Re: [VE][64] Nesting of Option and Span

Jukka K. Korpela,

Thanks for your informative, thoughtful, and even 
humorous reply.

I hadn't realized that Select controls are not 
line-wrapped, but I can see why the designers made 
this choice. Thus, as you point out, my <span> is not 
even needed. However, I would think that other style 
changes inside <option> elements might be justified, 
such as changing the color or weight of some of the 
characters. And sometimes <option> text must be 
lengthy (even though not in the case of my example), 
so some form of line wrapping is sometimes needed; one 
could argue that text should never be truncated or 
otherwise discarded unless specified by style 
information. But, of course, issues like these have 
nothing to do with validation, as you point out, and 
are therefore immaterial here.

However, the helpfulness of error messages is of 
material concern. Giving excuses about the internal 
design of the validator software makes me feel 
uncomfortable. They are also not a substitute for 
admitting that some error messages may not be very 
helpful and promising to fix them.

I worked in the 1970's and 1980's as a compiler 
developer. I believe that a validator for HTML is 
probably less complex than a compiler for a language 
like FORTRAN or C++. Keeping around sufficient 
information to determine that the current context is 
an <option> element and that a <span> element is 
occurring inside it should not be at all difficult.

Moreover, since the validator is widely used and is a 
key component of the W3C strategy, rewriting it to 
improve its error messages would seem to have 
sufficient rationale.

Finally, providing HTML authors (many of whom are not 
experts in the design of browsers, form controls, and 
so forth) with a better error message, one that 
explains specifically that <span> elements are not 
allowed inside <option> elements, and giving the 
reason for this limitation, and providing a link to 
the actual governing standards text, would fulfil 
W3C's mission much more effectively.

In any case, you have answered my questions well, and 
I thank you.

David Spector
Springtime Software

----- Original Message ----- 
From: "Jukka K. Korpela" <jkorpela@cs.tut.fi>
To: "David Spector" <w3c@springtimesoftware.com>; 
<www-validator@w3.org>
Sent: Saturday, November 24, 2007 7:31 AM
Subject: Re: [VE][64] Nesting of Option and Span
....

Received on Saturday, 24 November 2007 14:07:39 UTC