While the purpose of this screencast is to explain a principle, not to focus on the particular piece of software that was implemented, I am left wondering about a couple of things;

The timeline as he explains it, is updated by creating a new timeline instance which consists of previous timeline + new tweets. It seems to me then, that his client does not remove from the timeline tweets that were deleted by its author subsequently to having been downloaded.

To some it might be a feature to capture as many tweets as possible, but at the same time, if your view of the timeline includes deleted tweets then you might find yourself trying to reply to a deleted tweet, and then you would get an error response from Twitter when you try to post your reply. (Though I don’t know if his client also does posting, or if it’s a read-only client.)

Furthermore, what about tweets that are edited by their author subsequently to having been downloaded? Seems that you would not see those edits.