jtara - 6:52 pm on Dec 23, 2006 (gmt 0)

You need to build super computers next to hydro electric dams with software written by 100s of phds to compete in that area

Or so Google would like us to think. IMO, Google hasn't used those PhDs very effectively. Nor have they used much of their talent pool very effectively - most people at Google are over-qualified for the work they are doing.

They need to have them working on semantic analysis, which is the only way to move search forward from here. Instead, they are trying to wring every last drop out of the tired concept of link analysis, and have the masses cowed into so thinking that keywords are the best we are ever going to have that it's starting to affect language, as we start to lose conjunctions - and, unfortunately, meaning.

Lets see, put a page full of content, submit it, get approved and in the index, then change the content on the page.

You simply re-visit, and queue for re-review if there have been significant changes in the page. Do this enough times, with a clear pattern of abuse, and the site gets banned.

But I'm wondering why one would do this in the first place. If one HAD good content in the first place, why on earth would you replace it with bad content?

They key for SE's is to limit that corruption of listings. Google can't do it 100% right,

Or doesn't want to.

-----
There are three huge problems with search as it exists today:

(1) Relevancy of results is really very poor.

(2) There is little or no understanding by the search engine of the semantics of either the search or of web pages. It's amazing how well search does work, considering the the search engine doesn't know either what the user is searching for or what the web pages are about. Keywords are a nice parlor trick, but it's time to move on to the real deal.

(3) Search engines need to evaluate trust, legitimacy, viewpoint, motive, etc. etc. and match those with the needs of the searcher. We've progressed very little along this line, with the sole move forward being link analysis.

The first two of these I think can be eventually handled completely by computer, most likely with human "training" involved. The third almost certainly requires much more human involvement and probably the invovement of the public at large. Fan or not, you have to admit that Wikipedia is the biggest and most successful project to date along these lines.

I wish somebody would seriously take on the first two challenges, but I applaud Mr. Wales for taking on the third. It's certainly the one he's most qualified to tackle, and I wish him the success he has enjoyed with Wikipedia.