mismatch of html translate and ITS translate

Hi Yves, Fredrik, all,

(Taking www-international into the loop),

Am 22.02.13 04:52, schrieb Yves Savourel:
> Hi everyone,
> Fredrik has found an interesting problem with the Translate data category and HTML5:
> The ITS and HTML5 semantics don't match.
> See:
> http://www.w3.org/html/wg/drafts/html/master/dom.html#the-translate-attribute
> There are at least two issues:
> -- 1) Empty string:
> In HTML translate="" is the same as translate="yes"
> That's probably not a big issue. We just have to make sure our processors handle this properly.
> Probably a new test file could help for this.
> -- 2) Translate and attributes:
> "When an element is in the translate-enabled state, the element's attribute values and the values of its Text node children are to be translated when the page is localized."
> So, if I got it right: <p translate='yes' class='text'>content<p> applies to both "content" and "yes" as well as "text".
> First this doesn't match the ITS translate semantics where we exclude the attributes, and second, it's obviously not useable as both translate and class are not meant to be translated.
> Note that in the current specification we don't explicitely say that HTML translate must be used in stead of its:translate, but we give an example (http://www.w3.org/TR/2012/WD-its20-20121206/#translatability-implementation) and hints at the use of the HTML version here in the first paragraph of this section: http://www.w3.org/TR/2012/WD-its20-20121206/#html5-markup
> I think we really need to get our mapping between ITS and HTML clear and consistent. There is the open issue 97 for this (https://www.w3.org/International/multilingualweb/lt/track/issues/97) but now I'm not sure we can map HTML translate with ITS translate.

What would your preference be for resolving this:
- asking HTML5 to change its semantics
- making a special case for "Translate in HTML5"?

Also, a question to the i18n WG: all tests available for translate in 
HTML5 so far only test element content
but no attribute content. Is this an oversight?



Received on Friday, 22 February 2013 05:55:54 UTC