google isn't useful to find mathematical documents

[matthias]

Google would really really score (at least for me :-), if they would find a way to easy the search for rather technical documents. While it is not impossible to search for something like "Euler's phi function", it can get quite hard to get any useful results in even more compicated cases (eg where a function isn't named at all).

What are the odds, that this will ever happen? Any ideas how to do this? Any tipps how to search for such things today?

[Macguru]

Hi matthias,

Did you try a search with quotes? I find the results pretty relevant.

A lot of math sites use images because some special caracters are not supported on the web.
Can this be part of the problem?

[RBuzz]

Give me some examples of how you're searching.

[matthias]

I tried to search for Lambda (in a certain context which I can only sticky mail after request). Results are mostly ps, pdf files, or sites using MathML. It should be possible to search these formats for formulars., shouldn't it.

BTW: Even if you put the query in quotes, things like braces are ignored.

[RBuzz]

Try this query:

Lambda (brace ¦ braces) (formula ¦ formulars) filetype:html

Modify it to your circumstance and let me know if that's any closer to your target.

[karakas]

Try Scirus [scirus.com]. Or Mathematics Web [mathematicsweb.org]. They provide search capabilities specially designed for the needs of mathematics researchers.

MathML was specially designed to make mathematics equations searchable. But with its current schizophreny of presentational vs. semantic notation, this goal does not seem to be achieved nowhere in the near future. This is work in progress and a fast moving target if you are trying to write math for the web.

For your purposes you would need a search engine capable of indexing semantic MathML. I don't know of any one around.