[dsdl-discuss] Re: Synchronization of the OASIS RELAX NG Compact Syntax and the ISO/IEC19757-2 AMENDMENT 1: Compact Syntax

From: Martin Bryan <martin@is-thought.co.uk>
Date: Fri Feb 02 2007 - 18:35:56 UTC

Murata-san

Should this document become a discussion item for the Oslo meeting?

Martin

----- Original Message -----
From: "MURATA Makoto (FAMILY Given)" <eb2m-mrt@asahi-net.or.jp>
To: <dsdl-discuss@dsdl.org>
Sent: Friday, February 02, 2007 4:16 PM
Subject: [dsdl-discuss] Synchronization of the OASIS RELAX NG Compact Syntax
and the ISO/IEC19757-2 AMENDMENT 1: Compact Syntax

> 1. Problems
>
> The compact syntax of RELAX NG is defined by two specifications, namely
> OASIS RELAX NG Compact Syntax (OASIS RNC for short) and ISO/IEC 19757-2
> AMENDMENT 1: Compact Syntax (Amd1 for short). Unfortunately, there are
> some important differences between these specifications. First, OASIS
> RNC does not have the media type registration of
> application/relax-ng-compact-sy ntax, while Amd1 does. Second, Amd1
> does not provide semantics, while OASIS RNC does. Third, references and
> bibliographies of the OASIS RNC are different from those of Amd1.
>
> 2. Proposal
>
> OASIS RNC should be complete, since developers rely on this spec. In
> particular, the media type registration of
application/relax-ng-compact-syntax
> should be added to OASIS RNC. OASIS RELAX NG TC can do this by
> publishing a new OASIS committee specification or OASIS standard.
>
> The addition of the semantics to Amd1 is a long term goal. It
> probably requires the republish of Amd1. Minor bugs should be
> fixed by technical corrigenda.
>
>
> 2.1 Minor bugs of Amd1
>
> 1) "RELAX NG DTD Compatibility" is not referenced.
>
> Since annotations are used in C.5 of Amd1, this document should ideally
> be cited as a normative reference. However, since it is merely an
> OASIS committee specification, it cannot become a normative reference
> of an ISO/IEC standard. All we can do is to add an entry in the
> bibliography of Amd1.
>
> 2) Unicode and ISO/IEC 10646 are not referenced.
>
> Since the compact syntax has a escape mechanism for Unicode characters,
> ISO/IEC 10646 should be cited as normative references. It is
> not clear whether Unicode can be added as a normative reference of an
> ISO/IEC standard, but 10646 can certainly be added.
>
> 3) "Guidelines for using W3C XML Schema Datatypes with RELAX NG" is not
> referenced.
>
> Since an example in C.4 references to a datatype in W3C XML Schema
> Part 2, this document should be referenced as an entry in the
> bibliography of Amd1.
>
> 4) "W3C XML Schema Part 2" is referenced incorrectly.
>
> This document is already referenced as an entry in the bibliography of
> Amd1. It is used to define "NCName", which is used in the BNF in Amd1.
> However, OASIS RNC does NOT use W3C XML Schema for defining "NCName",
> but rather relies on "Namespaces in XML".
>
> I believe that OASIS RNC is correct and Amd1 is incorrect,
> since NCName as defined in W3C XML Schema does not allow the
> escape mechanism of RNC. Moreover, NCName as defined in W3C XML
> Schema allows leading/trailing whitespace, which should not be
> allowed.
>
> 3. Misc
>
> 1)
>
> Amd1 says "Add the following sentence at the end of Clause 5", where
> the following sentence is "An alternative compact syntax is described in
Annex C."
>
> Since Clause 5 of 19757-2 is "Data model", this instuction is incorrect.
> Rather, this sentence should be added at the end of Clause 6, which is
> "Full syntax".
>
> 2)
>
> C.7 of Amd1 says that the media type for RNC is "being reviewed by the
> IESG". However, since it is already registered at IANA.
>
>
> Cheers,
> Makoto
>
>
> --
> DSDL members discussion list
>
> To unsubscribe, please send a message with the
> command "unsubscribe" to dsdl-discuss-request@dsdl.org
> (mailto:dsdl-discuss-request@dsdl.org?Subject=unsubscribe)
>
>

--
DSDL members discussion list
To unsubscribe, please send a message with the
command  "unsubscribe" to dsdl-discuss-request@dsdl.org
(mailto:dsdl-discuss-request@dsdl.org?Subject=unsubscribe)
Received on Fri Feb 2 19:40:18 2007

This archive was generated by hypermail 2.1.8 : Fri Feb 09 2007 - 02:03:03 UTC