Yesterday, [url=https://twitter.com/miketaylr]Mike Taylor[/url] from Mozilla kicked off a very interesting open discussion on the use cases for user agent (UA) strings on the modern web. The results of this document will ultimately end up on the [url=https://wiki.mozilla.org/UA/UseCases]Mozilla wiki[/url].
This was a really welcome gesture from Mozilla, which has previously tended to stick to the stance that UA strings shouldn’t contain information that identifies the device type. This stance culminated in the infamous bug [url=https://bugzilla.mozilla.org/show_bug.cgi?id=625238]625238[/url] in which Mozilla argued that Fennec shouldn’t reveal information about the underlying device. In effect, Mozilla was dictating the approach to used to publish content to devices running their browser. More recently, however, I am glad to see that the stance has softened considerably as evidenced by the UAs used by the recently released Firefox OS phones, all of which contain device identifying information.
Far from being a legacy burden, the UA string is an absolutely essential part of the web, quite separate to whether or not you believe RWD is the solution to multi-device websites. There are several other use cases for UAs strings that are nothing to do with adaptation. The authors of the HTTP protocol had good reason to include this feature in HTTP 1.0 and 1.1, and, as is often the case with the early RFCs, their foresight has been quite remarkable.
You can find the resulting document here and contribute to it if you wish. [url=https://etherpad.mozilla.org/uadetection-usecases]https://etherpad.mozilla.org/uadetection-usecases[/url]
Thanks again to Mike for starting the discussion.
Leave a Reply