Claus,
I do
not believe that is necessary as the case presented in D.2 has rather limited
scope and demonstrates well what it intends to.
My
point is that there is very little in the way of a meaningful strategy for the
inquiring application to determine which name is more suitable for
presentation to the user if multiple names in the same language are
specified. In my previous message I introduced the notion of content
context, which we may or may not want to be implemented in the spec. By
applying context qualifiers, the publisher may offer clues to the inquiring
application as to which name is most appropriate for the given context.
Also, acronymic form of a name is rather
standard, but is not accounted for in the schema.
I am
wondering whether we need to more thoroughly model the name structure to address
the issues of language, script, context (legal, "proverbial", marketing),
acronymic equivalence, etc.
Daniel
Daniel,
Do
you think that we have to add more examples and explanations to the UDDI
V3 Specification, Appendix D.2 Multiple names in the same
language?
Claus
-----Original Message----- From:
Daniel Feygin [mailto:feygin@unitspace.com] Sent: Freitag, 4. Juli
2003 09:11 To: uddi-spec@lists.oasis-open.org Subject: RE:
[uddi-spec] Omission of use="optional" on the description
element
SAP, whether it's written in Latin or non-Latin alphabet, remains
meaningless with respect to the language in which it's written and therefore
what's significant is the script - and not the language -
qualifier.
This issue touches upon the intended use of content in UDDI
registries. Content intended for developers has different
representation requirements than content intended for marketing or
legal use; there may conceivably be other contexts too. For
instance, legal name of OASIS is SGML Open, but it's doing business as
OASIS. Also, SAP and IBM stand for various things (some more obscure
than others <g>), which may or may not be specified in their complete
form depending on intended context.
Is
this something worth addressing in the spec?
Daniel
Claus,
I agree with all what you say, but I have
just one small remark:
1. Select the encoding for this message as
UTF-8
2. Read this
САП
This how Russians call SAP. It sounds the
same, but they use their alphabet. I suspect that Chinese, Japanese and
other non-Latin alphabets have their own
transcriptions/interpretations.
Cheers,
Max
----- Original Message -----
Sent: Friday, July 04, 2003 12:12
AM
Subject: RE: [uddi-spec] Omission
of use="optional" on the description element
There was a lengthy discussion about requiring
the xml:lang attribute for names during V3 development and the V3
authors consciously decided to specify it as being optional. Names like
"SAP" simply don't have an associated language. See section 3.3.2.3 of
the V3 specification for more details. Also, using statically defined or
out-of-band negotiated default values can invalidate digital
signatures.
A different thing are descriptions. I tend to
agree that descriptive text always has an associated language and
believe that the V3 authors also wanted to specify this. Interestingly,
the V3 specification (section 3.3.2.4) and the V3 API schema are in sync
and BOTH specify the xml:lang attribute as being optional for
descriptions. This is true since not specifying the use attribute for
attribute declarations (as is done for xml:lang within the description
element) results in an optional attribute (see XML Schema specification
at http://www.w3.org/TR/xmlschema-1/#cAttribute_Declarations).
In case we were trying to make xml:lang
mandatory within descriptions, we would have a problem for migrating
data from V2 to V3 (xml:lang is also optional in V2). Which language
should be applied to those descriptions that don't carry an xml:lang
value? Maybe this was the reason why xml:lang is still specified as
being optional in V3.
Regards,
Claus
I do not see how we can have a default value
without providing a way for the client to find out what it is. A
publisher may have a different idea of what the default value is than
the inquirer, which means that an out-of-band implied agreement must
exist between all users of a registry. A node, on the other
hand, is not involved in the interpretation of content (is it?), so it
is not a party to such agreement. A node may be used though to
establish the agreement between users by publishing a special "default
language" descriptor in the node businessEntity. It could be
either in the businessEntity's categoryBag or down at the service
level to allow for multiple different default languages on multiple
services or endpoints, if that's a realistic requirement. Sounds
like TN material?
Replication also complicates things
somewhat...
My opinion is that it would be
worthwhile to make xml:lang mandatory
everywhere. This would eliminate the need for users to agree on
a default. If the registry is used by more than one language
community, it would be very difficult to establish and meaningfully
enforce a default language.
Daniel
My reading of the standard suggested that it was
optional, and that its omission indicated that the text was in the
"default" language, not that the idea of default language is
adequately described - is it default for the node, or default for
the user?
It used to be that only one entry could use any given xml:lang
value (including default), but that restriction has been eased in
V3. As I see it, any number of entries could be coded to the
default, so optional seems a valid way of indicating this.
I'd be in favour of drawing it into line with the treatment of
xml:lang for names - the two are analogous, as I see it.
Tony Rogers
-----Original Message----- From: Tom
Bellwood [mailto:bellwood@us.ibm.com] Sent: Thu
19/06/2003 9:12 To: Luc Clement Cc:
uddi-spec@lists.oasis-open.org Subject: Re: [uddi-spec]
Omission of use="optional" on the description
element
Yes, the xml:lang is optional everywhere else but
for descriptions. I agree we weren't consistent in its
treatment. I also seem to recall that it was
intentionally left as required here because descriptions
are intended to be human readable text and having the xml:lang
can be important for such uses. We should consider
if this line of reasoning is important before making it
optional I think.
Other opinions? Someone with a
different recollection than mine here?
Thanks, Tom
Bellwood Phone: (512)
838-9957 (external); TL:
678/9957 (internal) Co-Chair, OASIS UDDI Specification
TC STSM - Emerging Technologies IBM Corporation
"Luc
Clement" <lclement@windows.microsoft.com> on 06/18/2003
01:39:47 PM
To:
<uddi-spec@lists.oasis-open.org> cc: Subject:
[uddi-spec] Omission of use="optional" on the
description
element
Tom / TC,
Please note that
the http://uddi.org/schema/uddi_v3.xsd
schema omits use="optional" on the description element. I think
this is an omission and recommend we correct this
definition as part of CR-002. The current schema is
declared
as:
<xsd:elementname="description"type="uddi:description"final="restriction"/>
<xsd:complexTypename="description"final="restriction">
<xsd:simpleContent>
<xsd:extensionbase="uddi:validationTypeString255">
<xsd:attributeref="xml:lang"/>
</xsd:extension>
</xsd:simpleContent>
</xsd:complexType>
... when I think we should have it
declared
as:
<xsd:elementname="description"type="uddi:description"final="restriction"/>
<xsd:complexTypename="description"final="restriction">
<xsd:simpleContent>
<xsd:extensionbase="uddi:validationTypeString255">
<xsd:attributeref="xml:lang"use="optional"/>
</xsd:extension>
</xsd:simpleContent>
</xsd:complexType>
For your
consideration.
Luc
Luc
Clément Microsoft
You may leave a Technical Committee at
any time by visiting http://www.oasis-open.org/apps/org/workgroup/uddi-spec/members/leave_workgroup.php
You
may leave a Technical Committee at any time by visiting http://www.oasis-open.org/apps/org/workgroup/uddi-spec/members/leave_workgroup.php
|