New Post: Best Practise for contiguous collection

I have been using userstream and filterstream and am very impressed with the API. Thanks.

This approach allows me to collect live data as it is presented on Twitter. However the problem I have is when my collecting process is stopped and then later restarted.

What is best practise for catching up with missed data ?

One thing I wondered about was if one needed to switch over between collecting old data and live data (say at the point catching up had been completed) how would one ensure that no events were missed when catch up is stopped and live data is started?

Thanks