Forms WG Virtual Day

01 Feb 2008


See also: IRC log


wellsk, nick, [IPcaller], ebruchez, CharlieW, +1.978.399.aaaa, John_Boyer, markbirbeck, Steven, Leigh_Klotz
John Boyer
Leigh, Steven, Nick




<nick> but to be save maybe we can allocate 12 ports...

<Steven> hi all

<wellsk> Hi Steven!

<CharlieW> Hi everybody!

<Steven> howdy

<wellsk> Hi Charlie

<CharlieW> hey steve, keith...will call in shortly...are we ok to use zakim?

<CharlieW> sorry, steven

<Steven> Yep:

<CharlieW> k

<CharlieW> great

<Steven> It's reserved for 150 mins (first half)

<Steven> I thought I'd wait for reserving the second half until we know how many are really coming

<CharlieW> sounds right

<CharlieW> i'm in dallas

<CharlieW> hotel rm

<Steven> I brought a speaker phone with me, only to discover it is 230V only.

<CharlieW> ahg

<Steven> So I'll be hanging on a regular phone all day too

<CharlieW> i just called in using the hotel speakerphone, but it's breaking up so will probably use my headset

<Steven> WHich web interface are we using, Yugaw or whatever it's called, or another?

<CharlieW> not sure...let's decide when folks get on

<Steven> ok

<CharlieW> mark and i have both sent out links

<CharlieW> not sure which is better

<CharlieW> the one i have can only originate sharing from windows, but anybody can see

<nick> Charlie you should say SameTime don't you ;)

<klotz> I have to leave 11:25 as I have to 400 elementary school students (science fair) but will be back for 10.

<Steven> I don't understand your last message Leigh

<Steven> Time warp?

<nick> zakim who is here?

<nick> fine for me

<Steven> mark?

<Steven> Which webconference are we going to use?

<markbirbeck> coming...

<Steven> Yabadaba or the other one?

<markbirbeck> up to you...Yugman is running now, if you want to use it.

<CharlieW> is yugma able to share from any client?

<Steven> ok

<Steven> coming

<markbirbeck> yes

<CharlieW> k, let's go with that then

<Steven> The system is pretty straightforward...just go to yugma.com and click

<Steven> on 'Join a session'. You'll then be prompted to create an account. My

<Steven> premium account does allow us to skip that step, but to do that I'd

<Steven> need to invite everyone individually, so it seems a lot easier for

<Steven> everyone to just create a free account.

<Steven> As part of the sign-up you put in the session ID which is 203-299-949.

<nick> wxhat is de sessiosn id?

<wellsk> 203-299-949

<CharlieW> crashes my firefox

<nick> it works for me

<CharlieW> leopard or tiger?

<markbirbeck> tiger here

<CharlieW> leopard here

<markbirbeck> Do people still use Firefox?

<markbirbeck> ;)

<CharlieW> it's the jvm

<CharlieW> same either way

<CharlieW> i'll have to use windows under parallels, hang on

<nick> it is quite lagging ....

<Steven> I'm on

<klotz> i will start

<John_Boyer> Scribe: Leigh

<CharlieW> moi aussi

<John_Boyer> scribenick: klotz

<John_Boyer> Chair: John

<John_Boyer> Meeting: Forms WG Face to Face Virtual Day

<John_Boyer> http://lists.w3.org/Archives/Public/public-forms/2008Jan/0072.html

<John_Boyer> http://lists.w3.org/Archives/Public/public-forms/2008Jan/0071.html

<John_Boyer> <select1 ref="payment/methods/@method">

<John_Boyer> <label>Choose method:</label>

<John_Boyer> <item>

<John_Boyer> <label>

<John_Boyer> <value>

<John_Boyer> <group>

<John_Boyer> ...

<John_Boyer> </group>

<John_Boyer> </item>

<John_Boyer> This is what I think Mark is proposing

<John_Boyer> Rather than putting switch/case inside select/select1

<John_Boyer> <payment>

<John_Boyer> <methods method="VISA">

<John_Boyer> <Visa>

<John_Boyer> <number>

<John_Boyer> </Visa>

<John_Boyer> <mastercarddata>

<John_Boyer> <number>

<John_Boyer> <expiry>

<John_Boyer> </masterdata>

<John_Boyer> ...

<nic1> you can only show the content of the case that is selected, bcz not selected cases are non relevant

<nic1> this will not allow to implement the firefox example

<nic1> that markB gave us a minute ago

<Steven-eee> http://www.w3.org/MarkUp/Forms/2008/screen20080201.PNG

<nic1> I like this much more then adding the extra switch

<nic1> I had the same problem as charlie

<Steven> Do we really need itemset?

<Steven> Can't we just say that <item nodeset="..."> does the job of itemset

<John_Boyer> Could do, but then the case is still inside the item, so you still get case repetition

<CharlieW> i still don't like calling the element <case>

<Steven> +1 to charlie

<Steven> take 5 for coffee

<John_Boyer> take 5

+1 to sausage sandwich

<wellsk> keith is here

<wellsk> apps

<nic1> I like the case in select and switch because the the case does the relevance behavior and the element around it (select, select1, switch)

<nic1> does the switching

<John_Boyer> <switch ref="some/tree">

<John_Boyer> <select1> .. </select1>

<John_Boyer> <case>

<John_Boyer> <case>

<John_Boyer> <case>

<John_Boyer> <switch ref="some/tree">

<John_Boyer> <select1 ref="@choice">

<John_Boyer> <item> label value...

<John_Boyer> </select1>

<John_Boyer> <case> label value

<John_Boyer> <case> label value

<John_Boyer> </switch>

<nic1> I find it a bit strange

<nic1> I don't understand why the select needs be inside the case

<nic1> and how you con be sure that the same values are in the cases

<nick> sorry network dropped out

<John_Boyer> <switch ref="payment/methods">

<John_Boyer> <input ref="@method"> ...

<John_Boyer> <case>

<John_Boyer> <case>

<John_Boyer> <case>

<John_Boyer> <switch ref="payment/method">

<John_Boyer> <xyzzy ref="@method">

<John_Boyer> <case>

<John_Boyer> <case>

<John_Boyer> Leigh says do the same thing as upload

<switch ref="payment/method">

<xyzzy ref="@method">

<select1> ...</select1>


<case> ...</case>

<wellsk> ha


<nick> minutes?

<John_Boyer> yes

<John_Boyer> Scribe: Steven

<scribe> Scribe: Steven

John: Ten mins to go, can we design a name in that time?
... the reason I suggested 'state' was that I use that name on an attribute now

Steven: How about 'using'

<wellsk> express?

John: 'selection'?
... Let's go with 'using' until we find something better

<CharlieW> +0.5


John: I propose we have enough to consider it for XForms 1.2

<scribe> ACTION: Jboyer write text for switch using [recorded in http://www.w3.org/2008/02/01-forms-minutes.html#action01]

<trackbot-ng> Created ACTION-449 - Write text for switch using [on John Boyer - due 2008-02-08].

John: Time for the break

<nick> buy

<nick> bey


<John_Boyer> will be back in an hour...

same code

<ebruchez> hi there

<wellsk> hi eric -- we will be starting up again about 12:30 EST

<ebruchez> thanks, are we calling zakim?

<nick> yes

<ebruchez> ok

<ebruchez> thx

<wellsk> yeah -- You didn't need me!

<nick> (not our regular conference code)

<ebruchez> yeah I tried that and was thrown out ;-)

<wellsk> zakiim, mute me

<ebruchez> are we also joining on Yugma?

<wellsk> re: yugma, I don't know

<wellsk> eric, we were using yugma.com this morning -- you can register for a free acct, and then "join session"

<wellsk> the session code is 203-299-949

<nick> now I'm muted

<nick> it really echo's

<markbirbeck> http://eric.van-der-vlist.com/blog/2008/01/31/html-5-turns-documents-into-applications/

<John_Boyer> Scribe: Steven

Default trigger

John: The idea is to simplify common cases
... but it's not clear if every implementation generates a DOMActivate for [enter] in an <input>

Steven: We could define tha an [enter] in an input generates a DOMActivate

John: But you might have to duplication the action for several controls

Steven: Well, the advantage of bubbling is that you can wrap the controls and put the action there

Erik: Is the issue just about submission?

<John_Boyer> <group>

<John_Boyer> <action ev:event="DOMActivate" whatever to say any time in bubble phase>

<John_Boyer> <send> default action

<John_Boyer> But what if the group contains multiple triggers?

Steven: You mean that in some cases you would have to stop the bubbling?

John: Yes. You can get it to work now, but it is easier to author in HTML
... and it should be as easy as that

STeven: Are you trying to capture some patterns in XForms now or just trying to capture the HTML [renter] in input case?

John: The latter
... How did it get to be like that?

Steven: I think that Netscape did that, and everyone copied them

John: How does the Google main page work, where there's two buttons?

Steven: There's a lot of script in there

Erik: You can do this with script

<nick> 4.4.12 The DOMActivate EventDispatched in response to: the "default action request" for a core form control, for instance pressing a button or hitting enter.

Steven: <input> doesn't take [enter], so the Activate may not even go to the control

Leigh: I just tried it on the google page, and it seems to activate the first trigger in lexical order

<John_Boyer> Scribe: Nick

<John_Boyer> scribenick: nick

<ebruchez> WG members get younger and younger

markB: Maybe the auther needs to add a form element and add an action handler

JohnB: Or maybe an id of the button that needs to be triggered
... How do we now if the user wants to trigger the default one, with some controls it is obvious
... for others it isn't

MarkB: Maybe each control has to say if it wants to buble the event up

John: But what if you hit a button

MarkB: But should it buble if you hit a button, to activate the default handler

John: You can use the tagname of the target

Charlie: we can maybe add extra payload
... a field that indicates if default action is taken

John: Maybe we need a different event
... xforms-default-action
... You can then write a handler at any level in the UI
... the handler can stop its propagation
... The attribute that points to the trigger can also be on a switch, group
... I'm focussing on a form or subform that can indicate a default submsission

<Steven> I think we need to be able to specify what a return in an input does

<Steven> it's not only about submission

John: But it is more general, bcz. you can write a handler on every level

Steven: I think we have this already, the only problem is that input doesn't generate a DOMActivate

Leigh: Select1 generates also a DOMAcivate and you want to submit the form when you select something

<John_Boyer> Suppose we have <group defaultcontrol="sometriggerID">

<John_Boyer> Then various controls like input could dispatch xforms-default

<John_Boyer> when you hit enter

ack, me

Steven: so input should not generate DOMActivate but another event

Nick: And use our existing eventing framework

<ebruchez> XForms 1.1 says "An example was added to show the use of a DOMActivate handler on an input to automatically initiate a submission once a user enters and commits input, such as a search query."

<ebruchez> so we would undo that in 1.1?

<ebruchez> http://www.w3.org/TR/xforms11/#rpm-event-sequences

John: It is just simplification with attributes, a shortcut for the action handlers, there is also the problem that we don't allow actions as child as for example switch. Then you have to use an event handler with target

Erik: There are good cases for DOMActivate

<ebruchez> http://www.w3.org/TR/xforms11/#intro-diffs-ui

MarkB: Input generating a DOMActivate is verry handy, and there are forms with no trigger so targetting a trigger isn't always possible

<John_Boyer> http://www.w3.org/TR/xforms11/#ui-input

Leigh: It is to complicated to solve the problem in complex form, sometimes you just want to go to the next input

<CharlieW> * have to hang up for 5min to call room service, back shortly

Erik: Some platforms don't have a default button on a dialog

John: Mac OS does the same as windows in relation to the default trigger handling
... Do we want to defer this to a future release

<klotz> in w3m, enter in text input allows you to edit the field; in lynx, it tabs to the next field.

<John_Boyer> http://www.w3.org/TR/xforms11/#ui-input

Steven: I tink we should try to solve it

<CharlieW> +1 on it being important

John: The simple case is already coverred in XForms 11
... What is verry deficult to solve : if you have a group of controls that has special default handling

<Steven> Oops, no errata for XForms 1.1?

John: Do we get DOMActivate when we make a selection in a select1

<John_Boyer> http://www.w3.org/TR/xforms11/#sequence-for-select

<CharlieW> it's particularly important for things we don't have a convenient way to do...

<CharlieW> boil down the inconvenient combinations of markup into a new simplified solution

<CharlieW> but mark is suggesting we start from an example of how we'd do it today, concretely which i think is a great idea

John: A select one minimal, tab into it, hit space, expand, select an item, hit enter, I don't want to do the default submission
... Go back to the select1 hit enter, now I want to submit

MarkB: Maybe we need to tune when DOMActivate is sent in Select1 but this is a small change

<CharlieW> it's a more evolve through experimentation and drive down into simplified markup

<John_Boyer> <group>

<John_Boyer> <label>

<John_Boyer> <send ev:event="DOMActivate">

<John_Boyer> <input>

<John_Boyer> <input>

<John_Boyer> <input>

<John_Boyer> </group>

<John_Boyer> What if the group contains a trigger or a select1?

<John_Boyer> Each trigger would have to stop propagating its DOMActivate

<John_Boyer> The select1 does not know whether to cancel it or not

John: Every dialog box has a default trigger
... It is a common pattern
... There are maybe forms out that use the DOMActive on a value change in slect1

MarkB: We should describe more carefully how controls behave

<John_Boyer> <group>

<John_Boyer> <dispatch ev:event="DOMActivate" event="DOMActivate" target="someTrigger"/>

MarB: The common pattern is a group that contains an event handler that dispatches an event to the action handler of your deafualt tigger

<John_Boyer> <group>

<John_Boyer> <dispatch ev:event="xforms-default" event="DOMActivate" target="someTrigger"/>

MarkB: You have to solve the two half problems
... We should describe more carefully how controls behave

John: Should we digg in the conntrols, or should we dispatch another event
... The fact that DOMActivate bubles, prevented us to put a dialog in an action, otherwise the triggers in the dialog will trigger the outer trigger

MarkB: You could define that triggers stops propogation of the DOMActivate event, but this breaks generic XML Events 2.0

<CharlieW> nice

MarkB: submit is symantic sugar

John: Stop propagation stops the bubbling, cancel prevents the default handling

Steven: Default action happens after bubbling

<Steven> according to the DOM

John: The real problem is the select1, we can not distinguish between selecting an item and pressing enter on the select1
... Can you write an handler with with no action in it

Steven: We have action

John: Action has to have at least one child

MarkB: You can have a handler attribute on an action

John: Can we change the + in a *
... is it XForms 1.1?

Steven: I hope

<John_Boyer> http://www.w3.org/TR/xforms11/#action

who gets the action item?

<CharlieW> +1 (agree)

RESOLUTION: Allow the action element to have 0 children in XForms 1.1

<wellsk> yep

<John_Boyer> <submit submission="X">

<John_Boyer> <label>Press me</label>

<John_Boyer> <action ev:event="DOMActivate" ev:stopPropagation="true"/>

<John_Boyer> </submit>

<John_Boyer> Test that the event doesn't bubble up to parent of submit, but the submission still happens because the submission is default processing

<CharlieW> * coffee too

<John_Boyer> take 5

<John_Boyer> resume

<John_Boyer> <action ev:event="something" ev:stopPropagation="true" if="some condition"/>

MarkB: I don't think doesn't stop propagation in XML events 1

<John_Boyer> http://www.w3.org/TR/xforms11/#sequence-for-select

John: There is no text that you get a DOMActivate when you press enter in select and select1
... You only get the DOMActivate if you change the control
... Concludes that you also get a DOMActivate when the value isn't changed.
... And don't get a DOMActivate when the value is changed

<markbirbeck_> dinner Sunday sounds good to me, too.

<John_Boyer> Section 4.6.3 bullet 1 says when you make a selection in a select/select1, you get value-changed, but you don't get DOMActivate

<John_Boyer> bullet 3 says if you "activate" the control (which a common UI experience says happens if you hit enter)

<John_Boyer> then the third bullet says you get the value changed before the DOMActivate

<John_Boyer> The spec assumes it's "obvious" that if you activate the control

<John_Boyer> you will get a DOMActivate

<John_Boyer> so it is only explaining the order of events

<John_Boyer> when the control has changed the value.

<John_Boyer> The user agent defines what "activate the control" means

Leigh: Select1 default incremental is true
... Does it activate the control, I'm not sure. I think it is dependant on the modality

<markbirbeck_> DOMActivate

<markbirbeck_> The activate event occurs when an element is activated, for instance, thru a mouse click or a keypress. A numerical argument is provided to give an indication of the type of activation that occurs: 1 for a simple activation (e.g. a simple click or Enter), 2 for hyperactivation (for instance a double click or Shift Enter).

MarkB: It is vague because it says that the event occurs when a control is activated, it doesn't say how it always activated
... Sso we can be verry specific

JohnB: When you click on a dropdown I would expect a DOMActivate but when you click in a textbox I wouldn't expect a DOMActivate

<John_Boyer> Steven, would you type that or say again?

<Steven> The number (1 or 2) is not normally exposed in the XForms event handling

MarkB: It is availbe with the event function

<ebruchez> yup

<John_Boyer> http://www.w3.org/TR/xforms11/#evt-activate

JohnB: We should add context info to the DOMActivate event

MarkB: The hyper activation is the solution

<CharlieW> in previous work, we've distinguished between 3 states: interest, selection, execution

<CharlieW> this applies across all UI controls

MarkB: We need to add an if to Johns example of an handler on the group and use the context info

<Steven> Keith, Doug Schepers will attend all three days

MarkB: And stop propagation of the DOMActivate event on triggers

<wellsk> ok

<klotz> http://xformstest.org/2008-02-01.txt

<ebruchez> bye

Summary of Action Items

[NEW] ACTION: Jboyer write text for switch using [recorded in http://www.w3.org/2008/02/01-forms-minutes.html#action01]
[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.133 (CVS log)
$Date: 2008/02/01 20:04:29 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
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/bubbling/bubbling?/
Succeeded: s/his/its/
Succeeded: s/sthe/the/
Succeeded: s/to have 0 children/to have 0 children in XForms 1.1/
Found Scribe: Leigh
Found ScribeNick: klotz
Found Scribe: Steven
Inferring ScribeNick: Steven
Found Scribe: Steven
Inferring ScribeNick: Steven
Found Scribe: Steven
Inferring ScribeNick: Steven
Found Scribe: Nick
Found ScribeNick: nick
Scribes: Leigh, Steven, Nick
ScribeNicks: klotz, Steven, nick
Default Present: wellsk, nick, [IPcaller], ebruchez, CharlieW, +1.978.399.aaaa, John_Boyer, markbirbeck, Steven, Leigh_Klotz
Present: wellsk nick [IPcaller] ebruchez CharlieW +1.978.399.aaaa John_Boyer markbirbeck Steven Leigh_Klotz
Regrets: Uli
Agenda: http://www.w3.org/MarkUp/Forms/wiki/Agenda_Feb._1%2C_2008_%28virtual_day%29
Got date from IRC log name: 01 Feb 2008
Guessing minutes URL: http://www.w3.org/2008/02/01-forms-minutes.html
People with action items: jboyer

[End of scribe.perl diagnostic output]