AW: XHTML Modularization

Hello,

> On Sat, 1 Sep 2001, Christian Wolfgang Hujer wrote:
>
> > 3. Validation of XHTML impossible
> > It should be mentioned in XHTML Modularization, that a full
> validation of
> > XHTML is impossible at the moment since datatypes of attributes like
> > Character cannot be checked for proper use. Using <a
> > accesskey="aksueztn">...</a> will still result in no validation
> errors with
> > XML DTD validation.
>
> Just as a note, I was reprimanded once for using the term ``validation''
> incorrectly
> <http://lists.w3.org/Archives/Public/www-html/1998Apr/0076.html>
> <http://lists.w3.org/Archives/Public/www-html/1998Apr/0083.html>. It would
> seem that ``validation'' refers specifically to the process of validation
> thats (SG|X)ML specifications talk about.  So one can create erroneous
> but valid documents.  Some other term needs to be used.

okay, I give in...

[Definition: An XML document is valid
if it has an associated document type declaration and if the document
complies
with the constraints expressed in it.]

And I am just too lazy to look up what XHTML says about validation and data
types.

My understanding of validation was that there must be a document type
definition and validation is a process of checking wether a document
conforms to that document type definition or not, which means that I even
could have a document type definition in my mind and let my brain check the
document against it to find out wether it is valid or not. I wonder if my
interpretation of validation too idealistic or if my interpretation is
correct and the definition given by XML is just too simple.

But since the definition given by XML is in a recommendation, I see no
chance other than giving in and using the term checking instead of
validation for this purpouse in future.

But I do not give in completely ;)

I thought that it should be mentioned in the XHTML specification more
clearly, that current validation via DTD is definitely not sufficient to
prevent the authors from creating erroneous documents.
It just helps a lot (a whole lot really).

But now I already think different about that. If the XHTML specification
would state this:
- XML users that know that problem already are reminded about that again,
which is not neccessary,
- HTML authors, that do not know anything like always, might just think "why
use validation if it doesn't help me anyway" and completely forget about
validation.

So now I think, let them validate, let them find out the benefits, and if
they discover the limits of validation, tell them, that validation just
helps, but does not prevent them from creating erroneous documents, but
don't tell them in one of the XHTML recommendations :)

Greetings

Christian

Received on Wednesday, 5 September 2001 03:58:17 UTC