Forms WG Weekly Telecon -- 15 Jul 2009

<wiecha> Scribe: Steven

<scribe> Scribe: Steven

<wiecha> http://lists.w3.org/Archives/Public/public-forms/2009Jul/0005.html

Actions

Charlie: I think the actions agenda item is stale; you're OK on that Nick?

Nick: Yes

Charlie: I shall be updating my actions before I go on vacation
... anything else?

Nick: no

Charlie: There is an item on the backplane, maybe we can report on that

Backplane

Charlie: Our original charter was extended 6 months to October, which is a month shy of TPAC
... we want to present some of our work there
... Initially we identified low-level technologies that needed unifying over W3C
... events, and model-views
... for instance
... That developed into an architecture for opening up the browser to allow apps from different spaces
... such as the Ubiquity library
... Jack Jansen of CWI has been implementing SMIL3 into the framework
... MarkB had already put some SMIL bits in
... Jack has added the more basic timeline stuff
... it starts to look like a more generic technology
... I'm planning to add some ODF
... We're combining the stuff into a demonstrator
... (a finance app)
... there's some SVG in there too
... but the message is that we can extend the browser with new technologies

Steven: An XG is to define a work area
... we started off abstractly
... but now have a concrete demonstrator for pluggin bits of functionality together
... I believe that the XG should recommend that W3C formally defines that architecture as future work

Charlie: We have a couple of months left
... anyone want to join in?
... We were thinking of Google mapstype stuff

Steven: Some of that is already in there
... a range control bound to a pair of numbers
... with a certain class
... displays a map segment that is draggable

Charlie: Would be a good kicking-off point
... there's lots of other things too
... SVG, SMIL,XForms, give pretty good coverage of functionalities
... ODF as an online technology will look like a surprise too.
... Is there a forum?
... at TPAC

Steven: Yes, extensibility

http://www.w3.org/2009/11/TPAC/Overview.html#Group

Steven: Oh wow, registration is already open!

Charlie: HTML5 extensibility will be a hot button

XForms 1.1 PER

Steven: I'm afraid that Ralph didn't reply
... It looks like we'll have to wait now until I am back

Charlie: Away for three weeks?

Steven: Yes

Charlie: Any black-out periods?

Steven: No, none threatened; there is no reason to have to hurry

Custom event information

http://lists.w3.org/Archives/Public/public-forms/2009Jun/0073.html

http://www.w3.org/MarkUp/Forms/wiki/PassingEventContext

Erik: What kind of attributes do you want to have?
... .there are two kinds, some are bindings like ref and some multiple nodesets
... in the XForms context element, we should be able to carry any type of context information
... I propose a new attribute, @select
... taken from XSLT
... do we need this?
... or could we use @value?
... we might introduce variables in the future
... and it would be good if it looked like the XSLT variables
... however we have a large number of attributes
... you could argue against adding more
... though I don't have an actual solution
... XPath 2.0 has non-homogeneous sequences
... with no XPath 1.0 equivalent
... but nodeset would be a bad name if it doesn't return a nodeset

Charlie: Deprecate nodeset?
... use @select in an XPath 2.0 environment

Erik: We use nodeset on bind and repeat
... XSLT foreach does something similar
... We have to be restricted to a nodeset there

Charlie: It depends on how we introduce variables

Erik: If you want to iterate over 1..10, you can do it in XSLT, but not in XForms
... you need to change the processing model to handle the case
... in XForms 1.1 we don't have too many ubiquitous attributes, but @context is coming
... that makes 6
... itemset isn't very different from repeat
... replacing @nodeset with @select, it would make sense to allow other things than nodesets being returned

Charlie: Are there usecases where the nodeset changes radically?

Erik: Updating dropdowns for instance
... where the implementor has to do the work, it is not defined in detail in the spec

Charlie: Nick, do you want to use @nodeset here

Nick: Yes, but I want to expand what a nodeset holds
... so I agree that the name is misleading
... and @select would then be better

Steven: There is a problem that we already use the name "select" in XForms for something else

<nick1> we could introduce ;) itemset

Erik: It is going to be confusing for people who know XSLT
... If it is not consistent

Steven: I understand, it is just a shame that we confuse the XForms user, to prevent confusion for the XSLT user
... There is another issue, that I think we have been consistent in the past, is that we use nouns for attributes, not verbs

Erik: That would argue for @value
... which we use now for bindings

Steven: On <output>

Erik: It is used as a string in general
... On setvalue too, case
... and some nested elements, like setfocus
... name childelement, etc

<nick1> from the output element : 'An XPath expression to be evaluated. The string result of the evaluation is rendered by the form control.'

Erik: <case> is the surprise
... in all of them they are used as strings
... but I think we can make it compatible

Charlie: <output> with a media type?

Erik: No need, it will still be used as as string

Steven: It sounds as if we can define it consistently

Charlie: Sounds like it could work

<nick1> I'm leaning toward value

Charlie: Looking at the open issues on the wiki page

Should the element support the "model", "ref", "bind", and even "value" attributes as well?

Other actions can support xf:context, including xf:send, and dialog xf:show/xf:hide if introduced.

Erik: The show action in our implementation supports nested context element
... to pass information to the dialog
... related to the encapsulation pattern in OO systems

<nick1> sounds a good idea to add it to those actions

Erik: The wiki example uses @origin; that is a mistake should be @value

Charlie: The potential for XPath 2.0 is interesting

<ebruchez> the setvalue/@origin in the example

[ADJOURN]

- DRAFT -

Forms WG Weekly Telecon

15 Jul 2009

Attendees

Contents

Actions

Backplane

XForms 1.1 PER

Custom event information

Summary of Action Items

Scribe.perl diagnostic output