Re: HTML5 Authoring Conformance Study

On Sun, Mar 21, 2010 at 3:01 PM, Maciej Stachowiak <mjs@apple.com> wrote:

>
> On Mar 21, 2010, at 12:57 PM, Shelley Powers wrote:
>
>
> On the other hand, I can imagine it will be a frequent desire for
>>  otherauthors to check that they have not used any presentational markup.
>> Many developers choose to use all CSS for styling, and would like a tool to
>> help enforce that.
>>
>>
> I find that View->Source and using Edit->Find works well in finding things
> in a page.
>
>
> Is that really a tool you would recommend for finding all use of
> presentational elements and attributes, and to use repeatedly while
> maintaining a page?
>

What page?

Do you mean template pages, which are what is used for most larger sites
now? One can easily find existing uses of presentational attributes and
elements using command line tools with versions of grep, or using any
variety of other tools. I certainly don't believe that running a validator
such as validator.nu or the W3C validator over _each and every page_ to be
an effective way to find problems at a larger site. Would you?

The days of single hand crafted web pages are over, done, a thing of the
past. It's all templates now. Templates and tools that generate the markup.
Validators aren't worth squat when it comes to ensuring templates and tools
consistently deliver the correct markup. Not unless they're somehow built
into the tools, directly.

So, yes, my approach is just as useful as taking the same page and running
it through validator.nu.


> Or to check that a large team of Web developers are all following policy on
> this?
>


> Do you think it would be an improvement over the HTML4/XHTML1 solution for
> this (separate Strict doctype that bans presentational markup)?
>
>
That's not the original context to my statement. You said that people could
find it useful to use the validator to find the presentational attributes
and elements in their web pages. I provided another approach to how people
could find these items, rather than running the page through the validator.

Now what I believe you're asking me is whether I think the strict doctypes
are unnecessary?

I never used to think so, but that was until I started mixing RDFa and
inline SVG in my web pages. About that point, I stopped using the W3C
validator.

The problems with lock-in when it comes to page vocabulary, is you never
differentiate between an accident (left the 'f' off the href, which I've
done), an error (duplicate ids, which I, or I should say my faulty template
coding, did), and a new innovation (mixing perfectly legitimate RDFa AND SVG
in XHTML, which I do).

Once you've lost the ability to differentiate between an accident, an error,
and an innovation, all you have is too much red background--all you have, is
noise.

I'm not sure, though, that this isn't drifting away from the topic of this
subthread. In a way, though, it's quite relevant to the problems with not
having a truly good means for distributed extensibility in HTML5.

Regards,
> Maciej
>
>
Shelley