RE: ACTION-736: Check issue 100 to see canonical exi section 4.2.2 properly handles it

Hi Daniel and all,

The new order makes sense to me.

The item d in the newly ordered list is shown as "AT ( * ) [schema-typed value]". 

In the EXI spec, this event type is written as just "AT ( * )" (without trailing qualification) 
because the value is typed only when there is a global attribute defined for the 
attribute qname, otherwise it is untyped. 

I think we should also write it as "AT ( * )" in this list.

Thank you,

Takuki Kamiya
Fujitsu Laboratories of America


-----Original Message-----
From: Peintner, Daniel (ext) [mailto:daniel.peintner.ext@siemens.com] 
Sent: Wednesday, March 16, 2016 7:53 AM
To: Efficient XML Interchange Working Group
Subject: AW: ACTION-736: Check issue 100 to see canonical exi section 4.2.2 properly handles it

All,

I had the Action to look into issue 100 [1].

I think the Canonical EXI draft version [2] should be ok w.r.t. issue 100.

In step #1 productions which are not usable (e.g., EXI Profile) are removed.

In the second step the event is selected that matches most precisely.

While cross-checking I detected another issue.
I think we better change the order of attribute selection in this section [3].

The current order for attribute events is as follows:
    a. AT ( qname ) [schema-typed value]
    b. AT ( uri : * ) [schema-typed value]
    c. AT ( * ) [schema-typed value]
    d. AT ( qname ) [untyped value]
    e. AT ( * ) [untyped value]

Form the point of view of an implementation I think it is more straight-forward to move "AT ( qname ) [untyped value]" up.
The reason is that an implementation is likely to first check the qname and if the qname exists and it checks the value.
Hence my proposal would be to change the order as follows.

    a. AT ( qname ) [schema-typed value]
    b. AT ( qname ) [untyped value]
    c. AT ( uri : * ) [schema-typed value]
    d. AT ( * ) [schema-typed value]
    e. AT ( * ) [untyped value]

I am a bit unsure whether b or c should come first.

Any thoughts?

Thanks,

-- Daniel

[1] https://www.w3.org/2005/06/tracker/exi/issues/100
[2] https://www.w3.org/XML/EXI/docs/canonical/canonical-exi.html#exiEvents
[3] https://www.w3.org/XML/EXI/docs/canonical/canonical-exi.html#N66353





________________________________
Von: Efficient XML Interchange Working Group Issue Tracker [sysbot+tracker@w3.org]
Gesendet: Dienstag, 8. März 2016 17:09
An: public-exi@w3.org
Betreff: ACTION-736: Check issue 100 to see canonical exi section 4.2.2 properly handles it

ACTION-736: Check issue 100 to see canonical exi section 4.2.2 properly handles it

https://www.w3.org/2005/06/tracker/exi/actions/736<&smime=14.3.123.2https://www.w3.org/2005/06/tracker/exi/actions/736>

Assigned to: Daniel Peintner

Received on Monday, 28 March 2016 19:14:21 UTC