W3C home > Mailing lists > Public > www-style@w3.org > July 2007

Re: unite two space: attributes and properties

From: Dmitry Turin <sql40@narod.ru>
Date: Wed, 18 Jul 2007 10:38:33 +0300
Message-ID: <32675128.20070718103833@narod.ru>
To: www-style@w3.org

Good day, Martin.

>> MA> all presentational attributes in HTML should have equivilant properties in CSS ...
>> +1
>> 
>> MA> ... exact name, content model and usage may differ.
>> -1
>> I thought exactly about this, when i created topic.
MA> So as far as I can tell, you are proposing that:
MA>   * All CSS properties be allowed as HTML attributes, and
MA>   * All HTML attributes be allowed as CSS properties
  Yes.
  Let Ai is set of attributes of i-th html-element,
and let Pi is set of properties of i-th html-element.
I offer to append A-A*P into P, and to append P-A*P into A.
We can allow this now.

MA> I think really your main qualm is that the style="" attribute syntax is
MA> quite verbose
There are two ones; second is:
not presentational attributes should not be specified in CSS as far
as binding, aural style, behaviour.

---does not concern our topic: old mistakes

MA> For example, <TABLE
MA> BACKGROUND="#ffffff"> is not correct: the BACKGROUND attribute expects a
MA> URI, not a colour. CSS, on the other hand, considers "background" to be
MA> a shorthand for a number of other properties, including background-image.
Cross of attribute name and shorthand _does not concern our topic_:
this was mistake of planning (project design), because they crosses in mental!
So it was _necessary even earlier_ to change name of attribute itself to
full-meaning - to <table background-image= >.

MA>   * there is a massive overlap between CSS properties and HTML attributes.
As to A*P, if they have different names,
then this is clear mistake of planning !

MA> <table bgcolor="red" background-color="blue">, or you'd have to throw
MA> away either most of HTML or most of CSS.
Cross of attribute name and property name _does not concern our topic_ too:
as i said right before, this is mistake, and we _were duty_ to unify this earlier,
i.e. to recommend one name and to deprecate another.
And for compatibility, we should accept procedure of solving conflict.
I offer:
1) first name in attribute has priority (i.e. @bgcolor over @background-color)
2) name in attribute has priority over name in CSS (as it is now)

MA>  For example:  a#home { href: "/index.html"; }
MA> HTML HREF attribute expects a string
MA> containing a URI while CSS would normally use url("/index.html").
This was mistake (recall razor of Occam),
but thire is no problem:
  <a id="home" href="/index.html">
but
  <a id="home">
  a#home {
    href: url("/index.html");
  }

P.S.
As to unification:
  a#home {
    href: "/index.html";
  }

---in general

MA>   * there are plenty of HTML
MA> attributes that really have no place in CSS.
Orthodoxness, no more.

MA> for the reasons [...] popular recommendation
Next recommendation will be also popular.

MA> style information be
MA> ommitted from the HTML document completely
  After unification, nothing disturb you to omit names,
corresponding to mentioned information.
  But besides that, we can assign values to stay-ed names both in tags and CSS.


Dmitry Turin
HTML6     (6.1.2)  http://html60.chat.ru
SQL4      (4.1.2)  http://sql40.chat.ru
Unicode2  (2.0.0)  http://unicode2.chat.ru
Computer2 (2.0.3)  http://computer20.chat.ru
Received on Wednesday, 18 July 2007 12:46:47 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 27 April 2009 13:54:51 GMT