Re: kelvSYC's Thoughts on the new XHTML Draft

On 5/10/03 1:42 PM, "Henri Sivonen" <hsivonen@iki.fi> wrote:

> 
> On Saturday, May 10, 2003, at 22:40 Europe/Helsinki, Tantek Çelik wrote:
> 
>> On 5/10/03 7:49 AM, "Henri Sivonen" <hsivonen@iki.fi> wrote:
>>> 
>>>>> h1 to h6 Elements:
> 
>>> as
>>> deprecated. What's the point in introducing something that is marked
>>> as
>>> "should not be used" right away?
>> 
>> Comfort.  Adoption.  Understanding.  Transition.
> 
> The elements and attributes that were deprecated in HTML 4 exist in the
> XHTML 1 namespace because they were labeled "Transitional".

Yes, and this greatly aided the adoption/success of HTML4.

In some cases, it turned out HTML4 "Strict" made a few removal errors
('target','start','value' attributes), and the only W3C valid way of using
these removed features is to use the HTML4/XHTML1 Transitional DTDs.

> Now again 
> some old stuff that is labeled deprecated is being carried on to a new
> generation of (X)HTML and there will be more format variations to deal
> with.

You are missing a step.

H1...H6 were are NOT deprecated yet.

This (deprecating them in XHTML2) is the first step.


>> And more importantly, even if it *did* introduce complexity in the
>> software,
>> such complexity is *strongly* preferable to complexity in what web
>> authors/developers have to deal with.
> 
> Don't two different heading schemes constitute complexity for the
> authors?

Yes.  Hence my suggestion to place them in a separate module.  Then that
module as a whole can be deprecated.  We can even group all the deprecated
modules together so people know what is still supported, but should be
avoided in lieu of new & better solutions.


>>> Consider displaying a meaningful
>>> outline of a document that mixes h and h1...h6 for example.
>> 
>> There is nothing stopping authors from confusing themselves if they
>> really
>> want to.
> 
> Given two heading schemes two HTML 2 documents

I'm sure you meant _X_HTML 2 documents rather than HTML 2.0 documents.

> can't be combined using
> naïve copy-paste without ending up in a potentially confusing situation
> which could be avoided if only h and section were allowed.

Copy/pasting between multiple markup documents from different authors always
has potential for confusion.  If the documents were from the same author,
then that author has deliberately made those choices to use h in one case
and h1...h6 in the other case.

The most common case copy/paste case will likely be that of an author
copy/pasting from an HTML4/XHTML1.x document to an XHTML2 document.  In that
case, permitting h1...h6 is very helpful, and will aid transition/adoption.


>> Perhaps you mean they are cruft because "Their definition is completely
>> historical, deriving from the AAP tag set."[1]?
> 
> I mean cruft in the sense that they are redundant considering that
> there will be another preferred way (h and section).

We should not be so scared of a little bit of redundancy.  It's often more
helpful than hurtful.  Certainly if biology is any lesson.


>> Elementary psych.  All new is not as attractive as somewhat familiar
>> and
>> somewhat new (wish I could find the reference - I went to college
>> before
>> textbooks were hyperlinked).  Unless that something new is very very
>> simple
>> and fits (nearly) seamlessly into what you already know.
> 
> There are a number of other familiar elements in XHTML 2.

I was speaking of headings in particular.

But the way you put it "a number of other familiar elements in XHTML 2"
makes me think that you would prefer this approach as I mentioned in my
message:

  "create a brand new super cleaned up and purified language that simply
borrows a few things here and there from HTML"

In that case, rather than trying to twist XHTML2 into this other entity, why
not go all the way and simply propose such a new language (just don't call
it HTML)?  You could even build it from by picking and choosing modules from
XHTML2 and then adding your own.

Thanks,

Tantek

Received on Saturday, 10 May 2003 17:41:50 UTC