Re: possible validator bug - '>' not requred to close tag

Nick, Matt,


On Friday, Dec 6, 2002, at 06:18 Asia/Tokyo, Nick Kew wrote:
> On Thu, 5 Dec 2002 m_to-validatorlist-re_tagclosing@wickline.org wrote:
>
>> The above page has a menu option which is missing the closing > for an
>> option element's close tag.
>
> This is an FAQ.
>
> In the hope that we can assemble a proper FAQ document for the
> validator, let me try and contribute a fuller-than-usual answer
>
> Q: I found some nasty typo like <p<a ...> but the validator accepted 
> it!

Nick, do you remember we had discussed this something like a week ago?
I had updated the "Help" page at that time, and this question is 
already mentioned:
http://validator.w3.org/docs/help.html#faq-typo

> (Olivier, are you OK with assembling FAQs as-and-when someone asks
> a question in an intelligible manner and round tuits can be found)?

Sure. Doc is something I've started working on, but it still needs much 
more work, especially in these two areas:
- error explanations : the explanations are good, already, but could be 
improved by not only explaining "this is wrong" but rather "this is 
wrong, try to do this and this"
- FAQ : I've started to modify the "help page - read this first" into a 
"help page and FAQ - read this first"


> Short answer - If the validator didn't complain, then your usage is 
> legal!
>
> Technical answer - this is an SGML rule known as SHORTTAGS.  The 
> official
> SGML declaration for HTML permits shorttags, so the validator would
> technically be wrong to reject them.  This has been the subject of
> some debate within the validator team.   As we understand it,
> the decision to permit shorttags was deliberate, because dispensing 
> with
> it would also outlaw other constructs which are widely supported in 
> HTML.
>
> Political answer - because it is technically legal, the validator
> must allow it.  The decision is for the HTML working group, not for
> the validator team to make.
>
> Practical answer - though technically legal, this is not supported
> by most real-life user agents (such as browsers), and should not
> normally be used in a webpage.  OpenSP, the SGML parser used by the
> validator, has an option to generate warnings about shorttags.
> These warnings are available as a user option, and are enabled by
> default in some online validation services including Page Valet
> and the WDG Validator.

Nice set of answers. The current answer is much shorter. We could add 
your explanations to the FAQ (except maybe the political one <grin />).

Thanks


-- 
Olivier Thereaux - W3C - QA : http://www.w3.org/QA/
http://www.w3.org/People/olivier | http://yoda.zoy.org

Received on Thursday, 5 December 2002 17:39:46 UTC