Re: AUTOCOMPLETE=off is now a valid tag for input tags

olivier Thereaux wrote:

> On Jan 19, 2008, at 18:30 , Frank Ellermann wrote:
>> It's also a way to get rid of errors and warnings you're determined
>> to ignore.  The X in XHTML is supposed to mean eXtensible.
>
> The X in XHTML comes from the X in XML.

I thought I wouldn't comment on this discussion, as I've written my 
thoughts on validation on custom DTDs on my web pages and various 
messages, but now the discussion seems to get too wild.

Using custom DTDs has nothing to do with any "X" anywhere, in any 
meaning, including the buzzword-like practice of prefixing things with 
"X".

> The reasons behind the choice
> of Strict conformance versus extensibility via namespaces, and methods
> of extensibility being worked upon (modularization of XHTML, compound
> documents with the WICD framework) has *nothing* to do with taking the
> (X)HTML DTDs, adding non-standards stuff and pretending it's still
> standard (X)HTML. You know that, so do I, and my point is that it
> would be badly deceptive to talk of custom DTDs to someone who does
> not.

It is deceptive to talk of custom DTDs as relating to the "X" in 
"XHTML". It is not deceptive to talk of custom DTDs as such; it is more 
deceptive to talk of specific DTDs as something that need to be complied 
with, even so that compliance equals standards-conformance and 
interoperability. Deception of the latter kind is much more serious, 
especially since the W3C does that.

"To show your readers that you have taken the care to create an 
interoperable Web page, you may display this icon on any page that 
validates."

That's much worse than bullshit, since people not familiar with DTDs and 
validation (i.e., the great majority) do not easily recognize it as 
bullshit.

> IMHO pointing toward the HTML5 effort and its in-progress status,
> which indeed is the source of this thread (the original poster meant
> “why is this not validating now that it's going to be in HTML5”, I
> gather) would be much more constructive than giving people custom DTDs
> and letting them think they have valid HTML.

This is the first reference to "HTML5" in this thread, as fas as I can 
see, and I see nothing constructive either in the suggestion to refer to 
it or in "HTML5" itself.

When I have an SGML or XML document that is valid, then surely it is 
valid independently of whether the DTD originates from or is approved by 
the W3C. Calling it "valid HTML" is somewhat misleading but common 
parlance, even within the W3C, when the document belongs to the fuzzy 
set of documents broadly regarded as HTML. After all, the majority of 
HTML documents on the web aren't valid at all, and from the 
specification point of view, there is no way your document could be an 
HTML document without being valid, so "valid HTML" has an unnecessary 
attribute.

To take my favorite example, a document using <nobr> cannot comply with 
any of the W3C definitions for HTML; yet it can be fully interoperable 
(including, as usual, the possibility that some user agents will just 
ignore the tag), unlike many documents that comply with, say, HTML 4.01 
specification.

The only reason for validation is to check that the document uses the 
syntax you _want_ to use. Whether that syntax is useful, and whether the 
document is otherwise OK, are quite different dimensions. It is absurd 
to regard validation as a process that decides on your behalf whether 
you should use, say, the AUTOCOMPLETE attribute or not. After you've 
decided to use it, validation can be useful for checking against 
misspelloings in the attribute syntax, for example; and you surely need 
a so-called custom DTD for that, as long as the so-called standard DTDs 
don't cover it.

Jukka K. Korpela ("Yucca")
http://www.cs.tut.fi/~jkorpela/ 

Received on Sunday, 20 January 2008 11:17:27 UTC