Log in

No account? Create an account
Previous Entry Share Next Entry
Well, that didn't take long

Cavilling hack Andrew Orlowski sticks in the boot following the withdrawal of funding for the Web Science Institute.

Nice to hear one's research described as "webtastic wankery of dubious intellectual merit and zero commercial potential".

  • 1

I sympathise, but...

So, I can't tell how competent Orlowski is, exactly. But you can read his version of his encounter with Wendy, Tim and Nigel and they don't come off very well.

Orlowski, whether by accident or insight, presents them with the big unsolved problem. I know it's the big unsolved problem, Steve knows it, hopefully you recognise it. I'm not actually sure if Wendy, Tim and Nigel know, but if they don't that's a very good reason not to give them this money. Tim tries the "it worked last time" argument and quite rightly Orlowski doesn't buy it, because it didn't work last time, last time is what we have now, and it doesn't really work.

Try it: Google for eXfuze scam. You may find some plausible hits. Follow them, examine. Damn, the one you chose was telling you it isn't a scam. Try another. And another. And another. The eXfuze scammers have money and technology, and with it they've successfully blinded Google.

This provenance problem is really nasty, and Tim's toy systems do nothing about it. Without a _working_ solution the Semantic web is either the Wild West again or just the Semantic intranet and that sounds like something private enterprise should be funding, not government. I like Tim, Nigel and Wendy, but I can't say I'm displeased that they'll have to get on with something else for a while. A few PhD students might get supervised, for a start.

(Sorry, as a result of my choice of example you may get a lawyer letter, or LJ may just remove this entire blog post, eXfuze are just as effective in blinding legal systems as they are at blinding search engines)

Re: I sympathise, but...

In *his* version of the encounter they don't come off very well. That's hardly surprising - he's not going to write a story that says "I asked some stupid questions and the clever people pointed out that I was an idiot".

Yes, poisoning of search engines is a big, unsolved problem, but I'm not convinced that it's *the* big unsolved problem. Nigel, Tim and Wendy are aware of it, for certain.

Orlowski's take, which you characterise as "it didn't work last time, last time is what we have now, and it doesn't really work" doesn't acknowledge the constant co-evolution of spoofing/scamming techniques and counter-measures that are designed to sort the wheat from the chaff. Altavista did a good job of scaling out conventional term vector IR. The scammers misrepresented pages by adding extraneous terms. Google analysed hyperlinks to better rank pages and weed out the scammers. The scammers build networks of extraneous pages. And so it goes.

*One* of the hopes of Web Science is that, by applying computational techniques to the modelling, analysis and prediction of the behaviour of complex networks, we can better understand this sort of co-evolution. It's certainly not the whole of Web Science, but it's an example that most people can understand.

As for Orlowski's competence, consider this from the same article:

Hendler jumped in:

"I was the external reader for a paper for somebody here at ECS. [...]

The first time around, he'd attributed that quote to Danny Weitzner. I set him right on that much (I was the internal examiner for fides's thesis) but he kept the "external reader for a paper" phrasing that a) makes no sense and b) isn't something that Jim said (Jim knows and uses UK academic terminology).

Re: I sympathise, but...

I don't doubt Tim, Nigel, Wendy, or Jim's intelligence, having worked them at various points, however they all have a very laissez-faire attitude to security, which is common across semantic web academics, and many practitioners.

I was at Tim's inaugural too, and I didn't think they answered the point well.

It could be that real trust on the (semanitc) web is impossible, and personally I think that would be a price worth paying for the advances, but it doesn't mean we shouldn't try to do better this time round.

Re: I sympathise, but...

Granted, security on the SW has taken a back seat so far, mainly because there are more fundamental areas that needed to be addressed first (representation, reasoning, query, etc), and that security is concerned with a broader field than just the SW; developments on the Web should work for an extension like the SW.

As for trust - existing Web approaches don't work particularly well, and the different tack taken by early developments in the SW is an indication that CS researchers are *trying* to do better this time around.

Reading through some of his previous articles, Orlowski's main complaint seems to be that the primary goal of CS research isn't market-ready products, but that would be because research is not the same as development. For that matter, academic CS research is a different beast to industrial CS research (a dying breed), with the latter being more closely tied to product cycles and portfolios.

  • 1