Re: Proposal for developing HTML 5 materials for Web *authors*

Dylan Smith wrote:
> on 11/21/07 7:02 PM, Ian Hickson at wrote:
>> On Thu, 22 Nov 2007, Karl Dubost wrote:
>>> The fact for example to say you should write
>>> <p class="boo"></p>
>>> or <p class='boo'></p>
>>> does not change anything to the parsing algorithm, doesn't change
>>> anything to the implementation of browsers.
>> I've been having trouble working out exactly what you were proposing in
>> this thread so far. But it seems that you are actually proposing that we
>> change the spec to disallow unquoted attribute values (and presumably,
>> attribute names without attribute values), and in general disallow
>> anything that is incompatible with the XML syntax.
> <big ole snip>
>> While it is absolutely fine to have writing conventions and style guides,
>> they should not change the actual language itself. We have no interest in
>> the XML serialisation, and being forced to be compatible with it is of no
>> use to us. On the other hand, we _do_ want to make our documents as small
>> as possible. HTML's various syntax shortcuts are a big help here. They
>> should continue to be allowed.
> I'm not in favor of changing the spec, but refining a preferred convention
> and more precise style guide would be very valuable for those of us who deal
> with input from a variety of sources.
> Getting code from a variety of different producers, at a number of different
> companies, and having little feedback on what you're getting, is a reality
> for those of us in the news biz.
> We get ad code from all over, crazy RSS markup from all over, lousy iframes,
> output from the CMSs of various news orgs, and have to attempt to make it
> all play nice together.
> If there were a more focused "recommended" way to code, this might be a tad
> easier.
> Not that we should change what's allowed, or restrict rendering, but a
> smaller subset that says,"Please do it this way" is something I'm in favor
> of.

But of course the "Please do it this way" has never worked before on 
anything has it. I think having only the one way defined in the spec 
would be what's needed.
Having a webpage fail a validation test for not being interoperable 
would be a good way to get people to stick to the one method that works 
in all serialisations.

> FWIW, I'm for double quotes and the use of a solidus, myself.
> Dylan Smith

Dean Edridge

Received on Thursday, 22 November 2007 03:05:37 UTC