Standard way to detect mobile browsers in a web application based on the http request
We are beginning to go down the path of mobile browser support for an enterprise e-commerce webapp (Java/Servlet based). Of course there are many decisions to be made, but it seems to me the cornerstone is to be able to reliably detect mobile browsers, and make decisions on the content to be returned accordingly. Is there a standard way to make this determination (quickly) based on the http request, and ideally glean more information about the given browser and device making the request (screen size, html capabilities, etc?).
I would also appreciate any supplemental information that would be of use from someone who has gone down this path of taking an existing large scale enterprise webapp and architect-ing out mobile browser support from the development side.
[edit] I certainly understand the request header and the information about a database of standard user agents is a great help. For those talking about 'other' request header properties, if you could include similar standardized name / resource of values that would be a big help.
[edit] Several users have proposed solutions that involve a call over the wire to some web service that will do the detection. While I'm sure this works, it is not a good solution for an enterprise e-commerce site for two reasons: 1) speed. A call over the wire for every page request to a third party would have huge performance implications. 2) dependency/legal. We'd tie our website response time and key functionality to their service, which is horrible for legal and risk reasons.
Answer
Wouldn't the standard way be to check the user agent? Here's a database of user agents you can use to detect mobile browsers.