W3C home > Mailing lists > Public > whatwg@whatwg.org > July 2016

Re: [whatwg] proposal for new <input> inputmode: digits

From: Eitan Adler <lists@eitanadler.com>
Date: Mon, 25 Jul 2016 14:37:37 -0700
Message-ID: <CAF6rxg==6LU1Zokq94hBCqUXkpi9aBc4L_j=NiPsvsEfMK6GJQ@mail.gmail.com>
To: Nils Dagsson Moskopp <nils@dieweltistgarnichtso.net>
Cc: "whatwg@whatwg.org" <whatwg@whatwg.org>
On 25 July 2016 at 13:32, Nils Dagsson Moskopp <
nils@dieweltistgarnichtso.net> wrote:

> Eitan Adler <lists@eitanadler.com> writes:
>
> > At the moment if you'd like the user to enter *only* digits (no
> separators,
> > +, -, etc.) you must resort to a hack
> >
> > <input type="text" pattern="\d*" />
> >
> > This results in a correct "digits only" keyboard on some mobile keyboards
> > (and nothing on desktops).
>
> Why do you see a problem with that?
>

Since this is semantically confusing and quite magical behavior. I don't
expect a different keyboard if i provide pattern="hello+ (world|to you)"


>
> > There are several use cases for digits only, but the main ones that come
> to
> > mind are TOTP codes, CVV codes for credit cards, etc.
> >
> > <input type="number" min="100000" max="999999" step="1" />
> >
> > might work, but is non-obvious and still results in buttons for "+", "-",
> > and "." in some mobile browsers.
>
> This is wrong; text containing only digits is not a number.
>

I know its wrong. Hence this post.

> In addition, it may be useful to allow minlengt and maxlength for numeric
> > inputs. This can result in better error messages where the value to be
> > entered needs to be copied from somewhere, and so the minimum and maximum
> > are really proxies for length.
>
> Please continue to use text input elements and the pattern attribute.
>

See also the remainder of my email.



-- 
Eitan Adler
Received on Monday, 25 July 2016 21:38:49 UTC

This archive was generated by hypermail 2.4.0 : Wednesday, 22 January 2020 17:00:38 UTC