Re: change proposal for <main>: possible validation warning heuristic for misuse

Did you use the same dataset for analysis?
Silvia.

On Wed, Dec 5, 2012 at 6:22 PM, Steve Faulkner <faulkner.steve@gmail.com>wrote:

> Hi Silvia,
>
> >Let's be careful what we write into the spec. We should not codify
> something that is common usage right now, but won't work in the future.
> Also, conversely, >we should not ignore something (Scooby Doo) that doesn't
> work right now (because <header>, <nav> etc are not in common usage), but
> might work very well >in a few years when most Web pages have picked up
> <header>.
>
> I agree we should be careful. From the stats below [1] which are from the
> same data set that i used to check Scooby, we should be seeing a much
> better success rate for Scooby than is evident.
>
>  Rough stats on new element usage on HTML5 pages 28% nav, 16% article,
>> 31% header, 28% footer, 13% aside, 24% section. data set:http://www.
>> paciellogroup.com/blog/2012/04/h
>> tml5-accessibility-chops-data-for-the-masses/ … <http://t.co/LGRwZBW9>
>
>
> [1] https://twitter.com/stevefaulkner/status/270475603315146752
>
> regards
> SteveF
>
> On 5 December 2012 06:56, Silvia Pfeiffer <silviapfeiffer1@gmail.com>wrote:
>
>>
>>
>> On Wed, Dec 5, 2012 at 10:33 AM, Steve Faulkner <faulkner.steve@gmail.com
>> > wrote:
>>
>>> On 1 December 2012 22:19, Maciej Stachowiak <mjs@apple.com> wrote:
>>> > This would end up documenting the semi-mythical "Scooby Doo algorithm"
>>> for
>>> > cases when the <main> element is absent.
>>>
>>>
>>> From what I understand, the Scooby Doo algorithm concept was derived
>>> from section '4.13.1 The main part of the content' of the spec. It was
>>> named and further elucidated by bruce lawson [2] . Mike Taylor produced a
>>> script based on the concept [3]
>>>
>>> today I created a bookmarklet from the script (adds a dashed border and
>>> yellow background to what it identifies as the main content)
>>>
>>> Scooby Doo
>>>
>>> and tested it out on a hundred or so of the pages using <!DOCTYPE html>
>>> [4] I found that in approximately 95% of cases the algorithm identified the
>>> main content as either including all of the page content or an element at
>>> the very start of the page. In other words it is of little to no use in
>>> determining either what the main content consists of or where it starts.
>>
>>
>>> I think a much more useful algorithm would take into account id values
>>> commonly used to identify the main content.
>>>
>>
>> Let's be careful what we write into the spec. We should not codify
>> something that is common usage right now, but won't work in the future.
>> Also, conversely, we should not ignore something (Scooby Doo) that doesn't
>> work right now (because <header>, <nav> etc are not in common usage), but
>> might work very well in a few years when most Web pages have picked up
>> <header>. This is why I did my analysis assuming Web pages would use those
>> new elements - where would Scooby Doo end up?
>>
>> So, you could re-run your analysis by adapting your bookmarklet to run
>> Scooby Doo not just on the defined elements, but also on elements that have
>> a @class of header, nav, etc. That may be a bit fairer on Scooby Doo.
>>
>> Regards,
>> Silvia.
>>
>>
>>>
>>> [1] http://dev.w3.org/html5/spec/links.html#the-main-part-of-the-content
>>> [2] http://www.brucelawson.co.uk/2012/scooby-doo-content-element/
>>> [3] http://pastie.org/4663081
>>> [4] http://www.html5accessibility.com/tests/HTML5-main-content/
>>>
>>> regards
>>> SteveF
>>>
>>
>>
>
>
> --
> with regards
>
> Steve Faulkner
> Technical Director - TPG
>
> www.paciellogroup.com | www.HTML5accessibility.com |
> www.twitter.com/stevefaulkner
> HTML5: Techniques for providing useful text alternatives -
> dev.w3.org/html5/alt-techniques/
> Web Accessibility Toolbar - www.paciellogroup.com/resources/wat-ie-about.html
>
>
>

Received on Wednesday, 5 December 2012 07:33:35 UTC