pvdm - 10:33 am on Aug 31, 2002 (gmt 0)

I'd like to add some thoughts to this very interesting thread, but there are so many things to think about. As we can't discuss it all in one thread, I'd like to consider only the PageRank aspect. Let's 'abstract' the discussion to a higher level and forget about individual situations and sites. It doesn't matter whether Mr. Brandt has a personal agenda or not. His views on PageRank reflect thoughts and critics that are similar to many other's.

Of course Pagerank isn't perfect. There is a long way to go before an 'excellent' way of searching + ranking will be achieved. At least PageRank has the advantage to be better than the actual (large scale) alternatives. Yet. But things can change quickly these days. Where was this thread about the patents of IBM regarding P2P search algorithms?

I am confused about contradictory statements:
1. 'PageRank favourizes the big sites'. Is that so? Precisely the big sites have difficulties getting their deeper pages PageRanked higher. Where is the favour? 10.000 pages with no real Pagerank can't add much PageRank to a home page. So adding page ad infinitam is not really a clever way to play the game long term. Of course, if one publishes a site with a billion pages, statistically, you'll have more chances to see a page of this site than others. But that is statistics, not Pagerank as it is running now. And that's a good aspect of PR like it is now.
2. 'PageRank puts big sites at a disadvantage'. Is that so? It's normal that EACH new page starts at near zero, and build it's own PR by external quality links. A big site publishing 10.000 pages in a database should not get more Pagerank for each of those pages than my 5 pages published at the same time, if it had the same content. If a big site would get more PR because it has more pages, this would just motivate people to build sites containing millions and billions of worthless database generated pages. That would be absolute nonsense.

This whole discussion concerns the 'neutrality' of PageRank's calculation regarding small and big, old and new. I agree that in some cases PageRank favourizes established, big sites. But in other cases, it favourizes small optimized pages published by one individual. Pagerank will get better, and in the long term, I believe it will achieve better results.

Finally, I don't think PageRank is the only way to go. There are many, many other interesting variables that could be inserted and the algorithm. And the algorithm could even adapt itself to the type of search.

Semantics, thoughts, languages and meaning of words are among the most complex things one can study. In my opinion, actual search technology is just starting to discover how complex it really is. Giving a good computer automated answer or list to a human question or search, has a very, very, very long way to go, even with 100.000 multi-processing-hyper-threading computers...

I am sorry if my thoughts are not of high practical value in this thread, but I hope it opens other insights.