[UA-discuss] Progress on HTML and email...

[Chaals McCathie Nevile]

On Mon, 13 Nov 2017 06:07:44 +0100, Dr. Ajay Data <ajay at data.in> wrote:

> This is good initiative. I have few queries challs.
>
> 1. After @ , are you not checking / validating. What's the plan.?
> 2. Any specific reason, team decided to consider for before @ only?

We split this into two parts in HTML, so I will describe what we did for  
"after the @" and then what we are planning for "before the @"

Please note:
    1. We do not put things into an HTML Recommendation without real-world
implementation. Help from this group to identify concrete implementations
would be  very useful in getting support for changes.
    2. We aim to publish an HTML version about every year.
    3. We are only discussing here the "email" input type for HTML forms,  
specified at  
<https://w3c.github.io/html/sec-forms.html#email-state-typeemail>

==After the @

In HTML 5.2, which we have effectively finalised and hope to publish this
year, we tried to deal with some of the issues "after the @".

We didn't specify full unicode transfer, due to concerns expressed about
how compatible that would be with reality. We did try to encourage user
agents (e.g. browsers, or scripts in content) to accept "unicode" domain
names - roughly, anything that can be validated as a punycode domain, and
handle the conversion invisibly - so the user types and sees e.g.
chaals@яндекс.рф, but under the hood the representation is punycode which
is what gets transferred.

It seems that we should consider this issue again. If someone is confident
that there is sufficient implementation to justify browsers sending stuff
directly in unicode, please contact me and I will raise the issue on the
HTML spec - or feel free to do so yourself, as noted elsewhere, at
https://github.com/w3c/html/issues

==Before the @

We had a face to face meeting last week which had a short time for HTML
and at which we considered one technical issue - the question of what
should be allowed before the "@", and whether we should keep the email  
input type in the first place.

Note that if we were starting from the beginning, the email input type  
would be quite different - we have learned that there are numerous  
problems with the way it was designed. But it is useful, and its continued  
existence does not create new problems, so for now we resolved to keep it.

We also provisionally agreed that before the "@" we should accept a  
unicode label, not just ASCII which is what is currently required - and  
has been for over a decade without change.

In practice this means we need to change the specification to remove the  
current validation, and specify correctly what is allowed. Text for this  
would be appreciated, but otherwise it will say something like "blablabla  
NFC unicode string blablabla user agents should convert what they get from  
users appropriately blablabla"

This is currently being tracked in an issue:  
https://github.com/w3c/html/issues/845

cheers

Chaals

>
>
>
>
> Thanks.
>
>
> AD
>
> On 12 November 2017 08:43:54 GMT+05:30, Chaals McCathie Nevile  
> <chaals at yandex.ru> wrote:
>>
>> Hi,
>>
>> just to note that this week we reached a provisional agreement to  
>> change the HTML email input type, and make it accept a unicode value  
>> before the "@" instead of the current restriction to ASCII. Once we get  
>> this change made and published - hopefully in a few weeks - we would  
>> appreciate help encouraging browsers to implement it fast.
>>
>> cheers
>>
>> chaals
>
>
> --
> Sent from my Android device with XGenPlus.



-- 
Chaals is Charles McCathie Nevile
find more at http://yandex.com