Efficient XML Interchange Working Group Teleconference

15 Mar 2016

See also: IRC log




<scribe> scribe: TK

<scribe> scribeNick: taki

W3C TPAC 2016 - Will your group meet in Lisbon?

TK: Javier and Don will join by phone. DP and TK will like to be able to attend.

DP: One idea is to have a breakout session on Wednesday.
... We need to prepare some short presentation for that.

DB: There is still a chance for me to go and join. Just not sure for now.

Diagrams of EXI Compression for web Pages

DB: Tried to extend EXI-JSON pattern to other technologies.
... CSS is quite structured.

DP: I also plan to play a bit with CSS.

DB: JavaScript. Somebody expressed Java in XML years ago.

<dape> https://webassembly.github.io/

DP: There is some effort. It is based on abstract tree.
... WebAssembly is a community within W3C.

<brutzman> Possibility of asking Javascript ECMA effort what is official compression

<brutzman> Possibility of including a variety of JavaScript compression schemes (CBOR JSON etc.) to replace the text block, with annotation regarding compression used

<brutzman> Diagram should insert box showing JSON XML representation

<brutzman> Please type in any comments so that I can improve next version of this diagram

DB: HTML5 has two syntaxes. One is HTML. The other is XHTML/

<brutzman> Two syntaxes are defined in HTML5: HTML syntax (includes HTML4) and XHTML syntax

<brutzman> Both syntaxes have clear definitions how to map to DOM

<brutzman> Dotted lines mean something that is not yet done

DB: We need to find or create XHTML5 schema.

DP: Schema is useful for most of our use cases.
... When it comes to text-based messages, the benefit you get from schema is not that big.

<brutzman> Interestingly, schema is not required for EXI; can instead use schema-less EXI

<brutzman> Additionally I think that having a schema helps with quality assurance (QA) of input content

TK: I tried HTML schema about 4 years ago. I found the generated grammar very large.

<brutzman> Nevertheless I think schema-informed EXI offers major opportunities for performance and consistent EXI HTML output documents

<brutzman> Knowing all element/attribute tokens in EXI tables in advance offers numerous performance improvements; this greatly benefits Internet of Things IOT becoming Web of Things WOT

<brutzman> Summary: for HTML, schema is most likely to benefit performance (compression maybe not so dramatic)

DP: You need to decide upfront which set of schemas you use.

DB: SVG schema is also possible.
... Bundling might already be EXI2 requirement.
... Bundling is similar to ZIP and tar.

DP: Is it a container to contain several documents?

DB: Yes.
... Unbundling makes the original Web page.

DP: Archiving for Web pages, I am not sure if it works well or not.

<brutzman> Perhaps it is listed in prior EXI goals & requirements; will investigate further

DB: Compression and archiving is similar to tar+gzip.

<brutzman> Summary: zip compression + achiving ~= .gzip + tar ~= EXI with EXI2 bundling

DB: Bundle can be encrypted and signed to make compressed secure web page.
... To reverse the steps, we can first verify authentication, decrypt, then unbundle it.

DP: EXI bundle contains CSS, HTML, JavaScript files. Is that correct?

DB: Correct.

<brutzman> Bundle is a container that contains multiple EXI documents

<brutzman> Merit is that each individual contained EXI document uses a different compression scheme

DP: Some files cached. I wonder how bundle work in HTML web page use case.

<brutzman> Primary function of "bundling" is to collect multiple diverse documents, include metadata/structure information, and allow subsequent reassembly into source HTML document with all original sections

DB: We can use XSLT in the unbundling process.
... CSS to XML conversion is the only missing part.
... We can add more test cases to compare against Zip, cbor, bson, etc.

<brutzman> Thanks for reviewing this together... As the summary list of items on page 3 shows, there is more work to do. It is a happy surprise that most functional pieces are available already.

<brutzman> DP: important to reach out to other groups in W3C regarding what can be achieved

DP: We can reach out to Web folks. TPAC is a good chance to get feedback.

<brutzman> We should imrove these materials and list known gaps. We should brief Liam, Robin, other technical experts to seek feedback.

<brutzman> Discussed how prominence of Javascript efforts sometimes seem to overshadow importance of XML and document-centric markup.

DP: We need show benefits of EXI, would like to know feedback.
... One chance is TPAC.

Summary of Action Items

Summary of Resolutions

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.144 (CVS log)
$Date: 2016/03/15 15:35:20 $

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/DB/DP/
Succeeded: s/EXMA/ECMA/
Succeeded: s/Nevertheless/Additionally/
Succeeded: s/Zip/ZIP and tar/
Found Scribe: TK
Found ScribeNick: taki
Present: TK DP CB

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

Found Date: 15 Mar 2016
Guessing minutes URL: http://www.w3.org/2016/03/15-exi-minutes.html
People with action items: 

[End of scribe.perl diagnostic output]