IRC log of forms on 2015-09-30

Timestamps are in UTC.

13:01:44 [RRSAgent]
RRSAgent has joined #forms
13:01:44 [RRSAgent]
logging to http://www.w3.org/2015/09/30-forms-irc
13:01:46 [trackbot]
RRSAgent, make logs public
13:01:47 [Zakim]
Zakim has joined #forms
13:01:48 [trackbot]
Zakim, this will be
13:01:48 [Zakim]
I don't understand 'this will be', trackbot
13:01:49 [trackbot]
Meeting: XForms Users Community Group Teleconference
13:01:49 [trackbot]
Date: 30 September 2015
13:02:06 [Steven]
Agenda: https://lists.w3.org/Archives/Public/public-xformsusers/2015Sep/0011
13:02:22 [Steven]
Steven has changed the topic to: Agenda: https://lists.w3.org/Archives/Public/public-xformsusers/2015Sep/0011
13:02:35 [Steven]
Chair: Steve
13:02:41 [Steven]
s/Steve/Steven
13:05:04 [Steven]
Present: Steven, Alain, Erik, Nick
13:07:59 [Steven]
Topic: XForms day 5th Nov
13:07:59 [Steven]
[No link]
13:08:37 [Steven]
Steven: The event is coming on nicely, 4 varied speakers at present.
13:09:04 [Steven]
Topic: Editing
13:09:04 [Steven]
https://lists.w3.org/Archives/Public/public-xformsusers/2015Jul/0005.html
13:09:23 [Steven]
Steven: No progress to report. You're still having trouble Erik?
13:09:33 [Steven]
Erik: Yes, I've sent a message to Liam
13:11:22 [Steven]
Topic: public-xformsusers
13:11:22 [Steven]
[No link]
13:11:40 [Steven]
Steven: Nothing to report. I think next week Mark will join us on the call.
13:11:48 [Steven]
Topic: JSON-to-XML and CSV-to-XML mappings in XForms 2.0
13:11:48 [Steven]
[No link]
13:12:22 [Steven]
Steven: Not sure what to do about this.
13:12:43 [Steven]
... Let's research this week and talk about it next.
13:12:46 [Steven]
Erik; Sure
13:12:54 [Steven]
Topic: Comments on Delete action
13:12:54 [Steven]
https://lists.w3.org/Archives/Public/public-xformsusers/2015Sep/0010.html
13:13:11 [Steven]
"1. "The delete action is terminated with no effect if the delete context is
13:13:11 [Steven]
the empty sequence."
13:13:11 [Steven]
"
13:13:28 [Steven]
Erik: I reread this section, and found some surprises for me.
13:13:37 [Steven]
... some are more important than others
13:14:26 [Steven]
... This is a change since XForms 1.1, the context attribute is now very general.
13:15:05 [Steven]
... (Delete actually uses it in a different way, but that's another topic)
13:15:32 [ebruchez]
<delete
13:15:33 [ebruchez]
context="()"
13:15:34 [ebruchez]
ref="instance('my-instance')/bar"/>
13:18:11 [ebruchez]
<delete context="foo" ref="bar"/>
13:18:14 [Steven]
Steven: Step 2 already says there is no effect if the binding gives no nodes
13:18:34 [Steven]
Erik: If foo is empty, then so is bar.
13:19:15 [Steven]
Steven: Then let's delete the sentence from step 1
13:20:12 [Steven]
Nick: This was for XPath 1
13:20:34 [Steven]
Erik: I remember Michael Kaye saying something about this.
13:21:23 [Steven]
... in XForms 1 there may be no way of doing this, and that would be the reason for having this sentence.
13:22:06 [Steven]
Steven: So the sentence is there to stop XForms 1 implementations from throwing an error
13:22:29 [Steven]
s/XForms 1/XPath 1
13:22:32 [Steven]
s/XForms 1/XPath 1
13:25:07 [Steven]
Steven: Can we express it in a generic way so that it doesn't matter which version of XPath we use.
13:28:15 [Steven]
Steven: Such as in step 2 "If the expression uses the context and the context is empty, then terminate with no effect".
13:28:59 [Steven]
Erik: It would introduce an theoretical incompatibility.
13:29:36 [Steven]
... I am split. I am OK to leave this for now.
13:29:57 [Steven]
2. "The behavior of the delete action is undefined if the Sequence Binding
13:29:57 [Steven]
node-sequence contains nodes from more than one instance"
13:30:13 [Steven]
Erik: I don't know why this written this way.
13:30:33 [Steven]
... I don't see why it wouldn;t delete from both instances.
13:30:38 [Steven]
s/;/'/
13:30:41 [ebruchez]
<delete
13:30:42 [ebruchez]
ref="instance('my-instance')/bar, instance('your-instance')/baz"/>
13:31:06 [Steven]
Nick: Maybe XPath 1 again?
13:32:11 [Steven]
Erik: YOu would be very unlikely to do this in XForms 1.1
13:32:20 [ebruchez]
XPath 1.0: ref="instance('my-instance')/bar | instance('your-instance')/baz"/>
13:32:21 [Steven]
s/YO/Yo/
13:33:36 [Steven]
Steven: Is there any reason to exclude elements from more than one instance?
13:33:38 [Steven]
[No]
13:33:46 [Steven]
Steven: Let's remove this restriction then.
13:34:02 [Steven]
3. "Otherwise, the Sequence Binding is not expressed, so the Sequence
13:34:02 [Steven]
Binding node-sequence is set equal to the delete context node with a
13:34:02 [Steven]
position and size of 1."
13:34:37 [Steven]
Erik: So there is no ref and bind, so we delete the context node. Why?
13:34:40 [ebruchez]
<xf:delete/> or <xf:delete context="foo"/>
13:35:06 [Steven]
Nick: The first one would work in a delete
13:35:10 [ebruchez]
<xf:delete ref="."/>
13:35:29 [Steven]
Steven: I agree it's better to be explicit
13:35:33 [ebruchez]
<xf:delete iterate="foo"/>
13:35:44 [Steven]
Nick: Or an iterate action
13:37:17 [Steven]
Erik: Either remove this or provide examples.
13:38:23 [Steven]
Steven: I propose we remove it.
13:39:09 [Steven]
Erik: It's true it would be backwards incompatible.
13:40:20 [Steven]
Steven: IS it generally true that no ref means ref="."?
13:40:29 [Steven]
Erik: Not as far as I am aware.
13:40:50 [Steven]
Steven: <repeat ref="foo"><output/></repeat>
13:41:18 [Steven]
... is not allowed AFAIK
13:44:51 [Steven]
Nick: I would prefer that we kept it as is, and added an example
13:47:41 [Steven]
Nick: Alain what does your implementation do?
13:47:46 [Steven]
Alain: I'll check.
13:48:02 [Steven]
Erik: I would argue for being more explicit.
13:48:21 [Steven]
Steven: I support that, but I would never have written it without ref.
13:49:05 [Steven]
Erik: We could deprecate it.
13:49:20 [Steven]
Steven: That's a perfectly acceptable path.
13:49:26 [Steven]
Nick: I think that's better.
13:49:39 [Steven]
Erik: I support that.
13:49:45 [Steven]
Alain: I think that's better.
13:50:00 [Steven]
ACTION: Erik to deprecate <delete> without a node binding
13:50:00 [trackbot]
Created ACTION-2031 - Deprecate <delete> without a node binding [on Erik Bruchez - due 2015-10-07].
13:50:37 [Steven]
ACTION: Erik to allow delete on multiple instances
13:50:37 [trackbot]
Created ACTION-2032 - Allow delete on multiple instances [on Erik Bruchez - due 2015-10-07].
13:50:50 [Steven]
4. cannot delete "namespace node"?
13:51:07 [Steven]
Erik: I'm not sure about this one.
13:51:16 [Steven]
... not sure if we can solve it live.
13:51:24 [Steven]
... I think we support deleting them
13:52:55 [Steven]
... in fact we even have a use case.
13:53:35 [Steven]
Nick: We use XPath model, not a DOM model
13:53:42 [ebruchez]
http://www.w3.org/TR/xpath-datamodel/#NamespaceNode
13:54:12 [Steven]
Steven: Can I ask all implementations to check whjat would happen in this case?
13:54:25 [Steven]
"5. We must add what to do if the sequence points to some atomic values. We
13:54:25 [Steven]
should say that those are ignored just as readonly items are.
13:54:25 [Steven]
"
13:54:32 [Steven]
s/hj/h/
13:54:42 [Steven]
Steven: Let's do the namespace nodes next week.
13:56:37 [Steven]
Erik: It wasn't origianlly written with atomic values in mind. I think we should add a sentence clarifying.
13:56:48 [Steven]
s/origianlly/originally/
13:57:12 [Steven]
ACTION: Erik to add clarification about deleting atomic values
13:57:12 [trackbot]
Created ACTION-2033 - Add clarification about deleting atomic values [on Erik Bruchez - due 2015-10-07].
13:57:56 [Steven]
[ADJOURN]
13:58:02 [Steven]
rrsagent, make minutes
13:58:02 [RRSAgent]
I have made the request to generate http://www.w3.org/2015/09/30-forms-minutes.html Steven
15:32:29 [Zakim]
Zakim has left #forms
15:45:00 [ebruchez]
ebruchez has joined #forms
16:09:48 [liam]
Steven, still there?
19:53:27 [Steven]
Now I am
19:53:46 [Steven]
Liam
20:41:55 [liam]
Steven, oops, and I wasn't, sorry
20:42:22 [liam]
I'll try and make the call next week and talk/listen about where the wg/cg might go forward
20:44:39 [liam]
possibly what's needed is a tightly-edited spec as a cg report, together with a test suite & some results, as that might let us make a wg quickly without the politics. Doing that might actually be more valuable than publishing a rec I suppose, in the end, but test results are a necessary step in any case
21:18:19 [Steven]
Sure. Thanks.
21:23:02 [Steven]
Oh, and Erik Bruchez can't log in to the wiki. Any chance of finding out what's up?
21:37:19 [liam]
i''ve replied to his mail
21:37:31 [liam]
i *think* he needs to use a different account out of the 4 he has
21:37:53 [liam]
i can't have access to the WG wiki changed, only to the CG wiki