Alain Couthures, AgenceXML
Dan McCreary, Dan McCreary Associates
John Boyer, IBM
Philip Fennell, MarkLogic
Steven Pemberton, CWI/W3C
Leigh Klotz, Xerox
Kurt Cagle, XMLToday
Uli Lissé, DreamLabs
Steven Pemberton: It's not really a
bad idea.
Leigh Klotz: I think we need to get
some editors together, not to date, but to get started.
Steven Pemberton: I think it would
help. Would it be a real face to face?
Leigh Klotz: That would be ideal; it's
going to be a hardship.
Steven Pemberton: I'm always willing
to host in Amsterdam.
Uli Lissé: +1 for
Amsterdam
John Boyer: We've done this at F2F
meetings.
Steven Pemberton: Nick and I did that
at the last one, in the wiki. Nick and I could get together and
join by phone as necessary.
Leigh Klotz: I'm assuming it wouldn't
be during the summer.
Steven Pemberton: I will be in the Bay
area the last week of August.
Leigh Klotz: I'll be able to provide
space then.
Steven Pemberton: I'll talk with Nick
about some work here and then later in August.
Leigh Klotz: Then in August we can get
Erik, and then John is in the same time zone.
Steven Pemberton: I'm speaking at
http://uxweek.com/2011/
Steven Pemberton: For August 29-31,
2011, in SF Bay Area. And sometime before then in Amsterdam with
Nick, depending on his availability.
John Boyer: We might want to think
outside the box and have you and Nick lay the groundwork for the
mechanics of pulling specs together and follow closely with work
out in Pacific time to get together and do a work party.
Leigh Klotz: That makes a lot of
sense.
ACTION-1798 Steven Pemberton to contact Nick van den Bleeken about editorial meeting.
Resolution 2011-05-18.1: Editorial meeting August 29-31, 2011, in SF Bay Area.
Leigh Klotz: We need two functions.
What name should we choose for the second?
Philip Fennell: I'm not convinced; we
might be able to use one function. The fragement ID might not be
implementable, but makes sense for a form author. Alternatively,
you could have a URI scheme for resolving internal URIs, as in
Coccoon. ("coccoon:" and "resource:").
Leigh Klotz: We have an internal ref
XPath and an external URI ref, and we have two ways to unify them
(a unique scheme: and XPointer), and two was to dis-unify them
(separate functions, and special markers).
Philip Fennell: What is the likelihood
of needing an XPath expression for retrieving an XSLT Transform
from an instance given that an XSLT Transform is unlikely to be
embedded, but perhaps that's not...
Leigh Klotz: I do embed them.
Philip Fennell: I'd like an inline
transform, so you can use an ID.
Leigh Klotz: Did someone say it
doesn't work to use fragment ID?
Philip Fennell: It doesn't work in the
instance.
Leigh Klotz: Just put the transform in
xhtml:head and put it use fragment id.
Philip Fennell: But what if it's in
the instance?
Steven Pemberton: What's this about?
Can we just test the proposals with some examples and see which
ones we like most?
Philip Fennell: I'm quite happy to
write that up.
ACTION-1799 Philip Fennell to write up several use cases of transform function and means of referring to the transform (URI, ID, XPath) http://lists.w3.org/Archives/Public/public-forms/2011May/0014.html
Philip Fennell: There other issues:
parameters, error conditions and events.
Leigh Klotz: We discussed them last
week. Can you include those in the write-up?
Philip Fennell: Sure.
Philip Fennell: We can skip over this for now because the transform function needs to refer to the instance. Maybe another week.
Leigh Klotz: John, what is a
recommended plan for JSONx? Should we provide comments?
John Boyer: I'm involved with this now
and have contributed some comments. A couple of IBM products
provide JSON and XML conversion. This is the same kind of work
we've done with XForms and working with an external JSON submission
service, so we can process it and respond with JSON as a result.
The XML is a profile of XML with a schema and which is mapped to
JSON. So when we translate XML back to JSON, it's only XML that
came from JSON in the first place. This proposed XML format is very
mechanical and does completely cover JSON, but it doesn't produce
very friendly markup, and when you use XPath it doesn't look like
the JSON expression.
Steven Pemberton: So the same process,
but not the same answer.
John Boyer: Not as much of a process,
perhaps. The server products are heavy-duty: DB2 and IBM Data Power
accelerator engine. The authors using these products are J2EE
programmers and aren't markup web and form developers, who benefit
more from a simplified markup, if there's really no difference
otherwise. But also, there's a mismatch with JSON: the benefit of
JSON is that it's about JavaScript in the client. That's not where
these heavy-duty server things live.
Steven Pemberton: With us, it took
quite a while to crystallize and for us to understand what we're
trying to do. We have a simpler solution when the source of the
data is the JSON and we're not handling generic XML-to-JSON, and
not the other way around.
John Boyer: With the WG proposal we do
a natural mapping to the extent possible. They didn't want special
rules for non-XML characters.
Steven Pemberton: They didn't want to
look it up.
John Boyer: It makes it easier for the
programmer, not the user. The WG cares more about the authoring
syntax. Here we provide those as exceptions, and avoid obtuse
syntax most of the time. Are we 100% clean on all JSON?
Leigh Klotz: http://www.w3.org/MarkUp/Forms/wiki/Json
John Boyer: So it's better to have a
few exceptions expressed as exceptions.
Steven Pemberton: So we should
comment?
John Boyer: It is a request for
comments, yes. We should be clean on the edge cases ($ etc.)
John Boyer: This is also an issue for
XSLT and other groups.
Leigh Klotz: I should raise this at
XML CG.
John Boyer: The standards machinery
will move forward. And we shouldn't use this ugly XML, as people
will just stick with JSON.
Leigh Klotz: Our wiki page is not
complete.
John Boyer: The general principle says
to preseve the naming contention, and then the edge cases.
Steven Pemberton: Can you have any
character?
Leigh Klotz: There are characters in
JSON http://timelessrepo.com/json-isnt-a-javascript-subset
John Boyer: So is JSON a superset of
JavaScript?
Leigh Klotz: Not of all of
JavaScript.
John Boyer: But JavaScript
objects.
Steven Pemberton: So in the
"JSON":"Rocks" example, does JSON restrict the left part?
John Boyer: If we find illegal
characters on the left hand, then we need a way to encode
them.
Steven Pemberton: The wiki page is a
bit hand-wavey, as it's the basis for discussion, but it says
"something like the following mapping."
Leigh Klotz: There are some Unicode
code points for Asian languages which are in XML 1.1 and XML 1.0 R5
but not in XML, and so they need to be transliterated into
something else.
Steven Pemberton: Yes, you put them in
the attribute and refer to them.
Leigh Klotz: I'll come up with the
Japanese characters for example.
Steven Pemberton: I'll firm up the
page.
John Boyer: So Leigh will bring it up
to XML Coordination Group.
ACTION-1800 Steven Pemberton firm up JSON wiki page.
ACTION-1801 Leigh Klotz to talk to XML CG about JSON to XML.
ACTION-1802 Steven Pemberton to send comment on http://datatracker.ietf.org/doc/draft-rsalz-jsonx/