Rob Church: a strong reliance upon text access is not too brilliant for a project running on the toolserver at present.

Yes, it may be not too brilliant, but the TOOLserver doesn't intend to be a nice set of always-on utils of beuty interface. It is a place where *useful tools* (Wikimedia related) can run.

They don't try to be brilliant, they ttry to work. You can see how heterogeneus they are, there's almost no two tools (of different users) seeing _similar_. [1] We even have SQL-injectable scripts!

Not needing to ask the main site would be desirable? Agree. But there's no better computer to have this (unless a toolserver-prime were set at pmtpa)

Gmaxwell: As i previously explained, revisions are cached, so if you could stress it with [[George W. Bush]] (note you only ask for the raw code, it's simply passing it from DB), querying would be only once.

However, a 'search in last X revisions' (with a maximun) would be a good feature :-)

Platonides

[1] Recently there have been tryings to list users & projects (.about.me files and so)