I just submitted a whitelisting request, along with a comment along the
following lines:

I think it would be good if Twitter did API call accounting based on the
HTTP auth header whenever one is sent - even if the call itself does not
require auth.

I'm at home right now and although my account (@terrycojones) is
whitelisted, my home IP is not. So while I can make a decent number of
authenticated calls, I can't do much with the app I'm trying to write as
it's making non-authenticated calls. The non-authenticated calls are
accounted for by IP. Even if my home IP were whitelisted, I'd be stuck if I
were in a hotel.

The concrete suggestion is: to always do API call accounting based on the
passed auth details, and if no auth information is passed, then fall back
to IP-based accounting.

That wouldn't change much and is backwards compatible. Auth requiring calls
would be just as they were. But in the case of non-auth requiring calls,
the developer would have the choice: send auth to get account based
accounting, don't send it to get IP based accounting.

Terry

Reply via email to