Comment 81 for bug 614742

Yes, this is the problem. We've known that and reduced our API calls as much as we can. The hard part of the problem is the allocation, apparently, is per application, not per user/application. So it is throttling based on all gwibber users, not just throttling the calls for the individual users that are exceeding the limits. The latest version in lucid-updates and in maverick all should be well under our allocation, but according to the statistics on facebook our API calls haven't gone down.

So either users aren't updating or there are other problems with the way facebook calculates the API calls. The facebook documentation is rather vague on how the allocation works, and how it handles mobile/desktop applications vs. web applications. I have been trying to get in touch with anyone from facebook that can help shed light on how it works and have been ignored. If anyone knows anyone at facebook, PLEASE try to get us in contact.