W3C

- DRAFT -

Forms Working Group Teleconference

06 Feb 2013

Agenda

See also: IRC log

Attendees

Present
nvdbleek, Steven, ebruchez, unl, +1.443.837.aaaa, Kurt
Regrets
Philip
Chair
Steven
Scribe
Steven

Contents


<trackbot> Date: 06 February 2013

<scribe> Scribe: Steven

The XForms Dialog Module + relevance

http://lists.w3.org/Archives/Public/public-forms/2013Jan/0019.html

Steven: The last remark last week was from Uli "We need to talk about relevance"

Uli: About show and hide events

<unl> http://lists.w3.org/Archives/Public/public-forms/2013Jan/0021.html

Uli: Point 2, suggested behaviour

- close the dialog if it becomes non-relevant (we can't do much else anyway)

- don't re-open if it becomes relevant (idea: non-relevant controls lose

state information)

Uli: Is this consistent with other controls?

Nick: More or less

Uli: I'm not sure that our controls lose state when they go from relevance to non-relevance and back
... it may be implementation-dependent
... does it get recomputed?

Erik: I'm not sure we ever finished that discussion
... compare switch-case which has changed state, what happens then? This would be similar.
... maybe if the spec says, we can do the same.
... In any case it seems odd that it should pop back into life.

Nick: Agree

Uli: OK

Steven: So agreed.

The bind() function

http://lists.w3.org/Archives/Public/public-forms/2013Jan/0022.html

<unl> http://www.w3.org/MarkUp/Forms/wiki/Bind_function

Uli: We had some wiki text. It's not hard to do

Nick: There are some problems.
... it implicitly creates dependencies

Uli: The dependency engine has to be aware of the bind function?

Nick: The text of the dependency engine says it only looks at arguments
... that would be hard to get right
... it would change the core spec, not only the functions spec

Erik: There are many places where you could use bind()
... it has many uses
... if it is to hard to revise the dependency text, we can say it is undefined

Uli: The instance() function surely has the same sort of problems

Nick: Except it returns the root node, which doesn't change

Erik: and a structural change gives a rebuild
... but I agree about the comparison between build and instance
... and there are problems with allowing it on bind refs.

Steven: Does anyone implement this yet?

Erik: We do, but we ignore the dependency aspects

Steven: So it has no effect on the dependencies?

Nick: We use it, but you have to be careful using it, it effects the ordering of binds

a/effects/affects/

Steven: So the ordering of binds becomes important, and not interoperable?

Erik: We should ban it on bind@ref

Nick: Also dangerous for a recalculate, since ti won't be automaitcally recalculated.

Erik: right

<ebruchez> <bind ref="foo" calculate="bind('other-bind')"/>

Uli: You could statically analyse xpath expressions

Erik: The spec has never required that, that would be really new

<ebruchez> <bind ref="bind(contact('other', id))">

<nvdbleek> <bind ref="foo" calculate="bind(../foo)"/>

Nick: That is a counter example to static analysis

Uli: So disallow it on bind expressions

Nick: Use vars?

Erik: If we prohibit it, then implementations can't even try to fix it.

<ebruchez> "If the function is used in a model binding expressionXF the XForms Processor should terminate processing after dispatching the event xforms-binding-exception to the model. If the function is used in a computed expressionXF the XForms Processor should terminate processing after dispatching the event xforms-compute-exception to the model."

Nick: We should watch out for adding new features to the last-call spec
... we need two interoperable implementations, and be comfortable with the definition

<ebruchez> (the above is about the valid() function)

Erik: We could use similar text to that from valid(), see above
... Note the "should" there
... sometimes the spec is too strict I find

Steven: value="bind('constant', xpath)"

<unl> value="concat(bind('a'), bind('b'))"

Uli: You also want to use it deep in an expression

<ebruchez> calculate="foo + bind(''bar)"

Steven: I think we need to work on this feature more before we add it to the spec. It is now too late for the LC spec

Uli: I don't think bind() is to hard too specify.

Steven: We have a time constraint.

The mediatype Element (for upload)

http://lists.w3.org/Archives/Public/public-forms/2013Jan/0024.html

<unl> http://lists.w3.org/Archives/Public/public-forms/2013Feb/0003.html

Uli: These things are two different things; I propose we change the attribute name to "accept"

Nick: I made an error. I should not have deprecated the mediatype element.

Erik: We need to fix that mistake.

<scribe> ACTION: Nick to fix the upload/mediatype deprecation error [recorded in http://www.w3.org/2013/02/06-forms-minutes.html#action01]

<trackbot> Created ACTION-1929 - Fix the upload/mediatype deprecation error [on Nick Van Den Bleeken - due 2013-02-13].

Uli: They should also have different names
... since they are different things.
... It would be nice to reduce some markup bloat

Erik: I would like to use attributes instead of nested elements
... I'm not against moving to attributes.

Uli: There are more places where we could do that
... we should do it at all places where possible

Erik: I thought I sent out a proposal for this last year

<ebruchez> http://lists.w3.org/Archives/Public/public-forms/2012Feb/0030.html

Uli: I shall reread the spec for all places where we could do this.

<scribe> ACTION: Uli to propose all places where child elements can be replaced with AVT attributes [recorded in http://www.w3.org/2013/02/06-forms-minutes.html#action02]

<trackbot> Error finding 'Uli'. You can review and register nicknames at <https://www.w3.org/2005/06/tracker/xforms/users>.

<scribe> ACTION: Ulrich to propose all places where child elements can be replaced with AVT attributes [recorded in http://www.w3.org/2013/02/06-forms-minutes.html#action03]

<trackbot> Created ACTION-1930 - Propose all places where child elements can be replaced with AVT attributes [on Ulrich Nicolas Lissé - due 2013-02-13].

Uli: What about accept attribute on upload?

<scribe> ACTION: Uli to propose text for @accept on upload [recorded in http://www.w3.org/2013/02/06-forms-minutes.html#action04]

<trackbot> Error finding 'Uli'. You can review and register nicknames at <https://www.w3.org/2005/06/tracker/xforms/users>.

tracknot, help

trackbot, help

<trackbot> Please see <http://www.w3.org/2005/06/tracker/irc> for help.

<scribe> ACTION: Ulrich to propose text for @accept on upload [recorded in http://www.w3.org/2013/02/06-forms-minutes.html#action05]

<trackbot> Created ACTION-1931 - Propose text for @accept on upload [on Ulrich Nicolas Lissé - due 2013-02-13].

[ADJOURN]

Summary of Action Items

[NEW] ACTION: Nick to fix the upload/mediatype deprecation error [recorded in http://www.w3.org/2013/02/06-forms-minutes.html#action01]
[NEW] ACTION: Uli to propose all places where child elements can be replaced with AVT attributes [recorded in http://www.w3.org/2013/02/06-forms-minutes.html#action02]
[NEW] ACTION: Uli to propose text for @accept on upload [recorded in http://www.w3.org/2013/02/06-forms-minutes.html#action04]
[NEW] ACTION: Ulrich to propose all places where child elements can be replaced with AVT attributes [recorded in http://www.w3.org/2013/02/06-forms-minutes.html#action03]
[NEW] ACTION: Ulrich to propose text for @accept on upload [recorded in http://www.w3.org/2013/02/06-forms-minutes.html#action05]
 
[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.137 (CVS log)
$Date: 2013-02-06 17:01:01 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.137  of Date: 2012/09/20 20:19:01  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

Succeeded: s/TH/Th/
Succeeded: s/TH/Th/
Succeeded: s/bond/bind/
Succeeded: s/SO/So/
Succeeded: s/ti/it/
Succeeded: s/TH/Th/
Succeeded: s/ us / use /
Succeeded: s/ to / too /
Succeeded: s/atti/attri/
Found Scribe: Steven
Inferring ScribeNick: Steven
Default Present: nvdbleek, Steven, ebruchez, unl, +1.443.837.aaaa, Kurt
Present: nvdbleek Steven ebruchez unl +1.443.837.aaaa Kurt
Regrets: Philip
Agenda: http://lists.w3.org/Archives/Public/public-forms/2013Feb/0007
Found Date: 06 Feb 2013
Guessing minutes URL: http://www.w3.org/2013/02/06-forms-minutes.html
People with action items: nick uli ulrich

[End of scribe.perl diagnostic output]