Re: Proposal of @ua

Brad Kemper <brkemper@comcast.net> wrote:

> The one I proposed suggested being able to detect both the rendering  
> engine and the browser, so that one could also account for browser- 
> specific items, such as how form fields are drawn (different in  
> Camino than in FireFox, for instance). But I agree that the renderer  
> (Gecko, Webkit, Trident, etc.) is more important. I would still be  
> ecstatic if that were included in media queries, even if the browser  
> name were not.

I wouldn't.


> I never said that the entire UA string should be used, as that is  
> already hopelessly muddied. I suggested a media query that the  
> software could provide simple, honest values to, not something that  
> is derived from javascript or HTTP headers (as we already have that).

> Any UA spoofing another in a media query, in order to pick up CSS  
> meant for dealing with another UA's rendering bugs (or differences of  
> interpretation), would be doing so to their own detriment. I just  
> don't see it happening. The implementors in this WG could keep it  
> from happening in the prevalent browsers.

Of course it will happen.  Implementors may not want it to happen, but
stylesheet authors will do it anyway.  Then we'll be in the same total mess
that we're in with HTTP User-Agent and navigator.userAgent.  What will
happen is that an author will think to themselves "Only BrowserABC version
1.234 and later implement this new standard feature XYZ, so I'll check for
that browser", and it'll be downhill from there.


-- 
Stewart Brodie
Software Engineer
ANT Software Limited

Received on Wednesday, 21 November 2007 10:37:07 UTC