W3C home > Mailing lists > Public > ietf-http-wg@w3.org > July to September 2010

Re: remove/deprecate User-Agent header?

From: J Ross Nicoll <jrn@jrn.me.uk>
Date: Thu, 19 Aug 2010 14:20:12 +0100
Cc: ietf-http-wg <ietf-http-wg@w3.org>
Message-Id: <05D6C0F8-02C1-4D68-8FFE-AE2B5B979827@jrn.me.uk>
To: Paul Wise <pabs3@bonedaddy.net>


On 19 Aug 2010, at 13:57, Paul Wise wrote:

> On Thu, 2010-08-19 at 14:38 +0200, Julian Reschke wrote:
> 
>> Clarifying: we aren't defining HTTP/1.2.
> 
> Hmmm, Ok.
> 
>> That doesn't mean there aren't good ones.
> 
> Do you have any examples? I'm unable to think of any.

The button tag in IE 6 is handled differently to IE 8 (unsure about IE 7). Under IE 6, the button tag cannot be used to submit a form (well, not without using some JS and onclick), while under IE 8 it actually follows the specification. We have an application that generates HTML for really nice big buttons with images and text in them under IE 8, Firefox, Chrome, etc. and a standard old <input type="submit"... button for IE 6 (and 7, because we weren't sure). It decides which to generate based on the UA string.

Less critically, we also decide whether to put HTML 5 date fields into pages based on the UA. I know they're meant to fall back onto text entry inputs elegantly, however I don't think our users would agree that "1979-10-14T12:00:00.001-04:00" is exactly a date format they're used to, so we only put HTML 5 date formatting in for browsers that will handle the fields correctly and display a usable UI.

That's just cases where we use the user agent string to work around differences, I'm sure there are plenty of others.
Received on Thursday, 19 August 2010 13:20:45 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 27 April 2012 06:51:24 GMT