Forms Working Group Teleconference -- 11 Apr 2012

<trackbot> Date: 11 April 2012

<Steven_> Scribe: John

<ebruchez> zakim +1.650.919.aacc is ebruchez

<Steven_> Scribenick: John_Boyer

<ebruchez> all, I am calling from my cell and will be in my car for ~15 mn, might be "listen-only" for that time

Co-chair

Finding co-chair in progress, two candidates

Editorial Meeting

Steven and Nick will be working in Amsterdam Wed and Thu next week

Work Wednesday, bring issues to next call, then work Thursday

Longer call next week, 2 hours, starting at normal time

<Steven> http://lists.w3.org/Archives/Public/public-xformsusers/2012Apr/0013.html

Updating wikipedia pages

Please check wikipedia pages in languages you understand

update as necessary

Leigh noticed that XForms is a Japanese international standard (JIS)

<nvdbleek> http://lists.w3.org/Archives/Public/public-xformsusers/2012Apr/0015.html

(on Japanese wikipedia page)

<scribe> ACTION: Steven please add Michael Sperberg-McQueen's XForms course to News Items [recorded in http://www.w3.org/2012/04/11-forms-minutes.html#action01]

<trackbot> Created ACTION-1886 - Please add Michael Sperberg-McQueen's XForms course to News Items [on Steven Pemberton - due 2012-04-18].

Users community group

<scribe> ACTION: Steven please ask Leigh if he would be interested in chairing users community group [recorded in http://www.w3.org/2012/04/11-forms-minutes.html#action02]

<trackbot> Created ACTION-1887 - Please ask Leigh if he would be interested in chairing users community group [on Steven Pemberton - due 2012-04-18].

Alain: Leigh already tweeted about chairing this group

<nvdbleek> https://twitter.com/#!/leighklotz/status/185033274500972544

Nick: formally we have to vote but I don't mind if Leigh is chair

Steven: +1

John: +1

Alain: I just voted for him

Nick: There is a poll, I sent a link last week

Alain: When you go to the page, it already says Leigh is chairing

<nvdbleek> http://www.w3.org/MarkUp/Forms/wiki/Main_Page

<scribe> ACTION: Steven please add Community group link to working group home page [recorded in http://www.w3.org/2012/04/11-forms-minutes.html#action03]

<trackbot> Created ACTION-1888 - Please add Community group link to working group home page [on Steven Pemberton - due 2012-04-18].

XPath 2.0 bindings to atomic values

<Steven> http://lists.w3.org/Archives/Public/public-xformsusers/2012Apr/0010.html

<nvdbleek> this is my reply http://lists.w3.org/Archives/Public/public-xformsusers/2012Apr/0011.html

John: In that email, Nick says UI bindings are not restricted, but I think they are. Did something get removed?

<Steven> Scribe: Steven

Nick: I did a reading of the spec, and couldn't find any restrictions

<inserted> Scribenick: John_Boyer

John: In input element for example http://www.w3.org/TR/xforms11/#ui-input
... Data binding restriction

<Steven> ScribeNick: John_Boyer

John: Most ui elements have this data binding restriction that says what they can bind to
... Even though ref attr can now do Nodeset binding, the actual spec for input still says "Single node binding" so the ref will be boiled down to one node.
... Then the data binding restriction says it has to be a leaf

<Steven> Data Binding Restrictions: Binds to any simpleContent (except xsd:base64Binary, xsd:hexBinary or any datatype derived from these).

John: All the controls have these restrictions in XForms 1.1

Erik: Yes I remember those, but question is whether or to what extent should be lifting those restrictions?

Nick: Especially for repeats to repeat over sequences

Erik: In our implementation we reached conclusion that we should not throw an error if a repeat binds to a sequence 1 to 10. #1 it works out of the box, and #2 it was a good use case
... Not having this meant you had to do loops that generated instance values, very messy
... It is most useful for repeat and itemset
... Opens up what used to be a Nodeset binding to more interesting use cases, but it is less useful for single node bindings.
... It could be useful for output element too, but not as big a requirement due to value attribute

Nick: It makes some sense for a range control

John: Really, where do you put the input result?

Nick: It is useful reuse the range control to display a value but it behaves as if readonly
... Could be useful for custom controls too

Erik: For implementer, better not to restrict because there is no benefit.
... You already have to support readonly controls, so there is no harm in binding in a readonly way to atomic values

John: Benefits implementers, but does it benefit or is it a detraction to authors?

Steven: are there any restrictions we should have?

John: OK to lift restrictions as long as they are replaced with description of what the behavior would be, and so far the only stmt I've heard is that binding to things that aren't instance nodes means behave as readonly.
... Any other behavior comments needed?

Erik: Something more to be said for repeats

<ebruchez> http://www.w3.org/TR/xforms11/#ui-repeat-processing

Erik: We say what happens to repeat collections in terms of when you insert or delete instance data nodes, and that would have to be updated to account for things that are not nodes

<ebruchez> repeat ref="1 to 10"

<ebruchez> repeat ref="1 to 11"

Erik: Also, unlike DOM nodes, atomic values can be identical without being the same object
... In above, the repeat object needs to be updated with the 11. The 1 to 10 are different objects but are identical
... If you implement XForms 1.1 already, it is not hard. If it is not a node, you just compare values.

Steven: What happens in your implementation when it goes to 1 to 11?

Erik: If the 1 to 10 are in same order, then those repeat objects aren't changed, and just one new repeat object shows up corresponding to 11
... If the sequence order changes, then more repeat objects get updated

Steven: Nick is there any more info needed to update spec?

Nick: We need a resolution on whether repeat and itemset can bind to atomic elements
... Need a separate one for whether other controls can bind to atomic values

John: Do you also mean XForms bind too? or just repeat and itemset?

Erik: We haven't done work on that?

Steven: Perhaps it should be in bind too?

Nick: We need to consider whether model item property attributes can result in atomic values too.

Erik: that could be a binding exception
... Are you saying that we need to consider whether MIPs would attach to atomic values?

John: No I misinterpreted what Nick was saying as the thing I was thinking about, which is that MIP expressions could result in an atomic values. The issue of attaching MIPs to an atomic value is separate

Steven: What's the blocking issue?

Erik: Can you write <bind ref="1 to 10" required="true()"/> ?

Steven: OK so we need to decide whether to ignore it or throw an error.

John: Also, can you say <bind ref="some/node" calculate="1 to 10"/>

Nick: that will be a space-separated string with values from 1 to 10. the result, even if atomic, is converted to string for calculate or boolean for other MIPs like relevant

Erik: It would be good if we could just point to XPath 2.0 to specify how the conversion happens

Nick: We already say that, it is converted as if by calling boolean or string function

<Steven> Proposal 1: Allow repeat/itemset etc to bind to literals

<Steven> etc = wherever @nodeset is used

<Steven> Proposal 2: When binding input etc controls to literals, the control becomes read-only

<Steven> Proposal 3: When using <bind/> to bind to a literal, MIPS will be ignored

<Steven> Proposal 4: When using <bind/> to bind to a literal, generate an exception

John: etc is bad, need to look at all uses of nodeset, e.g. insert and delete

<Steven> ISSUE: What to do with insert and delete on a literal

<trackbot> Created ISSUE-1 - What to do with insert and delete on a literal ; please complete additional details at https://www.w3.org/2005/06/tracker/xforms/issues/1/edit .

RESOLUTION: Allow repeat/itemset and other uses of nodeset to bind to literals (some details still to be worked out, e.g. insert/delete)

<scribe> ACTION: Nick to implement in spec the resolution to allow repeat, itemset and other uses of nodeset to bind to literals [recorded in http://www.w3.org/2012/04/11-forms-minutes.html#action04]

<trackbot> Created ACTION-1889 - Implement in spec the resolution to allow repeat, itemset and other uses of nodeset to bind to literals [on Nick Van Den Bleeken - due 2012-04-18].

<nvdbleek> update 9.3.3 http://www.w3.org/TR/xforms11/#ui-repeat-processing, section about bind,....

RESOLUTION: When atomic form controls (e.g. input) bind to literals, they behave as readonly

<scribe> ACTION: Nick to implement in spec "When atomic form controls (e.g. input) bind to literals, they behave as readonly" [recorded in http://www.w3.org/2012/04/11-forms-minutes.html#action05]

<trackbot> Created ACTION-1890 - Implement in spec "When atomic form controls (e.g. input) bind to literals, they behave as readonly" [on Nick Van Den Bleeken - due 2012-04-18].

RESOLUTION: When using bind to bind to a literal, MIPs attached to literal generate a binding exception

<scribe> ACTION: Nick to implement in spec "When using bind to bind to a literal, MIPs attach to literal generate a binding exception" [recorded in http://www.w3.org/2012/04/11-forms-minutes.html#action06]

<trackbot> Created ACTION-1891 - Implement in spec "When using bind to bind to a literal, MIPs attach to literal generate a binding exception" [on Nick Van Den Bleeken - due 2012-04-18].

- DRAFT -

Forms Working Group Teleconference

11 Apr 2012

Attendees

Contents

Co-chair

Editorial Meeting

Updating wikipedia pages

Users community group

XPath 2.0 bindings to atomic values

Summary of Action Items

Scribe.perl diagnostic output