Re: Draft minutes for 2010-12-01

Thanks Kurt for this proposal.

I think that it's missing some features that we have found very
important in our use of XBL in Orbeon Forms, including:

* ability to have XForms models/instances that are local to a component
* ability to build controls that look, from a syntax perspective, like
native XForms components (including using xf:label, etc.)
* ability to build controls that "copy" user markup (e.g. tabview
control) vs. use markup internally

Also, we have found that it is necessary to define:

* a lifecycle for the component (when its local models are created, etc.)
* an event model, whereby events can be dispatched to and from the component

Most of this is covered in our "Guide to Using and Writing XBL Components" [1].

I don't have especially strong feelings as to whether XForms must
absolutely use XBL, a slight variation thereof, or something more
radically different. We probably don't need to decide this
immediately.

For the record we have found XBL to be an acceptable general
architecture, including:

* the notion of binding a component to an element
* shadow trees
* templates
* event handling

But also that  extensions are needed. In particular:

* the template language in XBL is poor (in particular there is no way
to address attributes with CSS selectors), so we also support XSLT as
template language
* we use a concept of inner/outer "scope" to switch between an
internal and an external XPath evaluation context and to allow parts
of shadow trees to surface to the level of the user of the component
(case of tabview)
* to emulate XForms constructs like xf:label, etc., some boilerplate
code is needed (specific constructs are desirable)

-Erik

[1] http://wiki.orbeon.com/forms/doc/developer-guide/xbl-components-guide

On Wed, Dec 1, 2010 at 11:53 AM, Kurt Cagle <kurt.cagle@gmail.com> wrote:
> While I'm not likely to respond to all Action items with such alacrity,
> today was quiet, so I thought I'd post my thoughts:
> http://www.w3.org/MarkUp/Forms/wiki/XForms_Modularization
> I'll continue laying out a fuller normative section, but I'd be interested
> in thoughts that others might have on this.
> Kurt Cagle
> XML Architect
> Lockheed / US National Archives ERA Project
>
>
>
> On Wed, Dec 1, 2010 at 12:01 PM, Leigh L. Klotz, Jr. <Leigh.Klotz@xerox.com>
> wrote:
>>
>> Please respond with corrections.
>> Please start new threads for discussion.
>>
>
>

Received on Wednesday, 8 December 2010 06:09:04 UTC