I’ve spent a few hours making a new experimental world academic repository search tool, 2012 version. It searches 2,756 repositories. Enjoy.
How it was made: Lists from ROAR, OpenDOAR, BASE, Open Archives, and repository software providers were all URL-extracted and combined in Excel — then comma delimited on the / and the URL path trimmed back a little (not too far), returned to their normal state, de-duplicated, and the resulting list generally cleaned by hand/eye with the aid of a few very useful Sobolsoft addins for Excel. Any special OAI-PMH query/harvester URLs were excluded.
2,756 clean repository URLs remained. These were topped and tailed to turn them into hyperlinks, and all were put into a rough-and-ready on-the-fly Google Search engine.
Drawbacks: Users need to be aware that…
* the search results may sometimes default to being drawn from the main Google database, usually after the first few pages of results.
* you may see Google’s ads with results, if you don’t run AdBlock in your browser.
* the search tool is limited to what the Googlebot is able to “see”
* it may sometimes throw a tantrum and refuse to work.
* it likes quote marks. Idly typing in film noir will get a pitifully small amount of results, but “film noir” will get a hefty amount.
So it’s not perfect, but… you are able to use useful Google Search modifiers such as filetype:pdf or filetype:doc to discover full-text.
Re-usage: If anyone wants to use these URLs for a proper Google Custom Search Engine, feel free. Anyone can make their own 5,000-URL Google CSE, free.
