W3C home > Mailing lists > Public > www-html@w3.org > September 2005

Re: XHTML copy and pasting *with* format (and potentially, meaning)

From: Junk Account <avoid.spam.account@gmail.com>
Date: Mon, 26 Sep 2005 22:08:24 -0300
Message-ID: <a0d591205092618084b196d1b@mail.gmail.com>
To: www-html@w3.org
Hi, David

Excellent answer, I couldn't have hoped for more.
I believe, however, that you missed something in the thread's title.
The bit that says "and potentially meaning".

The class atribute is generic, a general purpose specification mechanism.
It is not for styling, and it is not for semantics. It simply is general,
abstract. (*)
With the *current* use of the class attribute (basically, 95% of the cases,
for styling, and nothing else), your answer was simply perfect.
However, the semantic web is coming, big time. And I believe it will be as
widely used as CSS is today (I *do* have reasons for stating that, we can
talk them separately if you wish).
And what are the consequences of the semantic web arrival on XHTML
specification mechanisms? Well, that the class attribute will, suddenly, not
be used for just *one* major aspect, namely, CSS styling. Another massive,
major use will exist, the semantic one. And up to today, we have just the
class atribute, being used by the only player up to now, CSS.
I believe we will need more. And I have just proposed exactly that.
I proposed to create two new attributes (style and meaning, or something
similar), each of them specific to one major, massive use (styling and
semantics), as oppossed to just one generic one, class. In that way, we
could separate any posterior use we could make of our markup, be it for
stylistic-only purposes or for semantic-only purposes. Having just class for
both, won't do, I reckon. And please note that I sustain that class *should*
be kept, nonetheless, and for exactly the same reasons as before: a general
mechanism for unforeseen purposes.

Returning to the problem of XHTML copying/pasting *with* formatting (and
potentially meaning)...Could you possibly extract CSS rules, if class was
used not just for styling purposes, as today? Or would you need an
styling-specific attribute?

In fact, let me please state my question at two different levels:
1) Level 1: Can you even pull it, at all, period?
2) Level 2: Even if you could...do we want to?

The reason for my level 1 question is that if an algorythm for CSS extrating
could not possibly be done (in this new scenario), then we have just
identified the first use case that turns my proposal of two new attributes
into a totally, indeclinable necessitty, even considering the minimalism

The reason for my level 2 question, is that even if by some clever trick,
such algorythm could be pulled out, at all, I believe we should not. I
believe that would cause endless confusion, increasing of the learning
curve, questions and questions for the years to come, and, worst of all,
functionality breaking in the real world.
I believe minimalism *is* desirable (and highly so!). But I believe there
are other criteria as well. For more on this, please see my post yesterday
in the other thread (specifically, my example).

Once more, thank you for your answer, David. Most excellent. Unfortunately,
I think you missed that bit in the title, which shifts the original question
a little. Perhaps my fault?

Fernando Franco

(*) I know that because a certain very knowledgeable character with nick
<Dorward> told me a few weeks back, on irc. :)
Received on Tuesday, 27 September 2005 01:08:29 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:06:11 UTC