It would indeed be useful Vesp. There are a number of journal articles i am dying to read. But they are so expensive. I Don't understand though, i would have thought that it would have been the IP address that the journal serves use as a basis to identify google bots. Not the header. I Would have thought that headers would have been too easy to spoof obviously. And it seems like the journals have now realized that flaw?

I'm not seeing it on BL, could you give a link?

There is a firefox add on that does it, so no need for programming.
IP is probably the case -- turns out the googlebots and the google translation page/server use different IP addresses, and so my experiment of using google translate as a proxy didn't work either.
I assume that you probably can't get a proxy using a google bot IP so its pretty hopeless.. unless their is a chance that msn, yahoo or other would use the same IP address for a translator as well as their bots, but I think that is highly highly unlikely.

Referrer spoofing

Some websites, especially many pornographic paysites, utilize referrer information to secure their materials: only browsers arriving from a specified set of approved login pages are granted access. This facilitates the sharing of materials among a group of cooperating paysites. Additionally a site may want users to click through pages with advertising content before directly being able to access a downloadable file - using the referring page or referring site information can help a site redirect unauthorized users to the landing page the site would like to use.
If attackers acquire knowledge of these approved referrers, which is often trivial because many sites follow a common template, they can use that information combined with this exploit to gain free access to the materials.
Spoofing often allows legitimate access to a site's content where the site's web server is configured to block browsers that do not send referrer headers. Website owners may do this to disallow hotlinking.
It can also be used to defeat referrer checking controls that are used to mitigate Cross-Site Request Forgery attacks.

From WIki.

Also the searches generate .txt files or just raw metadata from pdf files to parse and index, if the pdfs permissions are not readable by world even googlebots wont get access to them (just metadata in the form of .txt or even the contents of the metadata loaded straight into an array functioning as memory in the code not even an actual file which then gets automatically indexed). lol I have been doing a little bit of study on search engines lately. If you were lucky you would get access to the text generated from a pdf, we need to spoof .edu sites that get automatic access (imho). If we could find the approved referrers for example sciencedirect, and spoof them we could set up a web portal that lets you surf with those credentials (like an online proxy).

Sounds like an absolutely awesome idea to me too. But you need to find the approved referrers first. How can we find them??