W3C home > Mailing lists > Public > www-international@w3.org > January to March 2014

RE: <input type=datetime-local> HTML5 CR bugs [I18N-ACTION-279]

From: Travis Leithead <travis.leithead@microsoft.com>
Date: Fri, 21 Feb 2014 23:14:43 +0000
To: Tantek Çelik <tantek@cs.stanford.edu>
CC: "Phillips, Addison" <addison@lab126.com>, Paul Cotton <Paul.Cotton@microsoft.com>, "public-html@w3.org" <public-html@w3.org>, "www-international@w3.org" <www-international@w3.org>
Message-ID: <96ff95051e02411faf33333bc163a3fe@BLUPR03MB344.namprd03.prod.outlook.com>
> From: Tantek Çelik [mailto:tantek@cs.stanford.edu]
> I have a specific and a general concern.
> 
> 1. Specific concern: from reading the threads on local/floating, and
> datetime inputs here and on the WHATWG list, I don't see consensus among
> the participants, and if anything I see more questions than answers.

We at least have consensus among the i18n WG for the current proposal, which I don't take lightly. Yours is the first public-html response expressing concern, so thanks! I attempted to summarize [1] the previous threads [2][3] and show how I got to this proposal. By putting this proposal together, we at least have a stake in the ground for discussion.

> 2. In general I'm concerned about functional divergence of feature
> definitions between HTMLWG and WHATWG specs, e.g. for features like
> datetime <input>s.

In general I agree with you :) This topic might also warrant discussion at the f2f meeting. It's worth noting specifically that the WHATWG clones of these bugs have already been reviewed and Won't Fixed without addressing the i18n WG concerns; this proposal is an attempt to do so.

> Background: Years ago I spent a lot of time working on the definitions and
> functionality of the <time> element and datetime inputs in HTML, and
> working to achieve reasonable levels of consensus or broad understanding
> with research and change proposals:
> https://www.w3.org/wiki/User:Tantekelik/time_element
> https://www.w3.org/wiki/User:Tantekelik/time_input_match
> http://wiki.whatwg.org/wiki/Input
> http://wiki.whatwg.org/wiki/Time
> http://wiki.whatwg.org/wiki/Time_element_accepted
> 
> Accordingly, we had (fairly) consistent <time> and datetime <inputs>
> across specs (good for authors, implementers etc.).

Do you believe the current change introduces too much inconsistency between <time> and datetime <inputs> within the W3C spec? Or is this a concern about the consistency of their definitions between W3C and WHATWG?

> I'm worried that this issue (timezones, naming etc.) is just one case
> where such divergence may be happening on the definition of the same
> feature in different specs, and that seems like a regression.

I agree and want to keep the specs as similar as possible, however some differences are legitimate and exist to address the concerns of W3C membership and WG constituents. We have to decide whether this change is going to be one of them. I think it is justified but am seeking feedback. We have started documenting [4] all relevant divergences so that we can track and communicate them more easily.

> From the existing email threads, I don't think there is much to be gained
> in additional email discussion on this topic (though feel free to follow-
> up in email if you disagree).

OK. But I replied anyway :-) 

I think list discussions can help to build consensus, especially when some folks aren't able to attend f2f meetings, and would encourage folks to continue to provide feedback on these changes.

> Thus I've added both concerns to the agenda (with additional URLs to
> bugs/threads etc) for the upcoming HTMLWG f2f in the hopes that in person
> we have a better chance of reaching a broader understanding and consensus
> on these topics.
> 
> https://www.w3.org/wiki/HTML/wg/2014-04-Agenda#Potential_Topics
> 
> If there's additional background reading for these issues, please feel
> free to add more URLs to the Potential Topics.

I just want to clarify the bullets you posted to the agenda for my understanding:

* naming: "local" vs. "floating"?

The naming change I made in 5.1 rewords these terms in the description [5], but does not change any API surface or behavior. Note that Ian stated he was about to make this change himself [6], but refrained because it conflicted with the input type datetime-local.

* which datetime inputs are needed/useful? datetime, datetime-local, datetime-floating?

I'm not sure what you mean by datetime-floating. Are you suggesting the need for a unique third type?

> Thanks for your consideration,
> 
> Tantek

Of the three HTML5.0 CR bugs, the Health warning bug doesn't seem to be a point of concern and so I will backport that change to HTML5.0 CR and close the bug. Per your concerns with the other two, I will address the HTML5.0 CR issues by backporting those changes, but leave the bugs open for 5.1 discussion and feedback. Thanks for your feedback and let me know if that works for you.

[1] http://lists.w3.org/Archives/Public/public-html/2014Jan/0072.html 
[2] http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2012-November/037879.html 
[3] http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2013-July/040277.html 
[4] http://www.w3.org/html/landscape/
[5] http://www.w3.org/html/wg/drafts/html/master/infrastructure.html#floating-dates-and-times
[6] https://www.w3.org/Bugs/Public/show_bug.cgi?id=17855#c1
Received on Friday, 21 February 2014 23:15:24 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 21 September 2016 22:37:36 UTC