Re: Manuals and Spiders

> One can simply have the form posted to > google.com with "site:php.net " prefix, > and that's all... done. I saw many sites > do that... Khm, that is by far too amateurish in my eyes. We should settle down with some cool search engine to provide results for the whole php.net site family. As we have a close contact at fastseach, this is obviously alltheweb for the first choice.... BTW google or alltheweb provide too much noise on their pages which we would like to avoid... You need to ensure for example, that all the mirrors are equally indexed... Then if someone searches on a mirror, he should not be sent to php.net or any other mirror as a result... Then it is also a question how those search engines handle languages... Google for example returns manual/hu pages as hungarian, even if they are English. BUT Alltheweb intelligently finds out that manual/fi, manual/ro and other pages contain English content, and they return them as part of the result list for English restricted searches too, which is not what we would like to offer for our users.... What we need is a strongly backed seach engine, which indexes php.net fully, capable of providing results for any mirror site (official mirror sites have full content), and capable of handling some specialties of the php.net site (like that we have English content under non-English HTTP headers and URLs, but those should not be listed as results for an English only search). htdig can solve some problems, but it is definitely not hat versatile as the big search engines, and damn slow, if you index all of the manuals and pages on php.net... Goba