W3C home > Mailing lists > Public > www-validator@w3.org > August 2010

Re: Check existence of class names

From: <chukharev@mail.ru>
Date: Wed, 18 Aug 2010 15:26:45 +0300
To: "Siegfried Gipp" <siegfried@rorkvell.de>
Cc: "www-validator@w3.org" <www-validator@w3.org>
Message-ID: <op.vhmvivn7mhpy7y@vova-vaio>
On Wed, 18 Aug 2010 11:36:51 +0300, Siegfried Gipp <siegfried@rorkvell.de> wrote:

> Am 17.08.2010 18:35, schrieb chukharev@mail.ru:
>
>> This does not go wrong but is irrelevant, since it's correct only while
>> you are not using the semantics you have added. When you start to use it,
>> you need consistency between different usages of the classes. E.g. if you
>> collect/visualize/do_whatever with names of persons, you need them all
>> marked
>> with class="vcard fn". Those accidentally marked with class="wcard fn" or
>> class="vcard_fn" or many other possible typos, have different semantics.
> Exactly! But it is not an error per se. A class name different to "vcard
> fn" adds a different semantic to the content of the container, That's
> absolutely true. The problem here is, that automatic tools are, as of
> this day, unable to cope with semantics in any way.

You wanted the same semantics and you got different semantics, but this
is not an error? Why? Sorry if this sounds harsh.

> A way might be to enhance a DTD to contain valid class names. But at
> least in html this would be mostly impossible. For xml and xhtml this
> might be possible using namespaces and then there adding a proper
> description. Then an automatic tool could at least recognize correctyl
> tagged parts.

You are trying to limit everything by (x)html. Why? It has been stated
already couple of times that the discussion is not about simple validity.
It's about errors in interaction between valid html and valid css.
No need to say that an implementation if happens should not disturb the
usual functioning of the involved validators.

> What you are thinking about is, check class names for similarities.

Absolutely not.

> You see, in a programming language variables may be defined like:
> int index1;
> int index2;
> int index3;
> These are very similar, but not equal. Do we have a typo here?

No, we have not. But try to extend this example to two file program,
where the second file uses
extern index_1;
extern index_2;
extern index_3;
The program will even run correctly if you do not actually change the
values in the first file and the linker creates and zeroes the undefined
external names (a real situation!).

-- 
Vladimir Chukharev
Tampere University of Technology
Received on Wednesday, 18 August 2010 12:27:24 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 25 April 2012 12:14:43 GMT