- From: Tina Marie Holmboe <tina@elfi.org>
- Date: Wed, 19 Jun 2002 13:40:59 +0200
- To: Jon Hanna <jon@spin.ie>
- Cc: Andrew Johns <andrew.johns@jkd.co.uk>, w3c-wai-ig@w3.org
On Wed, Jun 19, 2002 at 12:08:49PM +0100, Jon Hanna wrote: > > I can assure you that server-side browser sniffing is tempting, but not > > an option I'd suggest for anyone wanting to save their sanity ;) > > Depends on how far you want to go for it. If you have only 1 UA specific > choice (i.e. you have some CSS for all, some CSS not to send NN4, and maybe > some CSS to only send NN4) you can code it in something like ASP in 5 > minutes and still have time to add in some intelligent cache-related > headers. I refer the honorable gentleman to http://www.greytower.net/en/archive/articles/customcss.html which details a slightly off-the-beaten-path experiment that we have been doing for some time now. The conclusion is still inconclusive, BUT: it did turn out that some potential accessibility[1] issues could be solved this way. It is still a technique I would be *very hesitant* in suggesting to anyone, even with a sub-problem defined as "one target UA". For a more generic situation, the implementation is far more extensive than 5 minutes with any method. For ASP/VBscript I'd go so far as to say it would take prohibetively long time. ASP/PerlScript, perhaps. I'll stick with CGI :) [1] I mean 'accessibility' in the broadest sense, now; ie. specifically avoiding valid techniques that give invalid results in a wider range of browsers than Netscape's 4-series. An argument could be made for NOT using said valid techniques, but that's when that old proof regarding the impossibility of making a non-trivial program without errors spring to mind. -- - Tina Holmboe Greytower Technologies tina@greytower.net http://www.greytower.net/ [+46] 0708 557 905
Received on Wednesday, 19 June 2002 07:30:31 UTC