Re: 'role' should be property

Hi Maciej,

> The problem with the idea is that it goes against the Architecture of
> the World Wide Web. In particular it seems to go against 4.3.
> Separation of Content, Presentation, and Interaction <http://
> www.w3.org/TR/webarch/#pci>. "Some data formats are designed to
> describe presentation... these data formats should only address
> presentation issues."
>
> [...]
>
> I'm not going to claim the TAG is automatically right about
> everything they have ever said. But I think this particular finding
> is widely shared, at least to the extent that styling languages
> should be considered primarily presentational. It may be that
> discarding this principle would have some value, but it would
> certainly be a big deal, not something at the level of a trivial
> syntactic change.

Agreed.

But to stress...the ideas I am playing with concern re-using the CSS
*processing model*. Say we factored out the processing model from CSS
and called it 'dynamic infoset', or some such. We would then have a
mechanism that allows us to create rules that say 'select these
elements and add this property to them and set it to this value'. That
would be incredibly useful in many contexts.

But then also, a new version of CSS itself would be written, that
says, 'this spec uses the concept of a dynamic infoset' and would
refer to this other spec. And whilst the dynamic infoset is about a
generic solution, CSS is about specific properties with specific
meanings, like 'color: blue;'.

(And to also stress, I wasn't imagining that this issue would be
resolved here. I only raised it because I felt that Dmitri's idea was
being dismissed for the wrong reasons.)

Regards,

Mark

-- 
  Mark Birbeck, formsPlayer

  mark.birbeck@x-port.net | +44 (0) 20 7689 9232
  http://www.formsPlayer.com | http://internet-apps.blogspot.com

  standards. innovation.

Received on Wednesday, 23 May 2007 19:59:52 UTC