Re: HTML 4.02 : minor bug fix and update

On Wed, 8 Feb 2006, VK wrote:

> It is urgently needed to fix two old bugs in HTML schemas:

I'm afraid it's completely pointless to suggest changes to HTML 4. It's 
frozen; all work on HTML specs is oriented towards XHTML and specifically 
XHTML 2.0. Even reported simple _technical_ errors (in the way things 
are presented, as opposite to the intended meaning) in HTML 4 specs 
haven't been fixed. And probably won't.

Yet, as a mental exercise, here are some comments of mine on your 
proposal. They might have some practical value to you.

> 1. IFRAME element needs to be added to HTML Strict.

It was intentionally omitted and would not be put there even if, by 
miracle, the W3C started evolving HTML 4. Instead, perhaps, the 
possibility of making an embedded object a target of links, the way you 
can target IFRAME elements, might be added. In the hypothetical world, 
that is.

> Currently even fully 
> conformant web-sites have to use HTML Loose or cheat on validator,

Cheating on a validator is like cheating when playing solitaire.
Well, it _might_ be worse, if you claimed validity (in public or to 
your customers) despite knowing that the documents are not valid.

A validator is just a tool, and a relatively trivial tool at that.

> 2. WRAP attribute for TEXTAREA has to be added to all schemas
> or at least to HTML Strict. It was originally excluded by
> mistake as it was thought that CSS can be used instead.

Perhaps. But a _good_ argument, in a sense, for including WRAP would be 
that the (currently nonstandard) attribute wrap="off" is the only way to 
make many browsers behave the way described in the "standard" (HTML 4 
specification). The spec describes TEXTAREA behavior in terms that imply 
that only user-supplied line breaks should appear.

Of course, this is water under the bridge. Any robust processing of form 
data has had to work under the assumption that a line break received might 
or might not be user-supplied. That is, the line structure of the input as 
intended by the user cannot be inferred. We can live with that.

> this attribute defines *form submission
> behavior* (with or without line breaks in the text) and cannot
> be replaced by CSS rules.

Actually, some values of WRAP affect submission behavior, some don't. 
"Soft" line breaks do not affect the submitted data (though they may 
mislead the user).

> Currently these two bugs are too evident and they putting in a
> difficult position even the best of your followers among
> browser producers. To not look ridicilous they are forced to
> skip on these errors in their own validators.

I have no idea of what you are talking about. Do you think browser 
producers create and use validators? HTML processing in browsers has 
always been tag soup processing, and even XHTML-enabled browsers are not 
required to perform validation.

> Please let know if HTML 4.02 Strict (with these two bugs fixed)
> can be expected any soon.

It can't, in this universe.

-- 
Jukka "Yucca" Korpela, http://www.cs.tut.fi/~jkorpela/

Received on Friday, 10 February 2006 10:18:18 UTC