W3C home > Mailing lists > Public > www-style@w3.org > August 2009

Re: [css3-ui]? Proposal for ::error pseudoelement

From: Aryeh Gregor <Simetrical+w3c@gmail.com>
Date: Tue, 25 Aug 2009 20:08:51 -0400
Message-ID: <7c2a12e20908251708l68ad7f9ei772a3e94eca231fa@mail.gmail.com>
To: Boris Zbarsky <bzbarsky@mit.edu>
Cc: "Tab Atkins Jr." <jackalmage@gmail.com>, www-style list <www-style@w3.org>
On Tue, Aug 25, 2009 at 3:52 PM, Boris Zbarsky<bzbarsky@mit.edu> wrote:
> Which leads to immediate internationalization issues, right?

No more than the rest of the page content -- including, for instance,
custom validity errors that the author provides, or inputs with both
pattern and title attributes.  If anything, it would be pretty weird
if the browser-displayed error messages were in the browser interface
language, but incorporated a bunch of text in the page's language.

> Suppressing display of the error message altogether seems like something the
> HTML spec should be defining, no?

I think the way to do this would be to set the novalidate attribute on
the form using script, and write your own custom submit handlers for
everything.  Then you're just using the convenient HTML declarative
attributes and DOM methods instead of having to hack them up yourself,
plus you get fallback if script is disabled.

The goal of allowing styling of error messages would be to strike a
middle ground between leaving all error handling up to the UA but
getting a very simple declarative interface; and writing the whole
validation system in script (which mostly, although not entirely,
defeats the purpose of HTML 5 forms to begin with).  The worry is that
if the HTML 5 form attributes aren't customizable enough, authors
might not end up using them much, or at least not without a whole
bunch of script.

Part of the problem might be specific issues with how Opera displays
the errors, though.  I think there'd be much less problem if the
message was clearly part of the browser UI, so that users didn't
expect it to match the site design.  Best to wait and see whether this
is necessary, perhaps.
Received on Wednesday, 26 August 2009 00:09:27 UTC

This archive was generated by hypermail 2.4.0 : Monday, 23 January 2023 02:13:39 UTC