W3C home > Mailing lists > Public > www-forms@w3.org > April 2006

Re: Digging through section 6.2.1 / multiple types for an element

From: Allan Beaufour <beaufour@gmail.com>
Date: Tue, 11 Apr 2006 16:01:59 +0200
Message-ID: <90d6cb0e0604110701o440f2d1nb32e33396e2150bf@mail.gmail.com>
To: "Lars Oppermann" <Lars.Oppermann@sun.com>
Cc: www-forms@w3.org

On 4/11/06, Lars Oppermann <Lars.Oppermann@sun.com> wrote:
> Allan Beaufour wrote:
> > I'm trying to follow you. You are saying that 1) bind type="" only
> > specifies type (validation) information which is handled seperately
> > from the schema attached information, and 2) it might be used for
> > presentation purposes, but you do not care how. Or?
> Oh my, I hope I'm not adding to the confusion...

If so, I'm sure we can handle it :)

> Allan, John, could you please comment on the following example:
> In a form, I have an instance consisting of just one element:
> <my:element>2006-04-11</my:element>
> In said form, there is also a control bound to that node
> <xf:input ref="/my:element"><xf:label>input</xf:label></xf:input>
> Right now, there is obviously no type information of any sort associated
> with our node. The type-MIP is thus xsd:string (6.2.1-4).
> (A) Now, if there was a schema associated with the instance, which set
> my:element to be of type xsd:date, would section 6.2.1-1 require that
> the type-MIP for my:element be set to my:date?
> If not, should a control be able to know about the type of the node
> (i.e. have access to the type information by other means then the type-MIP)?
> (B) What if the instance is <my:element xsi:type="xsd:date">...? Would
> this have an effect on the type-MIP and is this case any different from (A)?

That depends on whether we talk about the future version of the spec.
or the current spec. I'm still trying to figure out what to do for the
former. For the latter, this is _my interpretation_:

(A) Does not set the type-MIP, but the control would have a way to
figure out the attached type still.

(B) Same as (A)

My rationale is, that afaik we do not define that schema type
information influences the type MIP. We define it the other way around
(which is the cause of all the problems). The control should know
about it still, because imho that makes sense and users would expect

... Allan
Received on Tuesday, 11 April 2006 14:02:11 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:36:17 UTC