Efficient XML Interchange Working Group Teleconference

01 Dec 2015

See also: IRC log




<scribe> scribe: TK

<scribe> scribeNick: taki

OSS Nokalva implementation

DP: Objective systems at one point was implementing EXI too.

CB: It talks about last call draft. It may be old implementation.

<caribou> http://www.objsys.com/support/xbrelease23.html

CB: Current release is 2.3.

DP: It seems it no longer talks about EXI.

TK: I would suggest Alessandro if he might want to make an announcement on public list.

DB: We might want to ask OSS Nokalva to join the WG.
... I am going to New Jersey for 10 days so maybe i can pay them a visit? so maybe i can pay them a visit?.

CB: They are no longer a W3C member company.

<brutzman> it would be great to invite them to present at an upcoming EXI meeting, and also invite them to re-join W3C and participate in EXI working group.


DP: The idea is to keep the document very simple.
... To use EXI, and transform JSON documents to XML, to EXI.
... Similar to XQuery WG does, instead of inventing a mapping.
... I proposed a schema. We already discussed it.
... JSON to EXI, and EXI to JSON.

DP going through the document...

DP: The document is conceptually complete.
... datatype extension part needs to be described.

<brutzman> Very very interesting. Thanks for writing this up and sharing Daniel. Your concept is well described.

DP: I recently published Javascript implementation.
... I haven't accounced that yet. Once I got demo page set up, I will make an announcement of the implementation.

<brutzman> It looks like the approach is describing data structures. There is a lot of diverse work out there mapping JSON :: XML but it seems pretty inconsistent.

DB: Data structure. A lot of work has been done for mapping, however, have not been very consistent.

<brutzman> Given your excelent progress, I think there is an interesting potential opportunity to go to a next level of abstraction: mapping directly to XML, so that both XML and EXI mappings to/from JSON are possible.

<brutzman> This is possible because of the object-oriented nature of JSON.

DB: X3D, which is still evaluation stage, has done XML-JSON mapping.
... There, we used the same structure in both XML and JSON.

<brutzman> Paraphrase: it is possible to make JSON data structure look like equivalent XML, not pseudo- or intermediate representations..

DP: In the ideal case, we can map JSON to XML, XML to JSON.
... We, Siemens, wanted to use the same implementation for dealing with XML and JSON.
... CBOR, for example.
... People need to implement both XML and CBOR.
... In our case, we have one implementation.

<brutzman> The X3D working group has recently performed a full mapping between Extensibile 3D (X3D) and JSON.

<brutzman> Since the X3D schema is fairly rich, it appears to accomplish most XML mappings of interest.

DB: We, X3D, have a doctype and schema.

<brutzman> We worked through a number of differences and were able to resolve them. This can likely be generalized to any XML schema (i.e. any PSVI XML, or any EXI).

DB: Differences in structures of JSON and XML. We were able to combine some of the existing techniques.

<brutzman> We found a number of different techniques for XML :: JSON that had many common characteristics and a few differences... we adapted the best parts as might be needed.

<brutzman> It is curious that people keep solving this core problem of XML :: JSON.

<brutzman> The X3D to JSON Stylesheet Converter page describes our approach in detail. http://www.web3d.org/x3d/stylesheets/X3dToJson.html

<brutzman> Key section is Data Types to show mappings. A variety of examples are also provided.

DB: For example, "Route" element which repeats turns into an array. See Design Patterns section in http://www.web3d.org/x3d/stylesheets/X3dToJson.html

DP: In general case, I wonder how XML mapping would look like.

DB: Element order may change slightly when other elements were between Route elements originally.

CB: XQuery/XSLT has defined XML-JSON mapping. They are using schema types.

DP: Didn't they do essentally the same way as we do?

<brutzman> "W3C HTML JSON form submission" W3C Note edited by Robin Berjon http://www.w3.org/TR/html-json-forms

CB: If you really want to round-trip, you need some schema-typing. No WG has done that.

<brutzman> Speculation: if we can show general 2-way mappings, and identify limitations, then a mapping might be able to address.

<brutzman> Speculation: if we can show general 2-way mappings between JSON and XML Schema, and identify limitations, then an EXI mapping might be able to address the limitations.

<brutzman> Big applause for Daniel! BCNU guys next week, aloha.

DP: In the future, if there are going to be a mapping between JSON and XML, then we can simply use it.

TK: 3.1.1 and 3.2.1 could be merged into one subsection, for example.

DP: Let me think about it.

Canonical EXI

DP: I was trying to use second-level production in some applications.
... We define everything as infoset. Implementations that use databinding, empty CH is reported.
... Event sequence should be the same regardless of strictness.

TK: The benefit of being consistent needs to be discussed.

Summary of Action Items

Summary of Resolutions

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.144 (CVS log)
$Date: 2015/12/01 16:36:57 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.144  of Date: 2015/11/17 08:39:34  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

Succeeded: s/10 days/ 10 days so maybe i can pay them a visit?/
Succeeded: s/10 days/ 10 days so maybe i can pay them a visit?/
Found Scribe: TK
Found ScribeNick: taki

WARNING: No "Present: ... " found!
Possibly Present: CB DB DP Paraphrase Speculation TK brutzman caribou 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: 01 Dec 2015
Guessing minutes URL: http://www.w3.org/2015/12/01-exi-minutes.html
People with action items: 

WARNING: Input appears to use implicit continuation lines.
You may need the "-implicitContinuations" option.

[End of scribe.perl diagnostic output]