Where are we with the semantic web, that ambitious project to create ways for static HTML pages to interact better and produce a richer user experience? There are three answers to that question:

Pie in the sky

The problem is that many people – especially those who wanted to talk to IWR on the topic – are firm adherents of the first answer to the semantic web question: the semantic web will never work because it can’t.

Linguist, consultant and author David Crystal says: “Computer scientists have broadened the concept of semantics in far too broad a way. They have taken an encyclopaedic definition of meaning, by which, for example, anyone can call themselves a French speaker if they can use words like Paris or Marseilles, when, of course, actually knowing French is far more complicated.”

Crystal – and others – argue that the semantic web is trying to deal with philosophical problems to do with meaning and language whose very intractability condemn it to an early death.

“In theory it’s great but practicality is the issue,” says Thomas Herbert of hosting firm Hostway. “Patently, we can’t go on forever having nothing better than keyword-based searches, but we can’t even agree on one definition of what a ‘business’ is. And even if we all managed to agree in the English-speaking world, what about what ‘business’ means to a Japanese speaker, and so on? If it does happen, it’ll be a lot slower than you think.”

In short, although it would be nice – and probably very handy too – to have some kind of universal set of definitions of things like business and so forth to structure the use of information on the web courtesy of the semantic web project, we’ll just never get there.

The problem, though, is that Berners-Lee and the rest never said they were going to give us such a thing in the first place.

Seven and a half years ago the article that kicked it all off said: “The semantic web will bring structure to the meaningful content of web pages, creating an environment where software agents roaming from page to page can readily carry out sophisticated tasks for users... without needing artificial intelligence on the scale of 2001’s HAL or Star Wars’ C-3PO.”

In other words, the harshest semantic web critics are lambasting a fake target.

Well, somewhat fake. As Ivan Herman, semantic web activity lead at the World Wide Web Consortium (W3C), the group co-ordinating the drive to create the standards that will power the eventual thing, admits: “Any argument that equates the semantic web with the use of big ontologies is wrong. To be fair, the community has made lots of messaging mistakes, but that’s not what we are trying to do.”

Herman argues that from the outset the project was never about creating an intelligent web but a web of data.

He adds: “While it is entirely legitimate to use complex taxonomies in c ertain restricted areas to try to agree a common vocabulary, what the project as a whole is about is getting all the various types of data on the web in their many zillions of different standards to be able to talk to each other – just like the basic HTML document pages do already.”

And that universal connectivity – supplied by Berners-Lee when he created the HTML format in the 1980s – was the key to the web’s attractiveness in the first place.

So reports of the semantic web’s demise may be premature. It certainly seems unfair to be so critical about something the project is not even attempting to accomplish (although, as Herman notes, this is also partly down to poor communication of the semantic web’s goals).

Good, but not good enough

OK, so what of our second answer: that the semantic web is fine as far as it goes but isn’t much use to business or the real-word problems of the average corporate user?

That argument chimes strongly with Andrew Yates, CEO and founder of UK company Artesian, which claims its software is firmly based on semantic web standards, but doesn’t want to say so.

He tells IWR: “If you looked at our website 10 months ago, you’d have seen lots about the semantic web. But we’ve taken all that off. We quickly realised that business isn’t at all interested in any of that; it’s only interested in business productivity.”

Lest you think this shyness – the search that dare not speak its name – about semantic web affiliation is just cautious marketing, Yates adds: “The semantic web is starting to become useful and interesting but it’s just not yet useful or interesting enough for business to take seriously – certainly not for any organisation with many ‘moving parts’ that has to deal with lots of variables.”

Similarly, Elsevier’s Illumin8 is a heavy-duty web page and content semantic search tool launched last May that contains five years’ research and can do lo ts of semantic web type things, but it hasn’t a scintilla of any of that hippie nonsense in it.

Product manager Joe Buzzanga says: “Our approach is consistent with the goals of the semantic web but we have chosen to use a natural language processing-based approach, not any W3C standards right now, as we don’t think that’s the best way of us delivering value for our customers. While we are monitoring the W3C work and see this as a step on the way to that, our corporate R&D users need a solution that is reliable and highly scalable today in a way that it just isn’t yet.”

And finally on this note, there is this from Kal Ahmed, founder of NetworkedPlanet, which takes a totally different approach again to solving semantic web-style problems with its ISO Topic Minds framework: “The semantic web is about making machines talk to each other better, but right now we need to make it easier for human beings to be empowered to make better decisions. It’s all too rocket cars and tinfoil for today’s corporates, basically.”

According to this argument, the semantic web – if it works at all – is still very much a consumer thing and not for People Like Us.

Indeed, enterprise content management company Vignette says we can get what we want in the business space by using Web 2.0, an already established consumer web technology.

Its European senior product marketing manager Guy Westlake says: “We have a product, Vignette Recommendations, that is helping organisations like Nasa find relevant content and make useful links between different sources.”

This view of the semantic web project is fairer than the first critique, but it still has some problems, say commentators.

Paul Miller, semantic web evangelist for library systems supplier Talis, says: “Semantic web has a lot of academic baggage but 2008 has seen a lot of things come together.”

Miller and others point to the final coming on stream of three key building blocks for the data web that Herman and the W3C are on about:

“Sparql is the last piece in the jigsaw to finally let developers and companies start unlocking the value in all those static HTML pages out there,” Miller, as an evangelist should, enthuses.

The theme is echoed, not too surprisingly perhaps, by the W3C’s Herman: “Now we can really start to look at the web as a means to start linking all sorts of information in increasingly useful ways.”

Behold the future

This takes us to the third answer to our question: the semantic web as the next dotcom boom. Quite rightly, klaxon warnings are sounding very loud right now, but let’s just take it a bit further. A slew of startups have kicked into life in the last year or so. Like Artesian, they may not emphatically label themselves semantic web companies, but they are very much using that OWL-RDF-Sparql constellation of standards to get themselves going.

Mostly, these businesses are consumer-facing – a valid enough label if you hold that the only “real” technologies come from that space, unlike fripperies such as, er, the web, Web 2.0, etc. And these are very early days for some of them. But then heartland information companies such as Thomson Reuters and Dow Jones are also very serious about looking into all this.

Could semantic web be a new disruptive technology and not just a mad philosophical idea? One startup, San Francisco-based Twine, thinks so; its confidence probably has something to do with attracting no less than $26m (£17m) worth of venture capital in the last 18 months. Its director of marketing, Candice Nobles, says: “Semantic web is very much alive and kicking, and we are just one of a new breed of consumer-facing companies putting applications out there based on it.”

So which of our three answers is closest to the truth?

The simple fact is that mis-sold, misunderstood and probably mis-targeted as it has been, semantic web may well be what gives us Web 3.0 – a fact about which no information professional can afford to have out-of-date views.

WHO'S ALREADY DOING IT?

MORE ON THE SEMANTIC WEB

Official W3C pages, including a very useful FAQ and over 30 case studies of companies using what is already available, at:

Wikipedia provides not just a good overview of the project but clear definitions of the various components such as OWL:

The original May 2001 Scientific American article by Tim Berners-Lee (and others) that introduced the concept.

A fun YouTube video explaining the basics.

Recent BBC interview with Berners-Lee.

Other useful resources:

Nova Spivack at The Next Web Conference 2008
ZDNet's the semantic web

Cloud of Data blog

Talis's Semantic Web Gang blog

Nodalis blog