W3C home > Mailing lists > Public > public-html@w3.org > May 2007

Re: 'role' should be property

From: Dmitry Turin <html60@narod.ru>
Date: Wed, 30 May 2007 10:10:16 +0400
Message-ID: <688209309.20070530101016@narod.ru>
To: public-html@w3.org

Good day, all.

---Ortogonal axeses (against attribute and CSemanticS)

PHL> CSS defines style, and a role is not just about style, but about
PHL> meaning/purpose...so it should be tied to the content, not the presentation.
TB> "roles" have nothing to do with styling
PHL> I'd suggest creating a new language, with the same mechanism, 
PHL> rather than hijacking CSS.
MB>  doesn't mean I disagree with you
MB>  that it might be best to call it something unrelated to CSS

There are two ortogonal axes: syntax and semantics.
Syntax pass division grouped-ungrouped,
and semantics pass division for-visualization - for-understanding.
Thus we have matrix ('color' represents all items of similar kind):

grouped  ungrouped
|color |cellspacing |visualization
|role  |            |understanding (man-machine interface)

So we have choice:
to put 'color' and 'role' into two different section of document
 (for example, into <style> and <sementics>, i.e. CStyleS and CSemS),
to unite them in one section of document
 (in <style>),
to put 'color' into separate section and to put 'role' into each tags
 (where it should be defined)
If this list is not whole, i agree to listen you continuation.

I give the following arguments to decide this choise:
(1)   man loss from separation them into two section
(1.1) because he _must remember_, for what each item is intended
      (and this separation help by nothing to himself and machine,
      and machine separate them easy in the same time).
(1.2) because there is no guarantee, that third and so on spaces
      (except visualization and understanding) will not appear.
      If this happen, we will be proned to continue entering of
      new section in addition to <style> and <semantics>
      (result of which is described in #1.1)
(1.3) because presence of <semantics> need additional attribute
      alongside with 'class' to refer to CSemS -
      it has sense, only if there are at least two semantic properties.
(2)   man loss from putting 'role' into each necessary tags,
      because it's more comfortable to use CSS as brackets
      (described in my previous letter).

Keep in mind, that arguments #1.1 - #1.3 have right of veto,
i.e. even one of them is enough to refuse separation into two section.

Dmitry Turin
Received on Thursday, 31 May 2007 08:01:33 UTC

This archive was generated by hypermail 2.3.1 : Thursday, 29 October 2015 10:15:21 UTC