W3C home > Mailing lists > Public > public-html@w3.org > May 2008

Re: Strategic decision time

From: Aaron M Leventhal <aleventh@us.ibm.com>
Date: Thu, 29 May 2008 17:22:30 +0200
To: Smylers <Smylers@stripey.com>
Cc: "public-html@w3.org" <public-html@w3.org>, public-html-request@w3.org
Message-ID: <OFDCF508CF.2CB618E6-ONC1257458.0053C35C-C1257458.00549D5B@us.ibm.com>
public-html-request@w3.org wrote on 05/29/2008 04:01:48 PM:
> Apologies if I've misunderstood, but I got the impression from elsewhere
> in the thread that if the colon is used then the CSS selector syntax for
> it would be different in XHTML from HTML.
> 
> That is, if I migrate a page from HTML to XHTML, I also have to make CSS
> changes.  That's a barrier.
You would also have to change to use setAttributeNS() unless you use the 
method HT proposed to  setAttribute("aria:foo") within XHTML.

However, using setAttribute with a colon means you need to remember to 
double the number of CSS rules that are based on ARIA properties, in order 
to give the appearance that declaring a property and setting it does the 
same thing. Depending on where you are in [browser verison] x [mimetype] x 
[method used to set attribute] x [has attribute already] space, using 
setAttribute("foo:bar") works differently. So you have to remember all 
that in order to be 100% sure things will work, otherwise a given 
property/value pair will often end up with a different DOM representation 
even in the same runtime for a single page. You'll get no parser or JS 
console warning if you do it wrong. 

Either way there is a significant cost increase:
- extra work to change the code
- hidden bugs that don't make themselves apparent through errors
- extra testing across browsers
- violation of the simple trust that setting a property just works plainly 
and simply without having to memorize and understand arcane quirks

- Aaron
Received on Thursday, 29 May 2008 15:25:39 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 9 May 2012 00:16:17 GMT