Oh and also the benefit of users/followers is it includes all the
user information. If you are just
maintaining a social graph of ids, then pulling down all the ids via
followers/ids would be the way to go.
I think for most users this just requires a few requests.


On Sun, Mar 14, 2010 at 9:42 AM, Josh Roesslein <jroessl...@gmail.com>wrote:

> A method via the streaming API to get friendship / follower updates would
> be nice.
> Now it may be better to use the users/followers method instead of
> followers/ids. The reason
> is this is ordered from newest to oldest based on when the user followed
> you. So you would start
> paginating from the start and keep going until you reach a "known"
> follower. At that point
> you should have a list of all new followers. You would still need to scan
> the entire follower list
> to find unfollows (if you need that info).
> Josh
> On Sat, Mar 13, 2010 at 1:31 PM, Zero <zeroh...@qoobly.com> wrote:
>> I currently need to auto-discover new people who have started
>> following me.
>> Here's how I do it:
>> 1. Periodically pull in my followers using '/followers/ids.json'.
>> 2. Compare to my list of known ids to find new ids.
>> The slight downside of this is it seems somewhat inefficient (for
>> twitter).
>> If there was access to an event stream of follow/unfollow requests
>> this
>> would be much easier.  It also seems like it could be done with less
>> latency.  That is, if I have a lot of followers, I'm not going to want
>> to burden
>> the system by fetching the whole list at a high frequency.
>> However, if I were just fetching the latest follows, it seems like I
>> could
>> do this at a higher frequency and not affect twitter.
>> Questions:
>> 1. Is there a better way to do what I want with existing API?
>> 2. Are there emerging features that could make this better?
>> Thanks,
>> Zero

Reply via email to