See also: IRC log
<scribe> scribe: TK
<scribe> scribeNick: taki
https://lists.w3.org/Archives/Member/chairs/2016AprJun/0033.html
CB: I will take care of WebEX change.
TK: Telecons will be on Thursday till June 16th
TK: I measured some numbers using
one example JSON.
... The current form is 313 bytes and the new one suggested by
Daniel is 332 bytes.
DP: Element Fragment Grammar contains the list of all element names.
TK: We may want to do it for EXI2.
DP: The proposal I made at that time was to use XSD anyType grammar.
<dape_> https://www.w3.org/XML/EXI/wiki/EXI2#Fallback_grammar
DP: In EXI2 wish list, it already
say we want easier mechanism to achieve a similar effect as EXI
profile.
... EXI has version numbers.
<dape_> https://www.w3.org/TR/exi/#version
<dape_> 1 0000 Preview version 1
<dape_> 0 0000 Final version 1
DP: Version can be a preview version.
<dape_> 1 0001 Preview version 2
DP: We can create a note by using
preview version numbers.
... When implementation use the extension, you need to use the
preview version.
... For shared-string work, we use DTRM to signal the
change.
... I am interested in measuring how much effect the change
would make by modifying EXIficient if I have time.
TK: I will also take a look at how it is easy to make the change in OpenEXI.
DP: Once we feel comfortable, we want to propose as a Note using preview number.
DP: We processed DB and TK
comments.
... We need to settle schemaId. For UTC normalization,
introducing a URL is an idea that we need to ask.
http://www.w3.org/2005/06/tracker/exi/issues/109
DP: Internally, we thought it
would be sufficient.
... I can see some use cases that don't get much benefit from
schemaId.
... DB and Youenn had strong opinion about requiring
schemaId.
... The EXI spec itself was silent about schemaId.
... If we get to EXI2, we may need to rethink about this
aspect.
... we may want to use hash of XML schema, as Daniel Vogelheim
suggested.
DB: What if there was not a
schema?
... We need to provide an explicit warning about schemaId's
importance.
DP: I agree. EXI 1.0 spec
provides only a help by providing schemaId.
... In the next EXI, we need to repair it.
... XMPP uses multiple schemas. They create a hash over
schemas. Hash identifies the schema.
... We need to mention schemaId "should" be present in the
spec, and propose it to the public.
DB: The consequence of not doing it may be the recipient may not be able to decode it.
DP: It is a problem common to schema-based formats.
<brutzman> https://lists.w3.org/Archives/Public/public-exi/2016May/0004.html
<brutzman> RFC 3986 is the latest IETF reference for URI/URL. It has an excellent section on normalization and comparison.
DP: It is desirable to have
normalization for URIs.
... EXI implementations currently do not know when it encounter
URIs.
<brutzman> There is a "stairstep" of increasingly strict checks that are possible. These are mostly application dependent.
DB: We can refer to the section.
<brutzman> So we might refer to this section, point out that C14N of uris should follow that guidance, and leave it up to the application.
DB: And suggest applications should follow that suggestion.
DP: We have one section about
design decision.
... Is it sufficient that we describe it in there?
<brutzman> RFC 3896 only has a little about normalization https://tools.ietf.org/html/rfc3986#section-3.1
DP: If Don, you can propose a text for this, that would be great.
DB: Just a simple sentence, that would be fine.
<brutzman> Suggested sentence: "Normalization and comparison of url and uri values, if applied within a document, is best performed in accordance with [RFC 3986]."
<brutzman> If we could discuss on the mailing list how EXI supports the full range of float and double values supported by XML Schema, that would be great. Not sure I understand the draft prose, all help appreciated.
<brutzman> Seems like we are close... just need to finish checking the mail threads to ensure all points were addressed. Thanks for the steady great efforts!
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/2896/3986/ Found Scribe: TK Found ScribeNick: taki Present: Taki Daniel Carine Don Found Date: 10 May 2016 Guessing minutes URL: http://www.w3.org/2016/05/10-exi-minutes.html People with action items:[End of scribe.perl diagnostic output]