Re: [whatwg] Standardizing autocapitalize

> On 10 Mar 2015, at 23:05, Mounir Lamouri <mounir@lamouri.fr> wrote:
> 
> On Tue, 10 Mar 2015, at 20:56, Edward O'Connor wrote:
>> Hi Mounir,
>> 
>>> I would like to standardize the Apple's proprietary autocapitalize
>>> attribute. This attribute is widely used and it would probably benefit
>>> the platform to have a broader support for it. The implementation cost
>>> should be fairly low while it can be very beneficial for the user
>>> experience with the keyboard on Mobile.
>> 
>> Cool. We proposed it here back in 2011, though it didn't end up getting
>> adopted.
>> 
>>          https://www.w3.org/Bugs/Public/show_bug.cgi?id=12409
>> 
>> It might be worth reading the old discussion around this topic.
> 
> It seems that the main reason why it was killed is because it was
> considered that inputmode superseded autocapitalize. Unfortunately, in
> the wild, this is not the case and we (Google) have data about the usage
> of the attribute (cf Hixie's comment 1). I believe it is high enough to
> consider standardizing.

Indeed, there's no denying the difference in adoption between the inputmode and autocapitalize, so standardizing the later makes sense to me, and the proposal seems reasonable of on first read.

However, inputmode was aimed at solving a superset of this problem. I am not going to take a stand at which of inputmode and autocapitalize bests solves the thing they both address. Autocapitalize works and has traction, and that's what matters.

However, if we are going to standardize autocapitalize, we need to make a path forward to deal with the other things inputmode covers, especially when it comes to internationalization.

Whether inputmode can stay as it currently is and we merely need to define how the two interact, or whether inputmode should be redesigned (or replaced?) now that some of its intended functionality is handled elsewhere, is something that needs to be explicitly addressed when standardizing autocapitalize.

Your proposal includes this sentence:

"The user agent may ignore the hint, for example if there is no virtual keyboard being used, if the hint is contradictory with other information or if the system does not allow it."

If the "other information" part is intended to be how this interacts with inputmode, I think we should make it more explicit and detailed.

By the way, we recently obsoleted the ime-mode css property from the css-ui spec[1], as it was a poor design that only worked well on desktop computers running Microsoft Windows in a monolingual environment. We instead point to html attributes such as lang, inputmode, pattern, and type as being the right things for a UA to look at when deciding what kind of input UI to show.

I still think this was the right thing to do, as good internationalized input support could not have grown out of the ime-mode property, but the remaining parts of the platform (inputmode) need to be strengthened, and I wouldn't want autocapitalize to effectively be what kills inputmode.

 - Florian

[1] http://dev.w3.org/csswg/css-ui/#input-method-editor

Received on Wednesday, 11 March 2015 11:24:25 UTC