See also: IRC log
<wellsk> Good Morning Everyone -- is there a tetlphone conf set uo yet?
<John_Boyer> Not yet
<wellsk> Ok -- will monitor -- but I am awake and here!
<John_Boyer> OK, thanks keith
<John_Boyer> It's very early...
<wellsk> 3:00 here on EST
<John_Boyer> You should maybe get a few hours more sleep and join us in the afternoon...
<wellsk> It'll be ok... I am an early bird anyways :)
<John_Boyer> Are you able to use skype?
<John_Boyer> Keith
<wellsk> no, but I may be able to use Sametime voice suite -- I have international permissions
<klotz> wellsk?
<wellsk> yes
<klotz> ah thought you might be at home but you are on an ibm network and can' t use skype?
<wellsk> yes -- I do not have skype installed -- but should be able to hook in though the IBM intranet using their sametime tooling with voice suites
<wellsk> I'll give a shot
<John_Boyer> scribe: Nick
<John_Boyer> scribenick: nick
<Steven> scribenicknick: nicknick
John: There is an XForms 1.1 problem in how repeat works
<John_Boyer> ACTION: John to Fix descriptions of repeat index update on xforms-insert and xforms-delete for XForms 1.1 [recorded in http://www.w3.org/2008/06/09-forms-minutes.html#action01]
<trackbot> Sorry, amibiguous username (more than one match) - John
<trackbot> Try using a different identifier, such as family name or username (eg. jkugelma, jboyer)
John to fix XForms 1.1 wording for repeat processing to associate repeat index updating with capture phase on containing model, not target instance [recorded in http://www.w3.org/2008/02/05-forms-minutes.html#action01]
John: We have work on XForms 1.1, I added the
people who have actions on it to the agenda
... There are a couple of reviews too
... We need to decide as a group which specs we want to produce for XForms
1.2
... We have a message module, maybe we can go through it and see how we can
produce an instance module
Erik: What does the instance module do for us? Is it a requirement, does it solves problems?
Steven: We will do modularization to ensure
better consistency and compatibility
... We could even just do XForms 1.2 as a module on top of XForms 1.1
Charlie: It enables the integration of XForms modules in toolkits like dojo
<Charlie> perhaps using non-xml data, e.g. JSON
<John_Boyer> Nick: Asks Steven about doing a separate module for "implied submission"
<John_Boyer> Nick: Can you change the behavior of an attribute (e.g. whether it is required or optional)
<John_Boyer> Steven: You would modularize the submission attributes, then create a submission element module that consumes them in one way
Steven: You don't have to do them as separate specs
<John_Boyer> Steven: and an implied submission module that consumes them differently
klotz: How do you say which modules an implementation supports
Steven: so XForms 1.2 could be XForms 1.1 and
an extra set of modules
... There is a dtd, schema version and outside W3C there is a Relax NG
version
Klotz: There is sourceforge project that is an XForms 1.1 relax NG schema
Steven: Ideally we would modularize XForms 1.1 and create some extra modules for XForms 1.2
John: How many specs do we end up?
Steven: That is up to us, one spec or two specs (core spec and the XForms 1.2 addition modules)
Nick: How much work is it
Steven: XForms 1.1 is more or less
modularized
... It needs some work, but is partially already in there
Erik: What does it mean
Steven: You specify which elements have
extension hooks (elements and attributes)
... What you do you write it in a schema natural way (English) and then
generates schema or dtd
Nick: Is there a requirement to generate a dtd or schema?
Steven: No generating a dtd or schema isn't required
<John_Boyer> http://lists.w3.org/Archives/Public/public-forms/2008Apr/0081.html
<John_Boyer> http://code.google.com/p/lib-xh/wiki/MessageModule
John: Where do you put the MIP-events they are dispatched to UI-controls but always dispatched by the model , do you put those in the module-module or the ui-controls-module
Erik: We should fix the MIP-events to make them
usable
... We could add new events that are usable in a separate module
John: when you rewire controls we should send an initialization event
Charlie: If you remove the events from the module-module then you still needs something to communicate with the module
Steven: I don't see the problem with events,
javascript is also based on events
... DOM events is the base of the javascript events
John: The module should send xforms-model-construct, xforms-model-construct-doene, xforms-ready, xforms-refresh
Klotz: if we have a module that only does the data (no submission, no UI) xforms-model-construct-done signifies when other modules can start working on the module
Steven: You have all these modules and a driver that decides how these modules work together
Klotz: do we define the communication by events or as IDL?
John: I have users that use the the xforms-model-construct-done to toggle switches, ...
Steven: all w3c languages are events based
Klotz: maybe we should do an extra level of abstraction
Charlie: We could define the events as informative, and not doing the work in the default handler
Charie: We define an IDL that dispatches those informative events but does the work itself (not when handling the events)
klotz: it will enable the possibility for extra optimizations
<John_Boyer> http://www.w3.org/TR/2003/REC-xforms-20031014/index-all.html#evt-modelConstructDone
klotz: we should make events informal, and create an IDL and the communication between the modules is done by the IDL
John: The instance module would contain the instance element, the setvalue, insert and delete actions
Steven: There is real value that XForms 1.2 is XForms 1.1 and this extra module
John: where is behavioral change
klotz: that all events are informal
John: we already hacked the xforms-model-construct
Charlie: ...
Klotz: we are missing cancel then (if we aren't doing the processing in the default action)
John: you could cancel submission
klotz: we need to have some events that invoke the IDL method
John: In XForms 1.1 the repeat related stuff is put in the repeat module, the submission stuff to the submission module, the only thing that maybe went wrong that the actions went to an action module
klotz: It isn't wrong, the actions are just an interface on the module
<John_Boyer> break time
<wellsk> brb
<John_Boyer> back
John: Where do we go from here, we will stick with one spec, we need to rewrite it a bit
Steven: I would propose the get all elements and attributes and do a rough division
John: We can set a bit on an event to cancel the event
Nick: I don't like that, we should use XML events to cancel it
John: Why would the events be in the actions module? We have a lot of events that aren't triggered by actions
klotz: what do you need to change if you want to switch to XML events 2.0 (for example bcz. we are modularized they just change the version of XML events they use)
Charlie: it is just plug and play
<Steven> http://www.w3.org/MarkUp/Drafts
<Steven> http://www.w3.org/MarkUp/Drafts
<unl> xml events 2: http://www.w3.org/TR/2007/WD-xml-events-20070216/
http://www.w3.org/TR/2007/WD-xml-events-20070216/
<Steven> http://www.w3.org/MarkUp/Drafts/#xml-events2
John: How would someone catch the version of XPath that is used
Steven: They would be separate version of the
module
... They would have a data-type
<ebruchez> MS defines a mapping JSON-XML: http://msdn.microsoft.com/en-us/library/bb924435.aspx
John: explains problem with allowing different expression languages
<wellsk> (module breakdown listed in 2.0 features
<wellsk> also, wouldn't Mark's Message module posted earlier be a good starting place?
<wellsk> 1.2 modules: http://www.w3.org/MarkUp/Forms/wiki/XForms_Future_Features#head-59b64c2c07c13848069d48f8097ca1b28be16d9e
<Steven> http://www.w3.org/TR/xhtml-modularization/
Steven: is going to project some examples of modularisation
<Steven> http://www.w3.org/TR/xhtml-modularization/abstract_modules.html#s_xhtmlmodules
<Steven> http://www.w3.org/TR/xhtml-modularization/abstract_modules.html#s_intrinsiceventsmodule
http://www.w3.org/TR/xhtml-modularization/abstract_modules.html#s_common_collection
http://www.w3.org/TR/xhtml-modularization/abstract_modules.html#s_events_collection
Note that the Events collection is only defined when the Intrinsic Events Module is selected. Otherwise, the Events collection is empty.
<klotz> http://www.sagehill.net/docbookxsl/AddProfileAtt.html#AddAttRelaxNG
<John_Boyer> http://www.w3.org/MarkUp/Forms/wiki/XForms_Future_Features
Steven: Data Properties module could have the attributes type, constraint, ...
klotz: I would have thought that the
module-module would have an IDL method to ask for validity
... So data properties needs to modify the DOM to allow adding properties to
nodes
... it could be possible that the user-defined properties can only be added
by other modules
Steven: starts writing the modules on the wall
xforms
data
Instance data
type
Meta data
manipulators
logic
submissions
binding
constraint
calculate
relevant
controls
container Form controls
group
repeat
input Form controls
outut Form controls
label
messaging
<John_Boyer> lunch
<ebruchez> FYI XProc uses an xpath-version attribute to specify the version of XPath requested: http://www.w3.org/TR/xproc/#xpath-context
xforms
Model
data
Instance data
type
Meta data
Properties set, get, create
manipulators
IDL for set, get, insert, delete
bind
expression engine
logic
constraint
calculate
relevant
controls
container Form controls
group
repeat
switch
input Form controls
outut Form controls
label
messaging
submission
cross-cutting:
actions
events
IDL
elements + Attrs
expression extension functions
xforms
-Model
--data
---Instance data
----type
---Meta data
----Properties set, get, create
---manipulators
----IDL for set, get, insert, delete
---bind
---expression engine
--logic
---constraint
---calculate
---relevant
-controls
--container Form controls
---group
---repeat
---switch
--input Form controls
--outut Form controls
--
label
messaging
submission
http://www.w3.org/TR/xforms11/#structure-model-instance
http://www.w3.org/TR/xforms11/#structure-abstract
http://www.w3.org/TR/xforms11/#structure-model-instance
http://www.w3.org/TR/xforms11/#structure-abstract
<John_Boyer> http://www.w3.org/MarkUp/Forms/wiki/XForms_instance_data_module
discussion about adding setvalue, getvalue, ... to instance, there are problems if you use another expression engine (context, expressions), you also have the problem that readonly needs to be enforced which is in the logic module
john: you need to be able to override the IDL functions
klotz: maybe we need pre/post hooks
... how can we give access to the DOM interface and support read-only
... we can do this by specifying precisely what operations are allowed
<Steven> that is Mark typing
John: value change events are send to UI-controls and send during refresh, not when the instance changes
Charlie: I think we need to focus on refactoring to modules and not discuss the related new features
John: but they may change the modules
... We already have getInstance(), maybe we should shoot it
Charlie: lets modularize first then later fix
Uli: we have an XPath data model not a DOM model
John: we can use XFormsNode and define our operations
Erik: Don't create the 6th XML representation
model
... we can use XML Information Set (Infoset)
<klotz> http://www.w3.org/TR/DOM-Level-3-XPath/idl-definitions.html
in java they have http://java.sun.com/javase/6/docs/api/index.html?javax/script/ScriptEngineManager.html
http://java.sun.com/javase/6/docs/api/javax/script/ScriptContext.html
(wrong link)
klotz: explains that leaf nodes in JSON can't an object that holds the value
<markbirbeck> I'm not quite sure what you guys are talking about...but if you are discussing 'access languages' for different data models, I think it might not be the right approach.
<markbirbeck> XPath and XQuery are 'abstract' syntaxes, and the XPath data model is 'abstract' too.
<markbirbeck> So we can still access JSON data with XPath.
<markbirbeck> I could...sorry.
<Steven> :-)
<markbirbeck> The Ubiquity XForms processor uses Google's AjaXSLT library...
<markbirbeck> ...that implements XML an XSLT on browsers that don't have it natively...
<markbirbeck> ...so an XPath query there is effectlively querying against JSON.
<klotz> John asked if we could define IDL to get the references from an expression-lanuage string and have it return an object array; I said that in the JSON implementation, it would be necessary to return tuples of {hash,name}, and Uli said just returning an array of strings simply for the path.
<Charlie> not clear if that (xpath over json) is an attractive option for the current ajax community
<klotz> We've discussed that it's possible to apply Xpath to a JSON data model, and with suitable restrictions (first-node at each level, no mixed content, no attributes) it's possible to apply a JSON expression to an XML data model.
<markbirbeck> Charlie...I agree that there is a question there.
<markbirbeck> But note that lots of libraries now are letting you express CSS with XPath.
<markbirbeck> So we shouldn't assume that XPath is a 'no no'.
<klotz> So we follow YAGNI and decide not to worry about JSON while we are working on modularization today, not having sufficient examples to generalize from.
<klotz> http://c2.com/xp/YouArentGonnaNeedIt.html
charlie: lets make it work with XML first
John: what does getReferences return:
NodeList
... what does getDocument() return: DOMDocument
<klotz> I think if you implement the instance module in isolation you need DOM, but if you wrap instance up in the rest of XForms it's your choice as an implementor whether to expose it or not.
http://www.w3.org/TR/DOM-Level-3-XPath/idl-definitions.html
Erik: Is not enough for us, the context needs a
position and context-size
... Otherwise position() and last() wouldn't work
John: When repeat adds index() does it becomes available in all modules, or only in some modules
Erik: How do we define in our language which functions are allowed where?
<markbirbeck> I'm not following this idea of "which functions are allowed where".
<markbirbeck> The extension mechanism is via @functions, is it not?
<markbirbeck> So aren't the functions then available to that model?
<markbirbeck> Is there a suggestion that we need more than that?
<Steven> say it out loud Mark
<Charlie> we're in the instance module so we don't have @functions
<Charlie> unless we reintroduce it there...
<markbirbeck> You're trying to define an instance module that is independent of a model?
<markbirbeck> Isn't that pretty much the DOM? (Putting aside data model.)
<ebruchez> String positionAttribute,
<ebruchez> List collectionToBeUpdated,
<ebruchez> NodeInfo insertContextNodeInfo,
<ebruchez> List originObjects,
<ebruchez> int insertionIndex
<markbirbeck> JavaScript supports pretty much everything...including aspects, John. :)
<klotz> Aspect-Oriented Javascript: http://www.devx.com/webdev/Article/31695
<John_Boyer> end of day
This is scribe.perl Revision: 1.133 of Date: 2008/01/18 18:48:51 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: RRSAgent_Text_Format (score 1.00) Succeeded: s/instance model/instance module/ Succeeded: s/implemantation/implementation/ Succeeded: s/XML/DOM/ Succeeded: s/evenets/events/ Succeeded: s/infomal/informal/ Succeeded: s/w// Succeeded: s/infoset/XML Information Set (Infoset)/ Found Scribe: Nick Found ScribeNick: nick WARNING: Dash separator lines found. If you intended them to mark the start of a new topic, you need the -dashTopics option. For example: <Philippe> --- <Philippe> Review of Action Items Present: Nick Uli Charlie John Erik Leigh Steven Keith(remote) Rafael Roger MarkB Regrets: MarkS Got date from IRC log name: 09 Jun 2008 Guessing minutes URL: http://www.w3.org/2008/06/09-forms-minutes.html People with action items: john WARNING: Input appears to use implicit continuation lines. You may need the "-implicitContinuations" option.[End of scribe.perl diagnostic output]