I'd really like to create a thought through proposal for this stuff.
As I understand it, there are at least the following things that as a
developer you'd like to control:
* Should spellcheck be enabled.
* Should autocorrect for virtual keyboards be forced-off.
* Should autocapitalize be entirely off, should it autocapitalize for
"normal" text, or for names.
* Numeric entry (for example zipcodes, where <input type=number> is in
appropriate)
* For Japanese entry, should it use Hirakana or Katakana.
I don't know if other scripts or languages need other metadata. Does
chinese or arabic?
There's a lot of correlation between several of these. For example for
things like code entry you want all "aids" like autocaptialize,
autocorrect and spellcheck. And if spellcheck is turned off then
autocorrect seesm like it should at least default to being off.
Another thing to keep in mind is that a single <input> might need to
handle both latin text, and japanese text. This is something that
happens in particular to bilingual people who have names that uses one
script, but visits websites that uses a different script.
There is a proposal in the HTML spec right now. But it conflates a lot
of the above concepts, and also is pretty hard to understand. And it
doesn't allow specifying handling both for japanese and latin text.
I'd love to find a proposal that is reasonably easy to use, with
sensible defaults, but also allows a high degree of control if the
author so desires.
Mounir (formerly at Mozilla, now at Google) has been looking into
this. So if anyone can help out with the above, please do let him
know.
It'd be good for the mobile web, and make it more competitive with
other mobile platforms, if we can find a good solution to the above.
/ Jonas
On Thu, Nov 28, 2013 at 11:22 PM, Kevin Grandon <
kgra...@mozilla.com> wrote:
> Autocorrect and autocapitalize are two well understand HTML attributes. We should leverage these and do more work to standardize if we deem necessary.
>
> I have filed this bug to implement in gaia:
https://bugzilla.mozilla.org/show_bug.cgi?id=944607
>
> If you oppose this, please comment here or in the bug.
>
> Thanks,
> Kevin
>
> ----- Original Message -----
> From: "Rudy Lu" <
r...@mozilla.com>
> To: "Ehsan Akhgari" <
ehsan....@gmail.com>
> Cc:
mozilla-d...@lists.mozilla.org, "Jan Jongboom" <
janjo...@gmail.com>
> Sent: Friday, November 29, 2013 1:29:15 AM
> Subject: Re: To support autocorrect attribute for input element
>
> Hi all,
>
> Just want add a note to clarify the current status in Firefox OS,
>
> We have x-inputmode support implemented, and now it supports the following values:
> 'verbatim': don't alter the user's input at all
> 'latin': offer word suggestions/corrections, but no capitalization
> 'latin-prose': offer word suggestions and capitalization
>
> As Dale mentioned, it seems we don't have a value for "auto capitalization only".
>
> --
> Best regards,
> Rudy, Mozilla Taiwan
>
> On 2013/11/29, at 上午6:55, Ehsan Akhgari <
ehsan....@gmail.com> wrote:
>
>> On 11/28/2013, 4:08 PM, Jan Jongboom wrote:
>>> Can you give a scenario where a simple autocorrect attribute is not
>>> sufficient? As far as I know we only have one type of autocorrect (at
>>> least in FFOS)?
>>
>> Autocapitalization is one example, auto-grammar-correction is another. In the future we might do other types of corrections as well (such as iOS style phrase abbreviation.)
>>
>> Note that I'm more interested in what we may have in Firefox OS and elsewhere on the web in the future, what we have today should not be the only requirement for designing this feature.
>>
>> Cheers,
>> Ehsan
>>
>
> _______________________________________________
> dev-webapi mailing list
>
dev-w...@lists.mozilla.org
>
https://lists.mozilla.org/listinfo/dev-webapi
>
https://lists.mozilla.org/listinfo/dev-webapi