Angels and Pinheads and Search Engine Page Counting

When I was at Web Search University someone started talking to me about the negative-only searches I was doing in Google - like -the -the -the -the -the - and affectionately pointed out that I'm a total nerd. Which I knew.

But I'm not the only one who likes playing with search engines like this or likes trying to figure out how many pages might be on a particular engine. There's a lovely article at http://addict3d.org/index.php?page=viewarticle&type=news&ID=10263 that looks at negative-only searches to find pages in Google.

The recommended search in the article is -djfdkjkfjkdjdfk . You can also try negative searches which involve non-existent top-level domains ( -site:gahsdgsagdsa ) or non-existent filetypes ( -filetype:asgdisagdsadgdsa ). The filetype has to be used in conjunction with another search, however; it can't be used by itself.

At the moment negative-only searches for nonexistent words are generating about 9.5 billion results. The Encyclopedia Britannica was the first result for most of the nonsense searches I ran. Wonder why? It must be interesting to figure out relevance by word exclusion.

I couldn't get Yahoo to work with negative only searches, but it did return a search result/count for the word "the"; 10.2 billion results. And of course that result doesn't include many non-English results. Strangely if you build a search that might include these results (adding le, la, el, etc and separating them with an OR) you get far, far fewer results -- a few million.

So reassuring to know I'm not the only search engine nerd out there....

Posted to Search Engines | TrackBack