Re: TA document and Messaging Service spec v0-2 - charsets

[duane <duane@xmlglobal.com>]

All:

I totally agree with Jon Bosak.   The use of UTF-8 should be
encouraged.  The wording from the TA spec states:

"Partipating components in ebXML MUST facilitate  multilingual support. 
ebXML specification MUST be compliant with Unicode and ISO/IEC 10646 for
character set, and UTF-8 or UTF-16 for character encoding."

Another quick note:

There have been a few misgivings about the state of the TA spec.  We
accept ( and encourage) all criticism of the spec and will not take
anything personally.  However,  if something is not to your liking, 
please provide constructive suggestions.  We have worked very hard to
get the current work together in the format mandated to us.  The format
was not to our liking and it is the third such template we have
followed.  Our goals were to define what each component must facilitate
and how they interact with each other.  If an area is weak,  please give
us specific criticism on that part along with suggestions for change.

The TA spec is not complete.  The only way we feel we can get it
complete is to solicit feedback from the plenary.  We don;t wish to
design ebXML in private then spring it on the plenary, only to find the
direction is 180 degrees different from general consensus.

Chris has given us some excellent feedback along with suggestions.  We
want this document to proceed to the plenary tomorrow.  TO keep it back
may impact the ability to deliver other specs  by Vancouver, based on
the 12 week cycle for approval.

I will likely not be able to participate in the conference call tomorrow
due to travelling.  PLease send a copy of QR team findings to me so we
can get to work immediately.  

Duane Nickull


Jon Bosak wrote:
> 
> Just remember that by definition an XML document consists of
> Unicode characters.  The XML spec does indeed allow the use of
> encodings that don't know about Unicode, but it sure doesn't
> encourage them.
> 
> Jon
> 
>    From: "Brunner, Eric" <EBrunner@Engage.com>
>    Cc: tmcgrath@tedis.com.au, duane <duane@xmlglobal.com>,
>       ebXML-Transport@lists.ebxml.org, Jon Bosak <bosak@boethius.eng.sun.com>,
>       joebaran@extol.com, ebxml-coord@lists.ebxml.org
>    Date: Tue, 19 Sep 2000 11:02:08 -0400
>    MIME-Version: 1.0
>    X-Mailer: Internet Mail Service (5.5.2650.21)
>    Content-Type: text/plain;
>            charset="iso-8859-1"
>    Content-Length: 4080
> 
>    Chris,
> 
>    Thanks for pointing out what I missed. The charset reference in the MS Spec
>    does not require utf8, and does not preclude sjis or any other (relatively)
>    compact and de facto (if not de jure) local encodings. I'll look at the TPA
>    as you recommended, and I hope to find a mechanism for encoding negociation.
> 
>    ebcdic and sjis work for me, just as much as ascii and utf8.
> 
>    Eric
> 
>    -----Original Message-----
>    From: Christopher Ferris [mailto:chris.ferris@east.sun.com]
>    Sent: Tuesday, September 19, 2000 10:53 AM
>    To: Brunner, Eric
>    Cc: tmcgrath@tedis.com.au; duane; ebXML-Transport@lists.ebxml.org; Jon
>    Bosak; joebaran@extol.com; ebxml-coord@lists.ebxml.org
>    Subject: Re: Messaging Service spec v0-2 - charsets
> 
>    Eric,
> 
>    I don't understand what you mean by "declining to label the encoding".
>    If you mean declining to explicitly REQUIRE use of a specific
>    character encoding and merely RECOMMENDING that UTF-8 be used,
>    the rationale is that two parties MAY agree to some alternate
>    encoding and that ebXML should not preclude such practices.
> 
>    They do this through the TPA. Please refer to the
>    Trading partner team's work for specifics.
> 
>    Cheers,
> 
>    Chris
> 
>    "Brunner, Eric" wrote:
>    >
>    > I must have missed something. I'd like to see the rational for declining
>    to
>    > label the encoding.
>    >
>    > Two parties using SJIS (or Big5 or GB or ...) MAY convay this encoding to
>    > each other how?
>    >
>    > Two parties using SJIS (or ...) MUST convert their data to utf8 (or
>    > iso8859-*) because why?
>    >
>    > Cheers,
>    > Eric
>    >
>    > -----Original Message-----
>    > From: Christopher Ferris [mailto:chris.ferris@east.sun.com]
>    > Sent: Tuesday, September 19, 2000 10:06 AM
>    > To: tmcgrath@tedis.com.au
>    > Cc: duane; ebXML-Transport@lists.ebxml.org; Jon Bosak;
>    > joebaran@extol.com; ebxml-coord@lists.ebxml.org
>    > Subject: Re: Messaging Service spec v0-2 - charsets
>    >
>    > The spec wording of the charset has been revised in version
>    > 0.21 which is what we plan to publish. That version contains
>    > some edits which were omitted (due to time and other constraints)
>    > from version 0.2 which you reviewed.
>    >
>    > UTF-8 is RECOMMENDED, but not mandated. I don't believe that
>    > we can mandate its use. Fujitsu might have some comment on
>    > this based upon their experience. Bottom line, the charset
>    > used should be pre-agreed (and hence, in the TPA or equivalent)
>    > by both parties.
>    >
>    > Cheers,
>    >
>    > Chris
>    >
>    > Tim McGrath wrote:
>    > >
>    > > as we have concluded our report and suggested the spec now go to public
>    > > review, this thread should now be addressed directly to the TRP team.
>    > >
>    > > i shall forward it to their list.
>    > >
>    > > duane wrote:
>    > >
>    > > > Jon:
>    > > >
>    > > > Yes - you are right.  UTF-8 has been mandated by the Technical
>    > > > Architecture group as minimal.
>    > > >
>    > > > ISO 8859-1 does not represent Kanji characters correctly as far as I
>    > > > know.
>    > > >
>    > > > Duane Nickull
>    > > >
>    > > > Jon Bosak wrote:
>    > > > >
>    > > > > [tmcgrath@tedis.com.au:]
>    > > > >
>    > > > > | * lines 211-215 and then 261-268  (charsets)
>    > > > > |
>    > > > > | Both claim to be "The charset attribute is used to identify the
>    > > > > | character set used to create the message" and yet propose
>    > > > > | different values.  i think it means the outer and inner encoding
>    > > > > | methods but if so it should not use the same first sentence.
>    > > > >
>    > > > > The encoding method could use the iso-8859-1 default value of lines
>    > > > > 213-215 or UTF-8 as recommended in 264-266.  It seems to me that
>    > > > > we should mandate the latter.  Surely UTF-8 is the minimum level
>    > > > > of internationalization for anything calling itself an XML tool.
>    > > > >
>    > > > > Jon
>    > >
>    > > --
>    > > regards
>    > > tim mcgrath
>    > > TEDIS   fremantle  western australia 6160
>    > > phone: +618 93352228  fax: +618 93352142
>    >
>    > --
>    >     _/_/_/_/ _/    _/ _/    _/ Christopher Ferris - Enterprise Architect
>    >    _/       _/    _/ _/_/  _/  Phone: 781-442-3063 or x23063
>    >   _/_/_/_/ _/    _/ _/ _/ _/   Email: chris.ferris@East.Sun.COM
>    >        _/ _/    _/ _/  _/_/    Sun Microsystems,  Mailstop: UBUR03-313
>    > _/_/_/_/  _/_/_/  _/    _/     1 Network Drive Burlington, MA 01803-0903