I also agree that structuring an app to take full advantage of conditional GETs isn’t easy, which again is why I feel I couldn’t really have covered it effectively in that article. How one goes about putting such support in an application is highly dependent on the application’s resources and how their representations are computed. Nevertheless, putting conditional GET support in your application is well worth the effort, as the scaling benefits can be enormous.

I’ll certainly continue to ponder this deceptively simple question, though. :-)

And you’re quite welcome regarding my comments about your book. I think the thanks need to go in the other direction, though, as I use the book quite a bit, so thanks for that. :-) It’s a must-have for anyone doing RESTful development.

Any solution which avoids the computational overhead of computing the entire result before determining if the result has changed will scale better than one that does not. And structuring your application in such a way that you can do this is a considerably harder problem than correctly formulating and parsing the relevant HTTP headers.

If you focus on this problem, then the need for an different web server to handle purely static pages is significantly reduced.

That’s the conversation I attempt to elicit via this deceptively simple question.

Meanwhile: thanks for the compliment and the plug for the book. :-)