Then yes, I can get behind the idea of that such a communication method would be very useful.
On Feb 7, 2012, at 4:55 AM, Matthew Wilcox wrote: > Absolutely agree that we should not be concentrating on one specific thing. > What I want is the ability to have the server ask for whatever info it needs > about the client device/environment, and have that sent as a header. And only > sent if the server has asked. I don't mean we want "screen size" alone, that > is but one example of something the server might care about. Bandwidth is > another, as is viewport size, and I am sure there are a lot more. I want to > see a *communication method* defined here, not a one-issue semi-solution to > an overly specific example of what is in reality a much broader problem (the > server not having reliable information about the client).