<trackbot> Date: 06 October 2015

<scribe> scribe: TK

<scribe> scribeNick: taki

JSON support

DP: I sent the summary to XSLT group, Michael Kay forwarded to XQuery group.
... I am not sure how they handle it, whether it is in their telecon or email discussions. We will need to wait for their response.

CB: XQuery/XSLT, there is a weekly telecon. They have it on their agenda this week.
... XPath 3.1 is not yet a REC.

DP: With regards to schema, we can still decode the stream whether we use the original schema or our extended schema.
... I think.

Comments: Canonical EXI -- Last Call Working Draft

DP: Large integer issue, JS can live with it.
... As for datetime, both arguments are valid. JS recommends to use a switch.
... "preserveTimezone"
... similar to canonical XML's preserve comments, PIs

TK: How do we exchange those options?

DP: Fragment identifier. Similar to how Canonical XML does for comments.
... If it is the only option we would need to introduce, then I prefer not to have it.
... EXI options and schemaId in the header.
... I think YF described it well.
... I fear if we introduce many ways to do canonicalization, you need to exchange those options for canonicalization.

CB: In some applications, it can be exchanged static manner.
... JS is arguing to have flexibility to have or not have options.
... In his use case, applications know the settings in advance.

DP: It becomes less secure, I fear.
... We would need switches here again, right?

CB: In JS's use case, it may be hardcoded in application.

DP: His argument about stream bigger was not an issue. Additional code to handle was an issue for him.
... One goal I think is to eliminate the flexibility. We require sorting, for example.

CB: Canonical form is more constraint.
... It doesn't constrain the way we exchange options.

TK: Why didn't we converge on bitpacked?

DP: Changing coding mode requires a lot of more code.
... We require applications to exchange schemaId beforehand.
... We discussed, and decided it was safer to have schemaId.

TK: I think "preserveTimezone" is the real issue. This "options" issue is more design issue.

DP: For dateTime, at least we need to specify fractionalsec being zero is the same as no fractional seconds.

Support for Canonical EXI interoperability test in TTFMS

DP: I guess header options presence makes most of the differences.
... whitespace facet handling may be another issue causing some of the differences.
... schemaId is the location of the schema.

ACTION-725: Check whether we need to mention anything about which type to cast to in the case of profile stream

TK: I think it was a good catch.

This is scribe.perl Revision: 1.140  of Date: 2014-11-06 18:16:30  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

Found Scribe: TK
Found ScribeNick: taki

WARNING: No "Present: ... " found!
Possibly Present: CB DP TK dape exi joined scribeNick trackbot
You can indicate people for the Present list like this:
        <dbooth> Present: dbooth jonathan mary
        <dbooth> Present+ amy


WARNING: No meeting chair found!
You should specify the meeting chair like this:
<dbooth> Chair: dbooth

Found Date: 06 Oct 2015
Guessing minutes URL: http://www.w3.org/2015/10/06-exi-minutes.html
People with action items: