Hi folks, I am just getting started on developing a web application that I wish to interact with Twitter. Essentially, I want to embed Twitter functionality into an existing application - so users of the app can authorize the app to use Twitter on their behalf (via OAuth), and then set up filters to receive tweets, follow their friends, post updates, etc. I am trying to come up with some principles to use when designing my application, based on the way the Twitter APIs appear to work. The application needs to be scalable. I am wondering if these principles look like good ideas to people who have already used the APIs in the context of web applications, so any suggestions or criticisms would be appreciated So with that in mind, here's what I've come up with.
1. I plan on mostly using the REST API to retrieve data, and cache it to avoid hitting the rate limits - on the server when necessary (eg, for less capable web clients that I can't rely on to issue javascript requests to twitter themselves), and on the web clients whenever possible. 2. I think I should try to offload as much work onto the web clients as possible - eg, use javascript twitter APIs to retrieve data directly from Twitter, rather than having the webserver retrieve data from Twitter and pass it to the client. Hm. Actually, I think those are the only principles I've really got so far =). Anyone have other suggestions for me? Do these look reasonable? I haven't used the Twitter APIs at all so far, other than a small test program in Java that just retrieves tweets posted by all of the friends of a given user. The program uses the Search API to do this, via Twitter4J, Thanks! Mike Shea.