Now we have it.
But wait, there is more.
And in my opinion, that could be web-changing.
The common user is about to be empowered with an easy tool to add semantics to their content, because one of the companies behind Common Tags, by outputting Linked Data in a standardized format, is solving the last problem standing between us and a more semantic web: the user interface.
Building on top of its friendly integration with mainstream web publishing tools, Zemanta, the "publishing assistant" application, is now making it possible for users to integrate semantic tags quickly and easily into their content. And, if I understood that correctly, it is the only company behind Common Tag to do so. If not the only one, at least the user-friendliest. Smart.
To do that, Zemanta reads your content, and makes recommendations of common tags for different concepts. In other words, it transforms unstructured data into structured semantic data that's quite precisely defined and linked out at the concept level. Quite precisely, because it doesn't rely just on algorithms, or just on humans, but on both. Don't be fooled: quite precisely is a huge improvement over existing solutions... As a human, I can now afford to take some time to add precision, because a lot of the groundwork to identify relevant tags and their addresses has been done for me.
Today, it works for blogs (except Wordpress as it strips out RDFa, but they'll have to change that...) and emails (I'm just guessing about that last one since Zemanta does email too). Tomorrow... mmm, let me guess... it will work on webpages and other content too.
No surprise Andraz Tori of Zemanta is the driving force behind this initiative...
Coupling its easy-to-use interface with Common Tags, Zemanta makes semantic data finally accessible to non-programming folks, i.e. most of the web publishing world. Which ultimately means, if it continues to execute as nicely as it has in the past, a lot of web 2.0 users. World domination is near. Think "Enhancing the world's knowledge".
Two things stand in the way. First, competition. OpenCalais so far has proved to be much less mainstream-friendly (sorry to my friends there, but I think they know that - they also focus much more on the B2B market as their recent deal with CNET shows), but they could come up with a more accessible interface for tagging webpages with what I'm just conjecturing might be more precise tag recommendations algorithms. Joe the web user would adopt that to enhance his web content. Many other application developers will target this new market.
Second, the system gives an incentive to humans to tie to the right tags: being found. But that's a two-side carrot. Remember meta tags? What makes Common Tags any different? The URI linking is no quality control.
Let me ask that: in your opinion will Common Tags survive the test of system-gaming by users using the wrong tags for their content? Could it work if a third-party application was the one tagging your content, using some CSS-type tags instead of embedded tags (nothing prevents Common Tag CSS)? It could do it as well through a mix of both algorithms and human inputs.
Thanks to a tag I added to an obscure post I made some time ago (I know, that sounds as credible as a Bill Gates ad for open source), I can make the bold claim of having long been calling for a "semantization" engine to help us easily add meaning to existing and new content (I even bought the domain semantized.com in case I was to launch that myself, you know, one day...) Turns out I was wrong. But not by much: what we got today, which can change the web as we know it, ladies and gentlemen, is a zemantization engine.
![Reblog this post [with Zemanta]](http://img.zemanta.com/reblog_e.png?x-id=72e7e102-96ae-4e2f-81be-08fd125aedf4)
