<scribe> scribe: TK

<scribe> scribeNick: taki

W3C TPAC 2016 - Will your group meet in Lisbon?

TK: I submitted the form, indicating that we are going to meet.
... I will later ask for telecon equipment in another questionaire.

DP: We do not know yet which days WoT is going to meet yet.

DB: I am still trying to get fund to attend TPAC.
... I am studying CSS.
... Content-model in CSS. Element-attribute values, such as selectors.
... To find best to way to convert XML into CSS.

DP: I will show you how current approach does.
... It shows how you map css to xml, and back to css again.
... For each css style rule, there is an element in XML.
... It bcomes bigger, but with EXI, becomes smaller.
... There are repearing strings.
... Benefit is not a lot with smaller css.
... I also used real one from w3c site. It reduced to 10% of the original CSS.

<brutzman> Wondering about a property structure that is closer to what we use in JSON, and perhaps even more compressible too

DP: The mapping I used was the mapping used by DOM.

<brutzman> example: <property name='background-color' value='#d0e4fe;' units='rgb'/>

DP: I wanted to use something that was specified somewhere already.
... I just walked through DOM, and generated XML out of it.

DB: I am looking at DOM4. Is it at somewhere else?

DP: It is DOM 2.
... DOM4, mentioned by Carine. I am using DOM2 Style specification.
... Carine mentioned another specification that they are working on.

<brutzman> https://www.w3.org/TR/2000/REC-DOM-Level-2-Style-20001113/

DP: Carine suggested to use CSS Object Model (CSSOM).
... The version I used is RECOMMENDATION. The others are editor's draft.

<brutzman> Agreed that https://drafts.csswg.org/cssom is best to use, and we should match capitalization & spelling exactly. Really super!

<brutzman> Looking at things to compress better: we might create enumerations for all known property names, found in CSS and in HTML

DB: For known property names, we should define enumerations.

<brutzman> ... that would be better than just unstructured strings. More reliable too.

DP: That's definitely true.

<brutzman> Need to look closely at schema regarding use of a fixed enumerations vocabulary in concert with unstructured string alternatives.

DP: We don't have to use strict. DOM style 2 is open. It still works with values that were not defined at that time.
... It might be better to use shared strings.

<brutzman> One path to "go both ways" is to have one attribute that uses only enumerations while another attribute is unstructured string. This can add reliability to encoded stylesheets. An XML rule might only allow one alternative (not both) in a given CSS XML instance document.

<brutzman> There are multiple merits and few downsides to a "best of both worlds" approach.

<brutzman> Given the carefully defined data model you have found for CSS, we should be able to express this exactly in XML Schema.

DP: They are using WebIDL.

<brutzman> https://www.w3.org/TR/WebIDL/

DB: Here is EXI 2.0 possibility. Support for WebIDL.
... WebIDL to schema.

DP: The idea was to separate EXI spec from schema mapping.
... apart from EXI spec, you can have multiple schema mappings.

<brutzman> WebIDL Parser for Javascript by Robin Berjon http://berjon.com/webidl-parser

DP: WebIDL to grammar sounds good.

<brutzman> Curiously an "existence proof" mapping might exist already: WebIDL -> JSON -> EXI JSON's XML schema. Perhaps not optimal, but might have full coverage already. We could then improve the generalization to specialize itypes not handled by JSON.

<brutzman> so if we write an XML Schema for WebIDL, then it provides a path for general schema-informed EXI compression of WebIDL languages?

DP: If EXI core spec talks about only grammars and types, you can plugin various schema languages.
... WebIDL is another schema.

<dape> https://www.w3.org/XML/EXI/wiki/EXI2#Overall_spec_writing_approach_for_future_specs

<brutzman> added bullet there: " provide mappings for XML Schema (as exists now), WebIDL, perhaps the EXI JSON schema "

<brutzman> discussion: we wouldn't have to change formal EBNF grammer used in EXI Recommendation document definitions. Providing different schemas (XML, WebIDL, JSON etc.) which map to that formal grammar makes it directly consumable by various implementers.

<brutzman> WebIDL specification abstract: This document defines an interface definition language, Web IDL, that can be used to describe interfaces that are intended to be implemented in web browsers. Web IDL is an IDL variant with a number of features that allow the behavior of common script objects in the web platform to be specified more readily. How interfaces described with Web IDL correspond to constructs within ECMAScript execution environments is also detailed[CUT]

<brutzman> ... document. It is expected that this document acts as a guide to implementors of already-published specifications, and that newly published specifications reference this document to ensure conforming implementations of interfaces are interoperable."

<brutzman> So interestingly, WebIDL would appear to match an object model rather than the broader "document model" of PSVI.

<brutzman> Anyway, pursuing this approach seems like a very powerful possibility.

<brutzman> How about this addendum to 4th EXI2 bullet: " define a standardized EXI grammar format? Using EBNF in the Recommendation itself still seems suitable. "

<brutzman> Wow factor: if we accomplish this, we might claim something like "high-performance EXI compression is possible for any W3C Recommendation with a well-defined data model"

Canonical EXI

DB: I will try to review next week.

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/W3C Recommendation/W3C Recommendation with a well-defined data model/
Found Scribe: TK
Found ScribeNick: taki

WARNING: No "Present: ... " found!
Possibly Present: DB DP TK brutzman dape discussion example exi https 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: 05 Apr 2016
Guessing minutes URL: http://www.w3.org/2016/04/05-exi-minutes.html
People with action items: