[Mediawiki-api] acceptable usage policy?

Scott Wheeler wheeler at kde.org
Wed May 7 15:41:43 UTC 2008


Hi folks --

I'm building a proof-of-concept application that does some work on the 
Wikipedia data set.  I was excited to see the announcement of the new 
API since it would much simplify things for me.

However, as one recent poster pointed out, what is and isn't acceptable 
usage isn't particularly clear.  I'd expect once I put up announce the 
demo that things might hit (complete guesstimation) in the ballpark of 
10k hits per day for a couple days and then probably dropping off to a 
few hundred a day.  Given that Wikipedia averages 30-50k requests per 
second, it seems that such usage would probably be rounding error 
compared to Wikipedia's load.  I'd cache requests that had already come 
across on my server for speed / load reasons.

But what I'd like to avoid is building this nifty demo, announcing it a 
few places and then getting the plug pulled on it.  In the case of you 
know it accidentally becoming The Next Big Thing, I'd naturally move 
over to a DB dump hosted elsewhere.  For clarity, my project doesn't 
have the goal of being a Wikipedia mirror, the demo is just to show how 
the software works on a big data set.

What would even be fine from my side would be just a heads up from 
somebody at WP if we're pissing them off, so that we could rework things 
within a couple days to use a dump.

Is there a policy on acceptable usage anywhere?  I get the feeling from 
a similar question this week that this may be a frequent question.

Cheers,

- [[User:Scott.wheeler|Scott]]



More information about the Mediawiki-api mailing list