Re: [w3c] Old better than new ?

Miguel Casquilho (Lisboa) ( wrote:

>     If you look at this experimental page, where I introduced an error on
> purpose,
> *
> <>*
> you get an (obvious) error with "" and no error with "
>". I wrote "kenter" instead of "center".
>     So, is the scope of  "" only structural ?

It’s more complicated. As far as I understand the situation,
1) is mostly a DTD-based formal validator, with some
additional functionality (added checks), which is not publicly documented
and which has varied quite a bit
2) is a schema-based validator,
3) for a document that declares HTML5 (or ”Living HTML” or whatever you
call it) with <!doctype html>, switches to schema-based
validation (HTML5 cannot be described with a DTD), which is based on code but not identical to it; in particular, it may contain
additional functionality (as in item 1).

Both DTD-based and schema-based validation are “only structural” in the
sense of applying formal rules, using two different formalisms. But
additional functionality may involve things that are outside the scope of
such validation.

CSS code, even when embedded in a <style> element, is not HTML (or markup
in general), and it is completely out of scope for both DTD-based and
schema-based validation. However, the additional functionality in includes CSS checking, presumably by running (some version
of) the W3C CSS validator on the content of HTML elements and attributes
defined to contain CSS code.

Even though such CSS checking may appear useful, you can get a more
adequate CSS checking by using directly
which, in this case, will report the error similarly but also issues some
warnings about vendor extensions (which may be needed for better browser
coverage, but still deviate from CSS specs).


Received on Tuesday, 1 March 2022 06:27:21 UTC