W3C home > Mailing lists > Public > public-html@w3.org > December 2007

Re: Guide/AlternateElementTemplate

From: James Graham <jg307@cam.ac.uk>
Date: Thu, 06 Dec 2007 13:14:24 +0000
Message-ID: <4757F5B0.4080702@cam.ac.uk>
To: Karl Dubost <karl@w3.org>
CC: Geoffrey Sneddon <foolistbar@googlemail.com>, Steve Axthelm <steveax@pobox.com>, HTML WG Public List <public-html@w3.org>

Karl Dubost wrote:
> 
> 
> Le 6 déc. 2007 à 21:42, Geoffrey Sneddon a écrit :
>> On 5 Dec 2007, at 23:54, Steve Axthelm wrote:
>>> In the Tags section, do also need to document if an element's start 
>>> and end tags are optional?
>> I think this is vital: it's too tedious jumping around so much looking 
>> up the list of omittable tags.
> 
> Interesting. In terms of benefits, I see it on the other side.
> Regular syntax rules are usually easier to learn. If we do not mention 
> the omit tags, people will be encouraged to close them always, it 
> doesn't hurt interoperability.
> And the "übergeeks" can still do what they want by looking at the 
> hardcore specification.
> 
> I'm in favor of not promoting optional end tags, to just ask people to 
> always close their elements.

One problem with putting "lies to children"[1] in such a document is that it can 
undermine trust in tools that are written to the actual spec. For example 
consider the sequence of events:

* An author reads the authoring guide and is led to believe that all tags must 
be closed

* The author forgets an optional end tag in one of their documents

* They validate the document against validator.nu and find no errors

* They subsequently notice the missing end tag and assume the validator was in 
error; their subsequent trust in the validator is undermined.

* The author either does not bother to validate in the future or ignores 
legitimate errors it flags, believing it is not trustworthy.

Similarly, a tool that chose to output conforming code with optional tags 
omitted might be incorrectly accused of producing non-conforming markup, 
unfairly damaging the reputation of that tool.

Given this I strongly believe that any W3C-stamped authoring guide should cover 
the full language and that recommendations about the best way to code should be 
written as suggestions not (inaccurate) requirements. If people want to promote 
a restricted subset of the language, that's fine, but I think it should happen 
outside of the W3C so it is not confused with the actual specification.

[1] http://en.wikipedia.org/wiki/Lie-to-children
-- 
"Eternity's a terrible thought. I mean, where's it all going to end?"
  -- Tom Stoppard, Rosencrantz and Guildenstern are Dead
Received on Thursday, 6 December 2007 13:15:46 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 9 May 2012 00:16:11 GMT