See also: IRC log
<nvdbleek> nvdbleek: I think we should supported quoted values in csv (see http://tools.ietf.org/html/rfc4180 for csv format)
trackbot, start telcon
<trackbot> Date: 21 November 2012
http://lists.w3.org/Archives/Public/public-forms/2012Nov/0020.html
Steven: Do we want to update to the newest namespaces spec?
Nick: I see no problem in that
Steven: But I think we deliberately reference XML 4th ed, because of the problem of the new name characters in 5th ed.
Erik: I don't know for sure which parser we use, it's Xerxes, not very old, may be 5th ed, we need to check.
Steven: If we require 5th ed, then we need implementations that support it.
Nick: Xerxes C++, only supports 1.0 3rd.
Erik: An XSLTForms? Does it use the browser's parser?
Steven: Yes
Nick: Xerxes Java uses XML 1.0
4th ed.
... but the newest XML 1.1
Erik: Safer to leave it as it is for now.
http://www.w3.org/MarkUp/Forms/wiki/XForms_2.0#External_JSON_values
Steven: I've updated these parts, and people need to take a look. In particular there is now treatment of anonymous arrays.
Nick: I think we should write a
blog post about the rationale for this approach, rather than
the XSLT 3.0 mapping
... which may give us some more input.
... I think what we now specify is quite OK.
Steven: I think now it covers all cases.
Nick: And allows as-easy-as-possible editing of the JSON before resubmitting it/serialising it.
http://lists.w3.org/Archives/Public/public-forms/2012Nov/0011.html
Steven: Tough one, but maybe we
should bite the bullet.
... I get the feeling we think we should leave this out of
XForms 2.0 until it is more mature
Nick: THat's my opinion
sTh/Th/
Nick: It's not fully baked yet.
Steven: I tend to agree; there are problems we don't really have a good solution for yet.
Nick: Each implementation has problems, and are not ideal.
Steven: Well, I think if we have
no consensus, we need to leave it out for the time being.
... We can still work on it.
Nick: The limitations in HTML make it difficult; maybe the future shadow DOM will make it easier
http://lists.w3.org/Archives/Public/public-forms/2012Nov/0016.html
http://homepages.cwi.nl/~steven/forms/xpath-stable-scope-test.xml
Nick: This is a problem with
'stable' functions ('pure' in functional programming)
... when they get reevaluated
... and this is not interoperable, now() being a good
example
... We need to decide when the functions get reevaluated
... in the UI on refresh
... in actions on the outermost action handler
... binds are unclear
Steven: binds get reevaluated
depending on their dependencies
... and we plan to loosen that definition as well
... so that may mean it is by definition non-interoperable
Erik: In XForms we have many functions that may not be 'stable'.
Nick: It's not as bad as all
that.
... but we need to specify the minimal case that such a
function gets reevaluated.
Erik: I have no strong feelings.
Steven: Can't we be loose, and say "it gets evaluated when it gets evaluated, and that depends on many other things, such as the other values that change around it"
Erik: if you write
now()=something or now()=something-else, then you'd like to
know how it gets interpreted
... Maybe it's just a question of the definition for one XPath
expression.
Steven: I'm not sure; an implementation might optimise the whole dependency tree.
Nick: XPath says that dates and times are stable.
Steven: Woh. That surprises me
<nvdbleek> http://www.w3.org/TR/xpath20/#id-xp-evaluation-context-components
Nick: I have to say something in the spec, since I took the text from XPath
Erik: We could say that those things should be stable
<nvdbleek> http://www.w3.org/TR/xslt20/#xpath-dynamic-context
Steven: We need to say something,
we can't require an implementation to optimise; shouldn't we
just say "don't expect cases like this to give identical
results".
... Do you know what to doNick?
Erik: We need to keep this open
Nick: I'll keep it open, and add a note asking for implementation feedback
Nick: Please look at the format
token emails and give feedback.
... I want to finish tomorrow if poss.
... Publish a version of the spec without ed notes.
[ADJOURN]
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/And browsers/An XSLTForms/ Succeeded: s/DO/Do/ Succeeded: s/TO/To/ Succeeded: s/TH/Th/ Succeeded: s/kc/ck/ Succeeded: s/;// Succeeded: s/iufy/ify/ Succeeded: s/nh:/n:/ Succeeded: s/,ay/may/ Succeeded: s/yp/ypo/ Succeeded: s/ypo/yo/ Succeeded: s/n;t/n't/ Succeeded: s/sday/say/ Succeeded: s/?/Nick?/ Succeeded: s/DO/Do/ No ScribeNick specified. Guessing ScribeNick: Steven Inferring Scribes: Steven Default Present: +1.650.919.aaaa, Steven, Nick, Erik Present: +1.650.919.aaaa Steven Nick Erik Regrets: Philip Agenda: http://lists.w3.org/Archives/Public/public-forms/2012Nov/0019.html Found Date: 21 Nov 2012 Guessing minutes URL: http://www.w3.org/2012/11/21-forms-minutes.html People with action items:[End of scribe.perl diagnostic output]