W3C home > Mailing lists > Public > w3c-wai-ig@w3.org > April to June 2003

RE: Browser variations in a "stretchy" CSS design

From: Tom James <TJames@salisbury.gov.uk>
Date: Fri, 27 Jun 2003 16:46:53 +0100
Message-ID: <2EF56870C33AD611BC3E0008C745CA1B01D1571E@BHEXCHANGE1>
To: "'Scarlett Julian (ED)'" <Julian.Scarlett@sheffield.gov.uk>, w3c-wai-ig@w3.org

Julian Scarlett wrote ...

"Another alternative is to take advantage of IE's conditional HTML comments
to serve up a completely different style sheet for all or certain versions
of IE. I tried doing that as an experiment and it seems to work OK. If
anyone knows of any accessibility base reasons not to use conditional
comments then please could they let me know."

Funnily enough, I was looking into this only yesterday, pondering various
"cox model hacks" and other things on a redesign of my own site.

As I understand it, there are two syntaxes for writing these comments. One
reads something like:

<!--[If IE5]>
 IE 5 _only_ will get this code (including any amount of markup)

This is a valid HTML comment (from <!-- to -->) so I can't see an
_accessibility_ reason why it may be a problem. There may be problems with
bandwidth (because you are sending redundant code to browsers who can't use
it). There shouldn't be a forward compatibility problem, because even if MS
dropped support for this form of scripting in future versions of IE - if any
ever arrive - they couldn't retrospectively remove the support form the
installed base of IE5.

The other syntax for writing the comment is be:

<![If IE5]>
 All browsers will get this (including markup) _except IE5_

This looks a lot more problematic to me, as the comments are not proper
comments. You are making an assumption about how [Random browser X] will
treat these "comments" - for example, will they erroneously treat them as
comments and hide what is between; will they treat the comments as
unrecognised tags and render what is between, as MS hope, or some other
behaviour? So I think this method would be flaky on any kind of
accessibility / future proofing grounds.

If you are trying to specifically exclude code from IE5, you could always
rewrite to use the first syntax:

<!--[If !IE5]>
 All MS browsers except IE 5 will get this code. Of course, it is hidden to
non-IE browsers ...

So fairly rapidly it gets complex and server-side detection starts to look
enticing - if you have scripting rights on your server, that is ... See
rview/ccomment_ovw.asp for MS' documentation.

My view is that it could be useful for detecting IE5 unambiguously, but
probably not for anything more complex. The most commonly used situation is
indeed for some kind of Box Model Hack:

<link rel="stylesheet" type="text/css"
<!--[If IE5]>
 <link rel="stylesheet" type="text/css" href="IE5-boxmodel.css">

There are other ways of sending correct box model widths to various target
browsers (the Tantek Hack; the Caio Hack; the Simplified Box Model Hack
etc.) All have advantages and disadvantages, at least from a CSS point of
view. Fundamentally, however, they all require writing two widths for a
given box - one for IE 5, one for compliant browsers (including IE6 if you
force it into standards mode), so arguments about one mechanism or another
simplifying maintenance don't wash for me - all require maintenance of two
separate values, whether in two separate stylesheets or twice in one
stylesheet. For someone who in future has to maintain the site (perhaps not
the initial author) I think the conditional comments mechanism (or a
server-scripted variation) at least has the advantage of being intuitively
clearer than the variations that rely on some kind of CSS parsing bug to
work - but that is just my view.


Dr Tom James
Corporate e-Government Officer

Salisbury District Council
3 Rollestone Street

"Internet Communications are not necessarily secure, and therefore Salisbury
District Council does not accept legal responsibility for the contents
of this message. Any views or opinions presented are those of the author and
do not necessarily represent those of Salisbury District Council.
Anyone replying by email to the author of this message (or emailing anyone
else, using the "@salisbury.gov.uk" address), is advised that such emails
may be read by persons other than the intended recipient"

This e-mail has been scanned for all viruses by Star Internet. The
service is powered by MessageLabs. For more information on a proactive
anti-virus service working around the clock, around the globe, visit:
Received on Friday, 27 June 2003 11:47:20 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 20:36:16 UTC