W3C home > Mailing lists > Public > www-forms@w3.org > February 2004

RE: A note on <bind /> (similar to <group />)

From: David Landwehr <dlandwehr@novell.com>
Date: Thu, 05 Feb 2004 11:00:17 +0100
Message-Id: <s022144b.010@emea1-mh.id2.novell.com>
To: <Leigh.Klotz@pahv.xerox.com>, <mark.birbeck@x-port.net>
Cc: <www-forms@w3c.org>

That is interesting. Is there backup for this in the spec today, e.g.
should an implementation do this today?

Best regards,
David



>>> "Mark Birbeck" <mark.birbeck@x-port.net> 5/02/2004 10:45:02 am >>>

Hello Leigh,

> Just as <group /> can be used as a container, so can <bind />.
>
> I believe that <bind /> is legal and useful, just like <group />,
> and want to make sure that implementations aren't tripped up on the
> base case.

This is actually how formsPlayer works; everything that could take a
binding attribute (@model/@ref/@nodeset/@bind) will use the current
context if none is present.

So this:

    <bind readonly="true()" />

would set the root node of the first instance in the current model to
read only.

As a result of this, I have often thought we should drop the fact that
submission/@ref explicitly defaults to "/", since that is already
implied by these same rules for context. The following would still use
the root node of the first instance, even if we removed this explicit
default:

    <model>
      <submission action="http://..." method="post" />
    </model>

Of course, much more controversial would be to continue the
consistency,
and allow @ref to be optional on form controls!:

    <input>
        <label ref="." /
    </input>

    <output />


Regards,

Mark


Mark Birbeck
CEO and CTO
x-port.net Ltd.
Received on Thursday, 5 February 2004 05:01:03 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Saturday, 10 March 2012 06:21:57 GMT