- From: Karl Dubost <karld@opera.com>
- Date: Tue, 17 Jul 2012 16:24:24 -0400
- To: James M Snell <jasnell@gmail.com>
- Cc: HTTP Working Group <ietf-http-wg@w3.org>
Le 17 juil. 2012 à 16:05, James M Snell a écrit : > given how User-Agent's are currently used (and abuse), I have plenty of horror stories, if you want me to share and help understand the problem space. (It's partly my daily job to deal with those.) > define a new Capabilities header whose value is a list of meaningful tokens that identify stuff the user-agent is able to support. > > User-Agent: Foo/1.0 > Capabilities: u, mobile, something-else I have the feeling that will not really work. In the browser world, the way we recommend to do things lately is feature detection [1]. APIs are complex and not always implemented fully at the same time. So Javascript developers can programmatically determine if the device has this or that capability, down sometimes to one specific property of an API. The "Capabilities" would become huge if we wanted to it an exact picture of the user agent capabilities. So that would not really scale. On the other side, I see another reason why "user agent maker" will want to keep their identity visible: market shares statistics (Faust). Don't want to be put apart/Want to have an existence. -- Karl Dubost - http://dev.opera.com/ Developer Relations, Opera Software
Received on Tuesday, 17 July 2012 20:24:56 UTC