Is Web 2.0 killing the Semantic Web?

by Dan Zambonini

(Disclaimer: Every now and again, I like to break up my bad cartoon blogs with some provocative, opinionated, ill-informed ramblings. This is one such entry.)

I really want the Semantic Web (SW) explosion to happen, and sooner rather than later. But a
sinking feeling in the pit of my stomach tells me that it's still a long way off. And worse still,
that the Web 2.0 momentum could push it further back. Let me explain.

Web 2.0 example - Writely
Web 2.0: Beautiful but deadly

Without getting into a protracted argument over the exact definition of "Web 2.0", let's go with
the general consensus that it's all about people. That's it. It doesn't care about
technology or standards; use AJAX, SVG, FOAF, PHP, Ruby, XHTML, P2P or XSL - it doesn't matter, just
make sure that it's people-oriented. Let people create, collaborate, share and interact. Who cares
what the back-end uses, or how it does it - just give "Power To The People", quickly and

The Semantic Web is the polar opposite: standardise all your data in RDF; encode it in XML (OK,
so there's also N3, but it's probably mostly going to end up as XML); create your OWL. And then, once you
have all this standardised data, let the machines loose on it! Because this data is for computer
consumption, the SW should be more or less transparent to its users.

So whilst Web 2.0 is about high-level (user experience) and immediate benefits, the SW is a
low-level (data), long-term solution. Users are seeing all this cool, flexible new
Web 2.0 stuff, and it's making the SW look even more complex, rigid and unnecessary. Both
technologies appear similar to the outside world - share and aggregate data - but
Web 2.0 has a pretty interface, and is here and now. And thus the (finite) budgets of
organisations are being spent on wikis and blogs, rather than RDF database converters.

Semantic Web example: Haystack
Semantic Web: All hail the true king!

But don't write off the SW. What do we really want from the future web? I mean
want? Web 2.0 has given us more efficient maps. We can share photos. And
collectively criticise the same websites. But, you know something - so what? Are these the
impacts we dream about making; is this our legacy when we die? The SW could save lives. Because it could enable the identification of otherwise un-detected patterns in
large-scale, distributed data sets, it could help find medical cures and aid other problems in
life sciences. It could help detect and prevent organised crime and terrorist activity. It might
help analyse geological or meteorological data and limit the destruction of natural disasters. It
could help detect and contain viruses and outbreaks. It could help distribute and re-use important
educational resources. These are bold claims, but these are the goals we should be aiming for, and
this is why we need the SW to flourish. We can't let a fancy map get in our way.

What's the way forward? Well, we need the SW to take advantage of the Web 2.0 pile-driver.
As Daniel Weitzner recently told me, it's all
about finding the "sweet spot" between the formal SW
semantics and the flexible, free-form Web 2.0. GRDDL
is one such project hoping to help us find this elusive middle-ground, by re-purposing existing web content into
SW data.

We can also take advantage of the flexibility of Web 2.0. As it is technology agnostic, we
can use SW technologies in our Web 2.0 applications and get the best of both worlds (the FOAF RDF
vocabulary has already succeeded at being integrated into many social networking applications).

So lets push things forward. The Web 2.0 applications are amazing, efficient, and without doubt
interesting and a huge step forward. But don't let them distract from the benefits that the SW could
realise. Only 10% of the world population have internet access, and those of us who regularly use Web
2.0 applications a very small niche within this. The SW benefits are further reaching; giving us
developers new toys to play with, but also potentially impacting the lives of the other 6 billion
people in this world without internet access.


2005-10-08 01:34:12
A bird in the hand
Whenever I see articles about the Semantic Web, I keep getting remionded of the AI hype from the 1960s. I remember as a child in the 1970s seeing a TV program told me that some time in the 80's AI would be so advanced that all the important political and economic decisions in the world would be made by computers.

All the factors that make 'brute force' AI based on rules programming next to impossible, also apply to the Semantic Web. Problems such as usefully defining problem domains in a rigorous way, unambigously defining terms that are ambiguous in everyday language, etc, etc.

Untill SW advocates have real working systems to put on the table, I'll stick with the powerful, real web applications I can access now with the click of a button.

2005-10-08 02:40:30
Re: A bird in the hand
It's a catch-22 though, isn't it? Until people see SW benefits, they won't create SW data; but until there's plenty of rich SW data, you can't create the SW applications.

On the brighter side, there are lots of SW applications -- just not necessarily ones the public can make use of. Most of those examples I mention (e.g. the SW being used to detect criminal patterns, share educational resources, and in life sciences) ARE actually happening right now; it's just that most of it is niche, or behind closed doors.

You also have RSS (arguably SW based), FOAF, Mozilla (mainly RDF based, internally), P3P, Yahoo! Creative Commons (, and a load of others -- all "powerful, real web applications". Just not as pretty...

2005-10-09 03:45:52
you forget SPARQL
SPARQL and AJAX ( make for a very powerful combination. And everyone can easily participate. It even seems to have created some enthusiasm with Tim Bray ( . No I think
Web 2.0 is going to play fabulously with the Semantic Web.
2005-10-09 12:46:59
Survival of the fittest
It's about time some real-world pressures were exerted on the Semantic Web (SW) activities. SW was entirely conceived in an ivory tower, and it really shows in results.

Linus Torvalds recently went on a rant ( ) about how useless specifications are. While I don't entirely agree with him, most of the SW specs entirely support his position. Interestingly, he uses "SW" in that email, though he's talking about "software" when he refers to it.

I think the idea of trying to create brand new things from whole cloth inside a standards body is a very flawed notion. In much the same way that SGML served as a first draft for XML, I think that a radical simplification of SW is in order.

2005-10-10 01:16:52
Cannot kill what is not alive
Until now, SW has promised a lot theoretical benefits, but delivered very little practical ones. So if it dies it will not be because it has been killed by anyone, but because it was not able to ever born.
2005-10-10 05:35:35
Friendly co-existence
From what I hear Web2.0 is a lot about user interfaces, or user experience. SW, on the other hand is a lot about things that go on in the background. As someone else already said, the two go nicely together: You could build a nice SW application with a Web2.0 interface. Or, to put it differently, you could build a nice Web2.0 application with an SW backend.
2005-10-10 22:04:00
Negative power?
You mention all the great potential of SW, but venture capitalists will tell you that every truly great invention also has a negative side. The automobile saves lives, and in the wrong hands also kills people. The television educates brilliantly but can also be used to incite violence and hatred. What can the SW do in the wrong hands?

The automobile and the TV also provide immediate personal gain. What do I personally get out of investing all that work in authoring my content the SW way? You're asking me to do at least 10 times the amount of work for the same content. What is my immediate gain? And you can't tell me I have to wait till everyone does it to get any gain. Nothing in the history of this planet has started that way. You can rant and rave all you want, but until you provide immediate gain and greatly reduce the barrier to entry SW is a non-starter.

2005-10-11 01:05:37
Re: Negative power?
"You're asking me to do at least 10 times the amount of work for the same content"

If you were creating data from scratch then, yes, RDF can be a bit clunky and time consuming. But as many SW advocates have pointed out in the past: most data that the SW needs has already been created (usually in existing databases - relational or otherwise). Creating SW data is therefore greatly simplified - you just need adapters/converters to expose this existing data in a SW friendly (RDF) way. It's not really about re-creating data, or creating new data from scratch in RDF - just re-purposing what we already have.

2005-10-28 00:46:33
Web 2.0 will evolve to Haystack
Haystack is having many interesting ideas such as UI continuations etc layered one over other . Web 2.0 developers are now handcoding mashups, which will be automated soon so that end users can themselves do mashups on the fly. An intermediate step in this direction is, which provides tools for mashups..

I think Haystack is the approach for enduser mashup tool using RDF. It would be either as the end of web 2.0 or begining of web 3.0

2005-12-15 08:10:52
JSON and the "Semantic Web"
JSON comes close to being exactly what is needed for the "Semantic Web" to explode not unlike "Web 2.0" but it falls just short of the mark. See the following article/site:

Francis Hastings
2006-04-03 09:56:55
The Semantic Web and Web 2.0 are trying to solve completely different problems. SW is about backend integration between disparate data sets; Web 2.0 is about interfaces and interactivity. They couldn't be farther away on the design scale.
2006-04-14 16:40:06
I didn't realize there was a difference between web 2.0 and SW. It occurred to me that both were leveraging similar technologies (er w3c recommendations: xml, xslt, rss etc), the main difference between the two being web 2.0s' emphasis on improved user interface/experience (especially with ajax yada yada). Web 2.0 (like flickr, google reader, diigo, digg yada yada) content is more compelling and is almost always organized in intelligent ways by human beings, not computers. Computers aren't smart enough, yet.
RDF is compelling, but just how to translate data in RDBMSs to be useful with RDF is just not mainstream yet. I really don't think web 2.0 is doing anything to retard RDFs fruition. I give it until Jan. 7 2007 at 11:23PM PST.