Why the Semantic Web Marketing Message Has Failed
So some guy writes why the semantic web will fail and ends up on Slashdot. How slashdot picks their articles, I'll never know. The article is pure opinion and guesswork (as all predictions seem to be), and it's perfectly OK for this guy to blog his opinions.
I'm not going to argue that the semantic web (that's *small s* semantic) will succeed, although I think it will prove useful in a large sense in some form, even if that form isn't RDF. I think what's really telling about the doom and gloom post is that the marketing message of the semantic web has failed.
For example, a quote from the blog post:
> The Semantic Web will never work because it depends on businesses working together, on them cooperating
Where, in all of the W3C's semantic web literature does it says that companies must work together for the semantic web to succeed? I think this is one of the biggest misinterpretations about the semantic web. For some reason, people think that the semantic web requires these large agreed upon ontologies before anything useful happens. Not only is that near impossible (for anything but the most generic or free form terms and definitions) but as we all know, specifications born out of committee have an awfully hard time meeting the pragmatic needs of the masses.
For the semantic web to succeed, the W3C doesn't need more technical specifications (although a new RDF XML serialization would be nice). Instead, the W3C needs to completely revamp its marketing message. For instance, distance the semantic web from AI. AI, no matter how promising, leaves a bad taste in your mouth. We need to completely deny any relationship to AI. Secondly, the W3C needs to rebrand the semantic web as "Simply Putting Your Database On The Web. No More, No Less. Anything Else Is Purely Serendipity." Thirdly, the W3C needs to really drive home that the semantic web will succeed *only if* it is not built with large top down ontologies.
So repeat after me: "The semantic web is just an effort to help expose the database that you already have to the web as RDF. Primary keys become URIs, and the intersection of a row and a column is a triple."
Or, to put it another way:
Problem: I have data, most likely in a relational database, that I need to get on the Web.
Solution: Expose that data as RDF. URIs are the primary keys for the data.
I'm not going to argue that the semantic web (that's *small s* semantic) will succeed, although I think it will prove useful in a large sense in some form, even if that form isn't RDF. I think what's really telling about the doom and gloom post is that the marketing message of the semantic web has failed.
For example, a quote from the blog post:
> The Semantic Web will never work because it depends on businesses working together, on them cooperating
Where, in all of the W3C's semantic web literature does it says that companies must work together for the semantic web to succeed? I think this is one of the biggest misinterpretations about the semantic web. For some reason, people think that the semantic web requires these large agreed upon ontologies before anything useful happens. Not only is that near impossible (for anything but the most generic or free form terms and definitions) but as we all know, specifications born out of committee have an awfully hard time meeting the pragmatic needs of the masses.
For the semantic web to succeed, the W3C doesn't need more technical specifications (although a new RDF XML serialization would be nice). Instead, the W3C needs to completely revamp its marketing message. For instance, distance the semantic web from AI. AI, no matter how promising, leaves a bad taste in your mouth. We need to completely deny any relationship to AI. Secondly, the W3C needs to rebrand the semantic web as "Simply Putting Your Database On The Web. No More, No Less. Anything Else Is Purely Serendipity." Thirdly, the W3C needs to really drive home that the semantic web will succeed *only if* it is not built with large top down ontologies.
So repeat after me: "The semantic web is just an effort to help expose the database that you already have to the web as RDF. Primary keys become URIs, and the intersection of a row and a column is a triple."
Or, to put it another way:
Problem: I have data, most likely in a relational database, that I need to get on the Web.
Solution: Expose that data as RDF. URIs are the primary keys for the data.