Re: Explanation of relation between HTML and ITS 2.0

Hi Yves, all,

thanks a lot for your comments. I implemented all in the spec, see
http://www.w3.org/International/multilingualweb/lt/drafts/its20/its20.html#html5-existing-markup-versus-its
and added a note so that we don't forget about the wiki link. Just one 
comment about the "what guidance to give" topic:

Am 29.04.13 13:52, schrieb Yves Savourel:
> Hi Felix, all,
>
>> I have written some text to explain the relation between HTLM and ITS
>> 2.0. The aim is to replace the content of section 1.4 with that text.
>> Please provide comments by Monday evening. I will then make the replacement.
>
> --- In "...to set Translate behaviour in HTML5 explicitly via global rules, and to process local translate attributes in HTML5 with dedicated ITS 2.0 processors, to avoid unexpected behaviour."
>
> "behaviour" should be "behavior" (US spelling)
>
>
>
> --- In: "Some HTLM markup has similar...":
>
> "HTLM" should be "HTML"
>
>
>
> --- The example 10 is missing a '>' for <em>
>
> http://www.w3.org/International/multilingualweb/lt/drafts/its20/examples/html5/EX-its-and-existing-HTML5-markup.html
>
>
>
> --- The paragraph "The Translate data category has a direct counterpart..." should probably be with a bullet to match the other data category items above the example.
>
>
>
> --- For "...To avoid unexpected behaviour, users of ITS 2.0 are strongly encouraged to set Translate behaviour in HTML5 explicitly via global rules, and to process local translate attributes in HTML5 with dedicated ITS 2.0 processors."
>
> This doesn't feel right: In practice we cannot really make a distinction between "users of ITS 2.0" and other users for HTML5 documents. HTML documents are published on the Web and are processed by applications the authors have no control over. So ITS-tagged file will be processed by both ITS and non-ITS processors no matter what we recommend.
>
> How can an author is suppose to use the Translate data category when the translate attribute that has two different official expected behaviors depending on with which tool their document is used?
>
> We said we can't refer to the HTML5 behavior because it's not stable and we can't set it in stone. But we are setting in stone a recommendation which is bound to cause a conflict with the current and, very likely, the future HTML5 behaviors.
> I'm not sure which one is worst.
>
> Couldn't we point to the wiki page from here too? And say something to the effect that for HTML5 the defaults and behavior of the Translate data category are described in the wiki page. And we make sure any wiki page changes are dated/documented and we make the test suite works with the latest changes until ITS 2.0 is a REC?

As I understand other implementers, at least Tilde, see
http://lists.w3.org/Archives/Public/public-multilingualweb-lt/2013Apr/0076.html
"Our development efforts and that will be exceeded somewhere around the 
middle of May"
We cannot assume that there will be test suite work until ITS 2.0 is  
REC. Whatever we settle about defaults for Translate, it has to happen now.

I am not against a wiki page summarizing the latest development of HTML5 
defaults. That was also part of the resolution of issue-118, see the note at
https://www.w3.org/International/multilingualweb/lt/track/issues/118
I had thought that this was meant only as a non-normative pointer, but 
probably I was wrong. I am also not happy with the current state, but 
don't see a better way: how to fix (also in terms of testing, 
conformance etc.) the relation to a moving target?

FYI, I'll put that on the agenda for Bled too. Let's see what this 
week's mail discussion and Bled brings us.

Best,

Felix

>
> I don't know what is the best solution, but recommending ITS users to discard completely HTML5 translate doesn't sound practical.
>
>
> cheers,
> -ys
>
>

Received on Monday, 29 April 2013 12:49:54 UTC