As part of the visualisation tool I'm building I'm fetching the parsed revisions of an article. When the article is of a considerable size , eg latest revisions of Barack Obama it takes 10+ seconds. As the tool is interactive and it shows the edits made to an article as an animation the time taken by the server does not bode well. (The requests are only read )

I'm currently not making parallel requests. What would a reasonable degree of parallel requests. Are there other ways to get around this latency issue ?

https://meta.wikimedia.org/wiki/Grants:IEG/Replay_Edits talks about the tool and the project.

Thanks

Jeph