Forms WG FtF, Cambridge, MA, USA, Day 1 -- 5 Nov 2007

<Steven> Agenda: Agenda: http://lists.w3.org/Archives/Public/public-forms/2007Nov/0027

<John_Boyer> Agenda: http://lists.w3.org/Archives/Public/public-forms/2007Nov/0030.html

<raman> Morning Gang!

<Steven> Hey Raman!

<Nick> morning

<raman> At the TAG meeting, but hoping to drop in once in a while if you will let me in

<John_Boyer> Hey Raman!

<John_Boyer> certainly

<raman> will probably show up post-lunch

<John_Boyer> This afternoon will actually be working out the last few technical issues for 1.1 CR. This morning is 1.2...

<John_Boyer> Unless the plan changes

<raman> can you swap -- I'll get bored with 1.1 issues and end up playing sudoku instead ...

<raman> when are we meeting voice/mmi? would like to be there for "old-time's" sake

<John_Boyer> Later this afternoon. From 4pm to 5

<Charlie> * hi Raman!

<John_Boyer> Will meet again tomorrow, focused on MMI, too

<raman> Hi Charlie!

<Charlie> * today is everything but events...tomorrow is events with MMI

<John_Boyer> Let me talk to everyone about swapping things, Raman.

<raman> Thanks John!

<John_Boyer> Hi Raman, we'll talk about 1.1 now, and then 1.2 in the afternoon. Look forward to seeing you again, Raman!

<raman> thanks John!

<raman> MarkB is probably shuddering now that I wont be playing sudoku when I show up.

<Steven> Scribe: Steven

Issue 87/Schema problem

<John_Boyer> http://htmlwg.mn.aptest.com/cgi-bin/xforms-issues/Model?id=87;user=guest;statetype=1;upostype=-1;changetype=-1;restype=-1

John: We now say that all schema are applicable to all the instances
... without saying what 'applicable' means
... and that's the problem
... There are two kinds of schema
... 1) schema+NS+structure+datatypes
... 2) type library
... without any structure bits
... This second kind, which Leigh says that people have been doing for years,
... is a problem for the default processing for an XML Schema processor
... which by default does a strict assessment
... which looks for declared elements, and does a check based on that
... but a type library causes it to flag an error on the root element
... and then continues with lax assessment
... (the schema spec says that processors MAY do this)
... So this text on the screen attempts to explain what a Forms processor should do

Steven: And this is not a problem with bringing oin off-the-shelf processors?

John: No. And others have to solve this same problem anyway
... but Erik doesn't like this because others do it in other ways
... so do we define it this way, or add an attribute to control how it is done?

Charlie: This would be an attribute on the instance?

John: Yes

Charlie: You don't want to expose the standard author to this sort of stuff
... so I favour what is proposed in the spec

<John_Boyer> For each namespace, a test is performed to determine whether to instruct the XML Schema processor to perform strict assessment or lax assessment for items in that namespace. If there is no schema for a namespace, or if there is and it contains no top-level xs:element or xs:attribute declarations, then lax assessment is performed for each element or at

<John_Boyer> tribute in that namespace, and the test is applied to any children in other namespaces. If the schema for a namespace contains at least one top-level xs:element or xs:attributedeclaration, then strict assessment is performed for all elements and attributes in that namespace.

<John_Boyer> This is the text from Leigh

<John_Boyer> Erik's proposal was <instance validation="lax"> Default of strict

John: Will the author of a type library understand the strict/lax stuff?

Steven: Not necessarily; anyway, the user of such a library surely won't in general
... Is there no way for the schema to say that it is a type library, therefore lax?

John: No
... Leigh described a different implementation strategy
... some implementations build up a master schema from all available schemas
... and they somehow know when a schema is a type library
... (though neither Leigh nor I know how they do that)
... The problem with the solution proposed that I don't know how to create the 'super-schema' that can mark the type libraries as lax
... We will revisit schema validation in the future for future versions

Steven: We could simplify the text to just say "Data type library schemas should be vlidated in lax mode"
... and leave it to the processor to decide how

Keith: Why not delete the first and last sentences, wouldn't that be enough?

John: Yes, we could reword

<John_Boyer> If the schema for a namespace contains at least one top-level xs:element or xs:attributedeclaration, then strict assessment is performed for all elements and attributes in that namespace. If there is no schema for a namespace, or if there is and it contains no top-level xs:element or xs:attribute declarations, then lax assessment is performed for e

<markbirbeck> My IRC client has a weird bug...I am not getting any messages in the forms window, even though XHTML and RDFa windows work fine!

<John_Boyer> ach element or attribute in that namespace.

<markbirbeck> I type...nothing appears. :)

we see it

Steven: Can't you use "otherwise" for the second sentence to make it shorter

John: Yes! Then I can paste it into irc in one go! Watch this:

<John_Boyer> If the schema for a namespace contains at least one top-level xs:element or xs:attributedeclaration, then strict assessment is performed for all elements and attributes in that namespace. Otherwise, lax assessment is performed for each element or attribute in that namespace.

Steven: "10 01If the schema for a namespace contains at least one top-level xs:element or xs:attributedeclaration, then strict assessment is performed for all elements and attributes in that namespace, otherwise, lax assessment is performed"

<markbirbeck> Does that mean that if I define a type for an integer being between one and a hundred, the whole document gets validated?

John: We may never take the XSLT 2 route to control the applicability of schema

<markbirbeck> Oh...no...take that back. I see.

John: Especially since we may use more schema types than just XML Schema
... in the future

[Michael SperbergMcQueen, Schema Expert, enters room]

MSM: What you are running up against is not a problem with your imports, but elsewhere
... The schema spec is trying to do you a favour
... when a validator is invoked, the spec tries to stay out of negotiations between invoker and validator
... so you can start validation in a number of ways, and 1.1 makes it clearer
... you identify either an element or attribute and a schema
... and then validate in a number of ways
... with an element declaration, with a datatype
... and other ways
... the spec doesn't require any one of those ways
... it depends on context
... what you can do is say that validation needs to start in a particular way
... when you are matching lax, if you don't find a declaration, a 1.0 processor can move on without doing anything OR may fall back to lax processing
... so you can skip the subtree, or to process a number of levels, or process all the way down laxly
... So I think the validator that complains in the use of datatype libraries, is doing the "skip the subtree" step

[Discussion of section 5.2 in Schema spec, Assessing Schema Validity]

[John shows an example on the screen for MSM, showing all the schema usage possibilities in XForms]

MSM: What you need to say is that lax processing occurs, with recursion
... because in lax mode, once you have a declaration, you get into strict mode

[walk through the XML Schema spec at this point]

http://www.w3.org/TR/xmlschema-1/#validation_outcome

http://www.w3.org/TR/xmlschema-1/#cvc-assess-elt

[MSM shows that when a declaration is found the element gets processed properly]

John: This is fantastic. Eric has been trying to persuade us that this was the case, and Leigh and I didn't believe it

MSM: So some elements will be marked as validity checked, but validity unknown

Steven: So we are OK, *unless* the root element is wrong
... and then the error won't be given
... since we start off lax

John: Thank you so much Michael

[Michael leaves]

<John_Boyer> Hi Leigh and Raman, just got to a late break...

restart

<John_Boyer> The schema definitions for a namespace ar applicalble to instance nodes based on initializing processing in 28<21termref20 def28="def-lax-processing28">lax mode28</21termref28>. Note that schema processing for nodes with matching schema declarations is governed by its content processing definition, which is strict by default.

<raman> look for me at the lunch break -- so one of you can pick me up to join the forms meeting after lunch.

[Leigh discusses aspects of new decision]

Leigh: There are variations of Steven's problem; validity checking doesn't kick in until you find an element with a declaration

John: What we actually want doesn't seem to be supported by existing Schema engines

Leigh: The way to do it is like XSLT 2

John: Yikes

<John_Boyer> <xs:schema targetNamespace="&a;" ...>r structuresr declares e, but does not declare badness as a childr f is not declared as top levelr</xs:schema>

<John_Boyer> <xforms:instance id="Y">r <data xmlns="">r <f xmlns="&a;">...</f>r </data>r</xforms:instance>

John: In this example, this instance will not be found to be invalid

<klotz> Net answer to your question: conforming processors can be written to validate any element you like. Not all processors need provide this service: buy or use processors that validate the information you need validated. By the way, the detailed rules give the processor a choice of validating the element against some particular identified element declaration, some particular identified complex type, or to use the mechanisms of strict, lax etc. to determine what to

<klotz> validate based on what declarations happen to be available. All of this is explained at xmlschema-1.

<klotz> from http://www.schemavalid.com/faq/xml-schema.html#d4

Leigh: We need to adress these questions. The validation only occurs at submission time
... if there is no schema definition then we don't validate, and if there is, we go strict

John: You have to write a schema if you want schema validation

Leigh: I'm ok with all this. If we want more stuff, let's do it in 1.2

s/aql//all/

s/alll/all/

John: Proposed wording is above

Nick: But there is a bit in schema that is a MAY and it should be a MUST for us

Leigh: Use the wording from XSLT 2
... that makes it required

<Nick> http://www.w3.org/TR/xslt20/#validating-using-validation-attribute

Leigh: Yes, that

<Nick> In the case of an element, a top-level element declaration is identified whose local name and namespace (if any) match the name of the element, and schema-validity assessment is carried out according to the rules defined in [XML Schema Part 1] (section 3.3.4 "Element Declaration Validation Rules", validation rule "Schema-Validity Assessment (Element)", clauses 1.1 and 2, using the top-level element declaration as the "declaration stipulated by the processor", whi

<Nick> y to contained elements and attributes to the extent required by the schema definition.

<Nick> When the parent element lacks a declaration, the XML Schema specification defines the recursive checking of children and attributes as optional. For this specification, this recursive checking is required.

"Note:

When the parent element lacks a declaration, the XML Schema specification defines the recursive checking of children and attributes as optional. For this specification, this recursive checking is required."

John: Is that normative?

Steven: I think it explains the text above

John: OK. So we make that text normative in our spec
... Like this:

<John_Boyer> The schema definitions for a namespace are applicable to instance nodes based on initializing processing in 28<21termref20 def28="def-lax-processing28">lax mode28</21termref28>. When an element lacks a schema declaration, the XML Schema specification defines the recursive checking of children and attributes as optional. For this specification, this recursive chec

<John_Boyer> king is required.

<John_Boyer> Also, a note immediately following to clarify the following:

<John_Boyer> Note that schema processing for nodes with matching schema declarations is governed by its content processing definition, which is strict by default.

<John_Boyer> Finally, a note in revalidate event to say that "applicable" schema definitions is as defined in the above content (which is in the description of the model element)

[Lunch]

<raman> TAG still tagging along -- not sure when we break for lunch

<John_Boyer> returning now

<John_Boyer> scribe: wellsk

John: 1st question -- go with last XML schema discussion change?

Steven: will we have unavoidable behaviors?

<John_Boyer> Proposed Resolution: Validation starts lax; schema applicability defined as above. Note in revalidate event referencing schema applicability statements

Steven: do we have enough members to make decision?

John: yes -- thanks to Michael

<John_Boyer> Also Leigh agreed before break, and this is what Erik proposed originally

John: Is everyone ok with statement.

everyone: yes

RESOLUTION: Validation starts lax; schema applicability defined as above. Note in revalidate event referencing schema applicability statements

<scribe> ACTION: John_Boyer to amend spec to resolve issue 87 [recorded in http://www.w3.org/2007/11/05-forms-minutes.html#action01]

<trackbot-ng> Sorry, couldn't find user - John_Boyer

<Steven> trackbot-ng, list users

<Steven> trackbot, list users

<Steven> trackbot-ng, help

<trackbot-ng> See http://www.w3.org/2005/06/tracker/ for help (use the IRC bot link)

John: next topic: tweak insert action
... recent changes to insert damaged last call issue (Eric)

<Steven> trackbot-ng, status

<Steven> ACTION: jboyer to amend spec to resolve issue 87 [recorded in http://www.w3.org/2007/11/05-forms-minutes.html#action02]

<trackbot-ng> Created ACTION-430 - Amend spec to resolve issue 87 [on John Boyer - due 2007-11-12].

John: Appendix B.2 parent of new element
... insert attribute, or child element, need to use (Eric) nodeset attribute
... not using nodeset attribute, need to use context for that

if nodeset empty, context is used to insert parent context node

scribe: identify parent to identify attribbute from item
... insert attribute before/after element

last node identified by nodeset

scribe: eliminated homogenous nodeset problem
... disputing note for using the context attribute
... to not be using nodeset

example is incorrect due to processing model for insert

scribe: looking at insert processing
... what nodes copying, what nodes cloned -- they are from origin
... here is where you put the newly cloned data
... struck (e) lines in spec

other last call comments -- don't impose attribute order

raman joins us

On insertion, removed para, but added (a), (b) about when cloned node is an attribute

eric suggests remove (a), rely on part (8) target location due to node type conflict

if type conflict, noop anyway

consistent view of context attribute for insert

to identify element whose content space is used to insert something

nodeset is used to determine before and after for insertion

if special case (a) is removed

insert with nodeset and origin -- will perform no operation because it is attempting to insert before or after of content

correct way for example is to use context

Nick: should there be a testcase

John: several areas for step 8 to look at

Nick: implementations may have done this, so test case needed to catch

John: so we have to decide on this example

Charlie: not sure

John: we do something like this in other places

Raman: look at it more as data structure rather than XML
... changes perspective
... data structures more interesting than XML structure

John: data model is built on data model

Charlie: a runtime error instead of noop

John: not an exception
... we could
... is binding exception

Nick: future feature, exploit 2.0 dispatches in runtime error

Charlie: why I was skeptical

Nick: remove nodeset

John: change nodeset to context
... have to specify context, because no nodeset is a noop

Nick: write context dot

context=dot(.)

Charlie: ship it

John: reword

<John_Boyer> Resolve to eliminate special case target location processing for attributes and fix Example B.4

Nick: different versions of the spec
... b becomes a

John: not had to make a separate version yet

Proposed Resolution: Resolve to eliminate special case target location processing for attributes and fix Example B.4

<scribe> ACTION: JBoyer to do eliminate special case target location processing for attributes and fix Example B.4 [recorded in http://www.w3.org/2007/11/05-forms-minutes.html#action03]

<trackbot-ng> Created ACTION-431 - Do eliminate special case target location processing for attributes and fix Example B.4 [on John Boyer - due 2007-11-12].

RESOLUTION: Resolve to eliminate special case target location processing for attributes and fix Example B.4

John: issue related to input mode with Martin
... what to do to go to CR for this issue
... waiting for update to list
... go to CR and modify list later
... go to CR and then modify afterwards to modify the input mode issue

Steven: yes

John: other than input modes issue all last call isues dealt with
... 1000 diffs

446 diffs

from last call document

approaching 100 diffs

John: process issues tomorrow
... resolve to exit last call

Raman: do it done

Steven: not appropriate until we have the spec

not informative

Raman: make it an informative appendix

Proposed Resolution: make input modes an informative appendix

<scribe> ACTION: make input modes an informative appendix [recorded in http://www.w3.org/2007/11/05-forms-minutes.html#action04]

<trackbot-ng> Sorry, couldn't find user - make

<scribe> ACTION: JBoyer to make input modes an informative appendix [recorded in http://www.w3.org/2007/11/05-forms-minutes.html#action05]

<trackbot-ng> Created ACTION-432 - Make input modes an informative appendix [on John Boyer - due 2007-11-12].

John Scanning diffs 1.0 and 1.1

downside marking feature at risk -- opens comments

John: saying no featrures at risk, could end up with last call

RESOLUTION: make input modes an informative appendix

Steven: issue from Madrid local-date-time() vs now()
... we didn't make a good decision
... not obvious for authors

John: not new information

Steven: several people approached Steven on now()

John: end-of-day had to pick a way
... recorded a (Resolution) on that

UTC version of now()

clarifying now, examples for locat-date-time, adjusting to time zone

John: talk about process issues tomorrow for 1.1

John, go to next agenda -- xforms 1.2

John: 1.2 to be author simplification

Are we under a requirement for a 1.2?

Raman: better handle on what we need

<John_Boyer> for lazy authoring

Raman: xhtml 5 has xforms, clean bridge for lazy authoring
... how to integrate into xhtml 5 spec

Charlie: what does it do to write guidelines? what is example?

Nick: charter agreement differences on guidelines
... Mark says simplification is important
... guidelines to enable transition from xhtml to xforms

Raman: continue to have traction for xforms

Nick: experimenting with expression language

Nick/Raman: xpath expressions, implementing xpath

John: where would JavaScript live?

Raman: expression to expression, conformance
... xpath processor

John: SMIL -- referencing engine pluggable module

Raman: keep xpath as expression language

Nick: to support xpath 2.0
... no harm to use JS in XForms document

John: what is syntactic expression in JS vs xpath

Charlie: guidelines to xhtml forms

<John_Boyer> Need to get the same content to work, which isn't easy with a "pluggable modules" architecture

Nick: model in html forms
... web forms having a model

<John_Boyer> If content A contains XPaths and content B contains JS expressions, then implementations need to have both modules available

Nick: have model and check datatypes,

Raman: ends up looking like xforms data model

Nck: is it another impl issue?

Raman: seeing model, independent model is enough

Nick: if date is assigned to node -- is it auto a date control?

Charlie: hope to not introduce an explicit model

John: start with syntax, all design is one glass, more control over submission, so xforms submission shows up, and then declare data instead of implicit
... describe data model
... doing it by plugging in what you want to use
... we have data model

raman: bind in a data structure not just xml data

John: isn't that what Charlie said a minute ago?

Charlie: worried about overlapping soln

John: point to our charter pointing out what we are supposed to do.

Charlie: alternate instance formats, not lazy authoring, opposite of ;lazy authoring, plug n' play

Raman: authors likely to not notice difference

Charlie: 1.2 authoring, 2.0 -- modularity

John: pluggable engine: xpath 1.0/xpath 2.0

Charlie: XForms 2.0 take on xpath 2.0 features
... componentization -- cleaned up stack for developers/architectural value
... dealing with stuff we need to add, but need to get to other issues like comonentization

Nick: why not specify readonly on UI level
... why not specify datatype on UI?

not to much work for 1.2

scribe: scope of 1.2 vs 2.0

John: not achieve charter unless we do both 1.2 and 2.0 in parallel

Parallel XForms vs sequential for 1.2 and 2.0

Nick: Is it limitation for 1 editor?

John: 1500 diffs since 2003

Raman: much more solid

John: we have to go parallel if we are to meet our charter

Raman: we have to keep technology moving

John: having more editors on parallel work would make Editor be easier.

Raman: perhaps editing smaller specs is easier than larger specs

John: good guidelines for what goes on in these specs.
... merge recalculate/revalidate

<John_Boyer> has been discussed for 2.0 but it's not modularization

<John_Boyer> it may need to be 3.0

Raman: getting 2.0 early would hurt xforms 1.2 and 1.1

John: 1.1 has text submission
... other text support

Leigh: backplane work by Charlie, floundered on W3C
... parallel specs
... implementors think hard about how to take part in this
... where to get visibility, work like Mark has done

Raman: important to communicate

Leigh: happy with what raman/charlie/john have said

raman: worked because it was subdivided out -- divide editoring job

John: structure/submission chapters split
... reorganizing, everyone might see everything is in right buckets
... instance module with defined interface paying attention to that
... modularization -- XF 2.0

Raman: modularization is one thing must be done, but don't put it in a specific version.

Leigh: support JSON

Raman: data model on JSON

Leigh: Google made announcement, who is doing xforms for platform

Raman: entire platform is open source

John: have to do modularization to do the things we want to do with XF 2.0

Raman: what are we specifying
... and what are interoperable pointsw

Steven: XSmiles combines the two bits together well
... demo from SYMM group tomorrow

Charlie: looking for xforms technology into other areas, SMIL, etc

Raman: do spec as series of blogs and investigations

Charlie: what is difference for work for backplane XG vs XForms 2.0?

Raman: Looks like XForms WG selling data model to everybody else.
... look at google mashup editor

break time

<John_Boyer> Agenda: http://lists.w3.org/Archives/Public/public-forms/2007Nov/0032.html

<John_Boyer> Break, then discuss "Read Access from WAF"

John: for 1.2, for 2.0 have list of items for people to have to do

Nick: for XForms 1.2 allow properties on UI level

<John_Boyer> Leigh, will you be back to discuss Read Access?

Topic -- read access from WAF http://lists.w3.org/Archives/Public/public-forms/2007Oct/0002.html

<John_Boyer> Restart

the link: http://www.w3.org/TR/access-control/

John: associate access control of resources
... resource on server, can't trust client

Nick: client says what is done with it?
... different domains, this is trying to solve this
... changing domains

like firefox does when domain changes

Nick: trusted domains, so can info from the access control
... domain added to trusted list

John reading: When a user agent performs a request to which the access control mechanism applies it performs an access request.

John: what are they defining for user to access reosurce?

Nick: multiple frames, multiple documents, in Javascript -- not allowed. this is trying to solve problem. Trust other domains
... they can receive data for me ...

John reading: When a user agents has to make an access check for a particular resource it must then associate the following with that resource:

Nick reading: in introduction, To prevent information leakage, user agents, such as Web browsers, implement a same origin policy that allows a document (e.g. some JavaScript) to read, process, or otherwise interrogate the contents of another resource if, and only if, the other resource resides in the same domain.

John: an httpRequest

not necessarily from JS

Nick: two different resources, browsers allow listeners to allow communicate, this spec is trying to solve this problem.

Charllie: other foreign resources can react

Nick: javascript or another lang that has access to the DOM
... to change frame, that is not the problem

<Steven> http://www.w3.org/TR/2007/WD-access-control-20071001/

John reading: When a user agent performs a request to which the access control mechanism applies it performs an access request. The exact details of an access request such as how to deal with network errors, redirects, et cetera are out of scope this specification and must be defined by specifications using the access control mechanism

John Reading: Namely, requests using a non-GET HTTP method must be preceded by a request using the GET HTTP method.

Nick: web page consists of multiple resources, how to get with other means than a get?

John: have to be able to post to servlet, you filled out that form correctly, now fill in the next form

client doesn't know details of this

scribe: web service calls to other domains

Raman: JSON authors are not doing this correctly

John: Leigh give us understanding

<John_Boyer> of what is expected of us

Leigh: if we can give them feedback less tied to html

take on xforms submission

scribe: to make that comment to WAF

John: do we provide more formal format?

Leigh: negotiate who does work to make their thing work
... shared goal, a way of doing it, hacked together
... if they do like privacy people did
... applicable to XForms

John: account for XForms 1.1 submission, do we account for ways to submit data?
... nonget resource from trying to 'get" resource. If can't use get, have no business trying to get

Leigh: response header with particular type of request, take XForms into account
... like privacy work

John: what is next step?

Leigh: direct a message
... please put xforms in
... works well with other specs that are out

<scribe> ACTION: Klotz to write something for the group [recorded in http://www.w3.org/2007/11/05-forms-minutes.html#action06]

<trackbot-ng> Sorry, couldn't find user - Klotz

<scribe> ACTION: klotz to write something for the group [recorded in http://www.w3.org/2007/11/05-forms-minutes.html#action07]

<trackbot-ng> Sorry, couldn't find user - klotz

<John_Boyer> don't worry about not finding user klotz

<John_Boyer> Nick can find it

John: when Nick read spec -- access by arbitrary Javascript commands
... is it something more than a network request
... Leigh to write up, get with Nick and Raman and discuss, then get back to group with comments

<Steven-mobile> ACTION: charlie to circulate XG draft charter to VBWG and Forms WG [recorded in http://www.w3.org/2007/11/05-forms-minutes.html#action08]

<trackbot-ng> Sorry, couldn't find user - charlie

<Steven-mobile> Trackbot-ng, status

<Steven-mobile> trackbot-ng, status

<Charlie> * ping

SMIL feedback

http://lists.w3.org/Archives/Public/public-forms/2007Oct/0108.html

<Charlie> Scribe: Charlie

SMIL 3.0 feedback

Issue 8, how would authors receive notification if nodes are deleted?

John: given no full XForms data model, nodes can be deleted without any notification
... but nodes when inserted generate events, not clear why not parallel function for delete case
... they have decided to add delete action, but not corresponding event
... so, for example, why have events on inserting? they're either important/useful or not
... furthermore, index management in repeat module is now in terms of listening for insert/delete events
... both would be required...this module could be used independently if we factor it out
... not sure they have equivalent constructs, but for us we need both to properly manage anything operating over the data model

Issue 4: good

Issue 5, will add new attribute to specify position of new element

Issue 13: have picked up target attribute from submission

Charlie: can this be done coherently without the rest of the data model?

John: yes, parts of the model can be taken independently, perhaps with restricted (default) function...target seems to be one

Issue 3: are attributes supported

Issue 7: no bubble but do they have a capture phase? is this some level of DOM eventing or another processing model?

Issue 9: multiple contentControlChange events

John: why fire both? which comes first?
... we've had similar problems where as you're firing events want to fire two that are related to the same activity
... what happens if you dispatch first one, its action handler makes 2nd event stale...no longer relevant?
... maybe doesn't apply in SMIL???

Issue 15: submission is synchronous

John: doesn't SMIL have particular need for async?

Charlie: seems like the fork/join case handles this

John: would pre and post processing of submission be possible?
... is there a need for control over submission serialization or is this future requirement?
... perhaps by adopting more of xforms submission in the future

- DRAFT -

Forms WG FtF, Cambridge, MA, USA, Day 1

5 Nov 2007

Attendees

Contents

Issue 87/Schema problem

SMIL 3.0 feedback

Summary of Action Items

Scribe.perl diagnostic output