Web 2.0 and Semantic Web: Mars and Venus?

Web 2.0 and Semantic Web: Mars and Venus?

Summary: Kendall Clark from XML.com posted an interesting article called Web 2.

TOPICS: Enterprise 2.0

Kendall Clark from XML.com posted an interesting article called Web 2.0 Meet the Semantic Web. In it he talks about a new technology called SPARQL, which is an RDF query language and protocol. It's apparently an SQL for the Semantic Web. Kendall's post proposes that Web 2.0 be "wrapped, morphed, or bridged on to RDF", to use Shelley Power's words (kindly translating from SemWeb geek into 2.0 geek). Kendall wrote in summary:

"Well, Web 2.0 fans, builders, and advocates need more love from SW fans, builders, and advocates. These two worlds really belong together." 

I'm supportive of the notion to bring the Semantic Web folks and the Web 2.0 folks into the same orbit - anything that brings Tim Berners-Lee's Semantic Web vision for the WWW closer to the vision of Web 2.0 that I follow, is OK by me. But I feel I must point out the two worlds are not necessarily the same. I once crudely put it like this: Web 2.0 is 'The Web as Platform' and the Semantic Web is 'The Web of Meaning'. The platform could use more meaning, or it could continue to do just fine with the same 'good enough' approach that characterised the evolution of the Web in the 90's.

Morphing the two worlds has been tried before and wasn't overly successful. RSS 1.0 is basically an RDF version of RSS. It still has a lot of proponents, but its simpler cousin RSS 2.0 was always far more popular - and remains so to this day. Web 2.0 is all about keeping it simple - RSS, RESTful APIs, permalinks, etc.

So how would the complex (compared to RSS anyway), but powerful, RDF fit into the Web 2.0 ecosystem? Danny Ayers is a Semantic Web expert from way back and he seems to think the two worlds can work together:

"...this network wants to join together - the majority of local stimuli are pushing in that direction. There are many points of intersection, relatively few difficult barriers."

Correct me if I'm wrong, but Danny's basically saying to take the organic approach and see how the Web evolves. I agree with that.

Finally, Dave Winer responded to Kendall's post with this:

"The Web is real. The Semantic Web is an idea and Web 2.0 is a marketing concept used by venture capitalists and conference promoters to try to call another bubble into existence.

The hype is treating "Web 2.0" as more and more real, and the hypesters are getting further and further out on a limb. "

Dave has a point. Web 2.0 is becoming very hyped and the Semantic Web has never got off the ground. But in both Web 2.0 and the Semantic Web, there are very real and viable technologies that are defining the next evolution of the Web. People are using the technologies, whatever you want to name the overall trend. 

For some reason I'm reminded of that old Palmolive advert, with Madge saying "You're soaking in it!" It doesn't matter too much what people think of the term 'Web 2.0', the fact of the matter is: we're living through the middle of it right now.

Topic: Enterprise 2.0

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.


Log in or register to join the discussion
  • 3rd Planet

    Good post, and thanks for the ref. Your "Web as Platform"/"Web of Meaning" is probably as good a summary as there is. But note that it's meaning in the machine-readable sense, so an alternative might be "Web of Data", in contrast to the Web 1.0 "Web of Documents".

    The platform could continue along with 'good enough' approaches, but personally I think the increased net-wide NxN complexity that lies down that path will hamper progress. Making more of the Web's data available at a common shareable level strike me as a way of accelerating improvements.

    "Danny's basically saying to take the organic approach and see how the Web evolves" - yes (not that we have much choice). I'd add that despite misconceptions that one or two commentators have had, Semantic Web technologies are eminently suited to organic development, in fact they've been designed in such a way as to support distributed evolution consistent with the existing Web.

    There's a little information lacking in both Dave's "The Semantic Web is an idea" and your "the Semantic Web has never got off the ground". There is a difference between the technologies and the idea they facilitate. The vision is pretty well indistinguishable from what most people have in mind for Web 2.0. The base technologies - URIs for naming, XML for syntax, the HTTP protocol, the RDF model, are relatively mature. There are toolkits for virtually every programming language and environment. But getting off the ground in this context really means widespread deployment, and in that sense I think it would be valid to counter with "Web 2.0 has never got off the ground".

    Sure, the RDF/XML syntax version of RSS hasn't been as popular as the plain version, but it's worth noting that RSS 2.0 can still be interpreted in the RDF model, as can Atom, microformats and virtually ever other kind of data we see on the Web today. The data contained in the world's relational databases could be expressed as RDF, not something that would be much use simply in RSS (despite Adam Bosworth's hand-waving). There has been significant industry adoption of RDF, but much of it isn't as visible as RSS. For example, every image/document created by Adobe's tools (Photoshop etc) now contains embedded RDF. FireFox/Mozilla has RDF under the hood. There's big adoption in the life sciences and curious alcoves like the intelligence community. Oracle recently added RDF support to their database, making it available to the top end of enterprise. So although the Semantic Web may not have got off the ground, it is taxiying along the runway at high speed.

    Going back to Kendall's post, SPARQL really is a nifty piece of work! It allows you to do many tasks reasonably familiar from the SQL and XML worlds, but in a totally (Semantic) Web-friendly fashion. It's relatively easy to understand being more immediate than working with RDF programmatically. An RDF datastore with SPARQL is an exceptionally agile development data setup. I don't know whether we'll see the kind of ubiquity of the SPARQL protocol Kendall hints at in the near future, but because SPARQL lowers the barrier to adoption of Semantic Web technologies I do think we will see significant growth of their use around the grassroots.

    On your final point: "You're soaking in it!". Absolutely.
  • Semantic albatross

    The problem with (web) data is the absence of metadata. In order to get meaningful searches in a "semantic web", you MUST provide a "survey" of questions ABOUT your data, to be published WITH your data (i.e. metadata). Needless to say, this would be VERY tiring/tedious to do if you published a lot of data! IOW no one will do it so "semantic web" searches will be just as hit-or-miss as any other searches.

    And DON'T talk about how computers can look at ordinary data and EXTRAPOLATE metadata from it! If your "semantic web" extrapolator looked at Shakesphere's "To be or not to be, ..." - just what meaningful metadata would you glean?
    Roger Ramjet
    • Data vs. metada

      The current Web is essentially one big document repository. So in that context, one problem with the data on the current Web is the limited amount of metadata, true. Some metadata can extracted from documents, but it's far from a precise operation. Where there is explicit data (such as hyperlinks), search engines can take advantage (Google LinkRank).

      But there is such a thing as machine readable data (the kind of stuff found in databases). An example would be that the line "To be or not be..." was written by a person named Shakespeare. Semantic Web technologies offer a way for which such statements can be made in globally shareable form. It's not really a matter of extrapolating metadata from "ordinary data", it's more about making ordinary data usable on the Web.

      Note also that search is only one possible kind of interaction with data, one that happens to be preeminent on today's doc-oriented Web. Or to put it another way, the best way to find things is not to lose them in the first place ;-)
  • Interplanetary Communications (eat your heart out Vint Cerf)

    Thanks Danny and Roger for your comments.

    Danny said: "Your "Web as Platform"/"Web of Meaning" is probably as good a summary as there is. But note that it's meaning in the machine-readable sense, so an alternative might be "Web of Data", in contrast to the Web 1.0 "Web of Documents"."

    Richard replies: absolutely, I like the 'Web of Data' term (I used it in the Digital Web article I wrote earlier this year with Josh Porter). We could probably label both Web 2.0 and SemWeb as the 'Web of Data' -- thinking back to Jeff Bezos' definition of Web 2.0 at last years conference: "web 2.0...is about making the Internet useful for computers." So in that sense there's a lot Web 2.0 and SemWeb have in common.

    Perhaps it's more a question of approach or style. But I like what an Amazon web developer said recently when talking about REST vs SOAP approaches to doing APIs. Despite REST being used about 85% of the time by Amazon Web Services developers, the guy from Amazon said that in future the likely approach is to use a mix of REST/SOAP (the latter to achieve the increasingly complex things that the Web will enable).
    • p.s.

      Web20Explorer = Richard MacManus