[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: Minutes of the McLean UBL TC meeting, 2005.01.31-02.04
##################################################################
MINUTES OF THE MCLEAN UBL TC MEETING 31 JANUARY - 4 FEBRUARY 2005
##################################################################
Note that these minutes cover only the TC plenary sessions, during
which the TC was not quorate but operated under our standing rules
as if it were a committee of the whole empowered to report back
(in this very report) to the full TC via email. Meeting time not
devoted to these plenary discussions was mainly spent working on
NDR issues.
IMPORTANT: Voting members of the UBL TC have one week from the
posting of these minutes to register any disagreement with the
recommendations of the TC subgroup meeting in McLean. Items for
which silence will be taken as consent are signaled by the
appearance of "Agreed:" at the beginning of the item.
PARTICIPANTS
Jon Bosak (chair) [M-F]
Marty Burns [M]
Mavis Cournane [M-Th]
Mark Crawford (vice chair) [M-Th]
Michael Dill [W-F]
Scott Edson [T-F]
Jessica Glace [M-F]
Mike Grimley [MTuThF]
Betty Harvey [Th]
Anne Hendry [M-F]
Sue Probert [Tu-F]
Ken Sall [Tu-Th]
Paul Thorpe [M-Th]
By phone: Mikkel Brun [Tu]
Marty Burns [Tu]
Sally Chan [M]
Stephen Green [MWTh]
Tim McGrath [TuTh]
Marion Royal [WF]
Sylvia Webb [MTu]
Unidentified Lurker [Th]
##################################################################
MONDAY MORNING (OPENING) PLENARY 31 JANUARY 2005
##################################################################
HISC REPORT
The SC needs wider participation for conference calls but
appears to be making progress by mail.
SSC REPORT (Anne Hendry)
SSC spent most of the quarter identifying schema tool issues
and trying to resolve them. The process of aligning 1.0
schemas and spreadsheets to provide the basis for future schema
generation is now complete. The SC must still document the
differences between the 1.0 schemas and the tool. Regarding
the user FAQ, the SSC has captured questions and answers but
still needs to organize this material.
NDR TEAM REPORT (Mavis Cournane)
The team has taken a second pass at the issues list and
categorized them. The most critical are those holding up
EDIFIX [EF] work; these will be dealt with first.
CODE LIST TEAM REPORT (Marty Burns)
The requirements doc has been sent out again with the
requirements further elaborated to make them easier to
understand and a request for ratings of relative importance.
Nothing has been received yet.
ATG LIAISON REPORT (Mark Crawford)
ATG met face to face in Wollongong. ATG2 has approved its NDRs
and submitted them to the ATG for approval, expected today.
The NDRs include updated CCTS modules, largely based on input
from OAG. There is a new qualified datatype module with a
number of qualified DTs derived from XSD rather than from the
unqualified DTs; these are being submitted to CCTS as
candidates for additional representation terms.
On behalf of ATG2 to UBL: please look at the implementation
verification schemas. ATG2 believes that it has accommodated
most of the UBL issues. There are still differences, the most
important being (1) the base types and type definitions and (2)
the fact the the ATG2 modules are not imported, just
referenced. If there is any reason that UBL can't adopt the
proposed schema modules in UBL 2.0, ATG2 needs to know that
now.
MartyB: What is the difference between the ATG, UBL, and
Department of the Navy (DON) NDRs?
MarkC/MikeG: All three are implementations of ebXML Core
Components (ISO 15000-5) with the UBL NDRs as the source for
the other two. The DON version combines CBC and CAC into a
single schema module and features a more robust namespace
approach with two categories, Enterprise and Developmental;
"Enterprise" mimics the UBL namespaces. In UBL, we are
building a single standard and aren't yet trying to support
multiple lines of business in the library, whereas the DON has
about two dozen namespace coordinators aligned with different
functional areas, so the scope is much larger. ATG follows the
same basic approach as UBL but (like DON) combines CBC and CAC;
the big difference between ATG and UBL lies in the approach to
local element declaration.
MarkC: The UN/CEFACT CC library was released last week. The
registry is in progress but has not been built yet. A major
work item for UBL will be alignment with the emerging TBG17 CC
library.
JonB: The tasks are to (a) review our names, (b) review our
content models, and (c) identify any CCs missing from the
approved library. If CCs are missing, or the TBG17 content
needs expansion, we will need to petition them to add CCs or
more content; if the names differ, changes will have to wait
for UBL 2.0.
Agreed: A UBL 1.1 work item is to submit our new BIEs as CCs;
for any new BIEs created in 1.1, we should either identify the
underlying CC from the approved list or submit a candidate CC.
CALENDAR REVIEW
MoU/MG meeting: JonB to ask whether a phone presentation would
be sufficient.
Add Open Standards Day in Moscow 26 April; MarkC has submitted
a paper.
Hangzhou: TimM and MavisC are planning to meet in Shanghai and
travel together to Hangzhou by train; others are invited to
join them. MikeG and AnneH are trying for approval. JonB to
canvass the ubl list for commitments [this will happen after
the lunar new year holiday is over].
The CEFACT plenary is the week of 20 June (not May), probably
in Geneva.
The ATG interim meeting will be 27 June through 1 July in Oslo.
Adobe has confirmed hosting for the UBL TC meeting in Ottawa
8-12 August 2005.
CCTS will meet 21-25 February in McLean to kick off the next
version (CCTS 3.0).
Agreed: We will hold the TC meeting formerly scheduled for the
beginning of November in Santa Clara the week of 5 December
2005 instead.
The remaining X12 meetings for this year are 5-10 June in San
Diego and 25-30 September in Atlanta; SylviaW will attend.
##################################################################
MONDAY AFTERNOON PLENARY 31 JANUARY 2005
##################################################################
2005 WORK SCHEDULE
This discussion merged the UBL 1.1 schedule arrived at in Santa
Clara with the UBL 1.0 data dictionary schedule recorded in the
minutes of the 25 January Pacific TC call and adjusted the
latter half of the time line to include the shift of the
November UBL TC meeting to December (see minutes of the Monday
morning session above). A few details added during the closing
plenary Friday morning are included here for convenience.
Agreed: REVISED 2005 WORK SCHEDULE
01 Jan 2005 Completion of 1.0 translations
14 Feb 2005 Begin public debate on use of substitution groups
for codelist extension
23 Feb 2005 UBL TC completes NDR review and specification of schema
layout
24 Feb 2005 UBL SSC issues list of proposed non-codelist changes
to be made to the schema generator
28 Feb 2005 End public debate on substitution groups
02 Mar 2005 UBL TC signs off on proposed (non-codelist) schema
generator changes, sends list to GEFEG for
implementation
31 Mar 2005 GEFEG completes approved NDR and possible code list
modifications to EDIFIX
15 Apr 2005 Cutoff for any major additions to 1.1 submitted in
our format and accompanied by whatever resources
are needed to accomplish the work; cutoff for any
previously unforeseen changes to the UBL 1.1 NDRs;
cutoff for any changes to the UBL 1.1 codelist
spec, including the outcome of the current codelist
work
01 May 2005 Comments close on the Multilingual Data Dictonary
(MLDD) for UBL 1.0, including possible (but not
likely) suggested modifications to the English data
definitions; this must be properly collated for
work in Hangzhou
09 May 2005 All-week UBL TC meeting in Hangzhou hosted by the
Chinese National Institute of Standards: work on
content additions to UBL 1.1; review NDR 1.1 and CL
1.1; last MLDD 1.0 changes; consideration of
possible suggested changes to the English 1.1
definitions
01 Jun 2005 GEFEG implements final changes to schema generator
15 Jun 2005 Begin Public Review Number 1 (data model only)
(Summer vacations)
08 Aug 2005 All-week UBL TC meeting in Ottawa hosted by Adobe:
Comment disposition, Review Number 1; begin final
schema generation for Review Number 2
01 Sep 2005 Package assembly for Review Number 2
15 Sep 2005 Begin Public Review Number 2 (entire 1.1 package)
15 Oct 2005 Comment disposition and repackaging
15 Nov 2005 1.1 internal UBL CD vote begins
01 Dec 2005 CD approved by start of UBL TC meeting;
begin OASIS one-month public review (etc.)
CODE LISTS
Note that this discussion continued Tuesday morning (see
below). References are to the draft circulated by Marty Burns.
2.2.3 [R3] Data and Metadata model separate from Schema
representation
Agreed: We don't see the need to provide a recommended
schema AND in addition a more abstract representation of our
data model for code lists. If an abstraction is needed, it
should be possible to construct it from the schema. If
there are features not in the schema, the solution is to add
the requisite features to the schema.
2.4.9 Ability to prevent extension or modification
Agreed: This is a requirement that is already met by UBL's
use of the "final" attribute. So this is not actually a
future requirement.
2.5.2 [R20 (Future)] Multiple lists of equivalent values
(codes) for a code list
We don't think that the real req here is to support the
normative definition of equivalent code values but rather
the normative association of code values with metadata. We
don't see a need to actually support codes used as
equivalent keys. [See continuation Tuesday.]
2.5.8 [R26 (Future)] The ability to import, extend, and/or
restrict values and elements of other code lists
This needs further discussion [see continuation Tuesday].
Key questions: What are the alternative ways this could be
done? How onerous are they?
##################################################################
TUESDAY MORNING PLENARY 1 FEBRUARY 2005
##################################################################
CODE LISTS (CONT.)
2.5.2 [R20 (Future)] Multiple lists of equivalent values
(codes) for a code list
We can distinguish some different kinds of code lists:
1. Tabular, lots of metadata
2. International standard code list with version and agency,
but user wants to use differently; e.g., currency codes,
country codes
3. Like EDIFACT or X12 code lists; much more likely to be
changed by code list users; e.g., type of document, or
type of party sending a shipping container
4. User defined
We observe that there is a difference between (a) a code
list that has one set of keys and a lot of highly organized
metadata and (b) a code list (such as the example in the
draft) that has multiple keys, e.g. a currency code list in
which the alpha and numeric forms are merged.
Agreed: We accept the need to accommodate (a). With regard
to (b), we suspect that multiple keys would be expensive to
implement and not that interesting for small and medium-size
enterprises; also, simple schema validation would become
difficult or impossible. So we don't see that meeting this
part of the proposed requirment would benefit the UBL target
audience.
2.5.8 [R26 (Future)] The ability to import, extend, and/or
restrict values and elements of other code lists
Agreed: We will conduct a public discussion of the extension
question on ubl-dev starting in mid-February and ending at
the end of the month. JonB will moderate.
For a summary of the discussion that occurred at this point
in the meeting, see the message kicking off the ubl-dev
debate:
http://lists.oasis-open.org/archives/ubl-dev/200502/msg00006.html
##################################################################
WEDNESDAY MORNING PLENARY 2 FEBRUARY 2005
##################################################################
UBL 1.0 ERRATA
Known problems with the UBL 1.0 Standard:
- Wrong content for a file in xsdrt
- Extra space character
The first problem appears to be due to some kind of
installation error at OASIS, though we can't figure out how
this could have happened; the UBL 1.0 zip file we submitted and
the copy of that zip file at OASIS both have the correct
content.
JonB to contact OASIS to get this fixed.
The second problem (the extra space char) brings up the larger
question of how to handle such errata and exposes the fact that
OASIS has no procedure for correcting obvious errors during the
familiarization period leading up to an OASIS vote on
standardization.
Agreed: We recommend that OASIS implement a correction process
modeled on common practice in international standards bodies
such as ISO. For example, the standardization process could be
modified to allow the OASIS Administrator to correct what in
his or her judgement are obvious nonsubstantive clerical errors
in candidate standards in the familiarization phase upon
notification from the originating TC, with the assumption that
the Administrator will then notify the OASIS membership of the
change.
Agreed: Given our belief that the extra space is not a problem
in practice for conforming XML processors, we will address this
particular erratum through an errata document, referenced via a
jump table linked from the UBL TC page, that advises users to
delete the extra space from their copies of the file
UBL-CodeList-CountryIdentificationCode-1.0.xsd, notes that the
error also exists in the corresponding spreadsheet, warns them
that there may be an interoperability problem for nonconforming
XML processors, and notes that this will be fixed in UBL 1.1, a
notice to this effect to be posted to ubl-dev.
Agreed: Ken Sall is appointed our errata maintainer.
KenS to prepare an errata document for UBL 1.0 modeled after
the one used for SAML, its first entry addressing the trailing
space problem and based on mail from JonB to the ubl list Mon,
29 Nov 2004 15:35:29 -0800 (PST).
CUSTOMIZATION QUESTION
Background: See
http://lists.oasis-open.org/archives/ubl-dev/200501/msg00050.html
Agreed: Mark to post this response to Juha's questions:
| We've looked at your figure and think it's correct.
|
| > Case: I don't want to customize document schema, I just want to
| > import my own namespaces to allow use of customized components via
| > "xsi:type".
|
| You can either take this approach (which means that xsi:type
| in the instances now points to your customized type, which
| is defined in your customized CBC or CAC schema module) or
| you can customize the document schema to reference the added
| elements (with their own declared namespace prefix) from the
| namespace you've imported.
|
| > Is it UBL compliant if..
|
| We use the term "UBL compatible."
|
| > I just add those imports and change targetNamespace to my own?
|
| Yes, this is UBL compatible if you're using xsi:type in the
| instance.
|
| > Or should I extend the UBL document schema to sustain the
| > linkage?
|
| We think you're asking whether you should extend the UBL
| library schemas (CAC and CBC modules); the answer to that is
| No. Customizations should be done in your own industry
| schema modules. The UBL imported schemas should be left
| untouched.
|
| (By the way, if you created your own specialized data types,
| you would have to create and maintain your own specialized
| data type module. And that customized SDT module would have
| to directly import at a minimum the UBL UDT module, and
| might also, depending on the data type you are creating,
| have to import the UBL SDT module.)
##################################################################
WEDNESDAY AFTERNOON PLENARY 2 FEBRUARY 2005
##################################################################
SMALL BUSINESS SUBSET
Agreed: OK to recommend SBS to Altova. We may also wish to
recommend the UBL subset being developed by the JPLSC for use
in Japan.
JonB to convey this to Altova.
JonB to see whether Adobe, Blast Radius (Paul Prescod),
Topologi (Rick Jelliffe), and Tibco (Jason Scott) might also be
interested.
We discussed the place of SBS in the UBL work (all development
so far has taken place in ubl-dev). Alternatives ranged from
leaving it entirely alone to recommending the creation of a new
OASIS TC. We concluded that we approve of the idea of
officially recognized subsets and think that some kind of
formal recognition is necessary in order to give the spec a
persistent location, allow it to be referenced in trading
partner agreements, and put it under the same royalty-free
umbrella as the rest of UBL.
Agreed: We will form a Small Business Subcommittee (SBSC) whose
deliverables will be a set of XPaths, instance documents, and
prose.
StephenG and GKenH to work up a charter proposal for SBSC to be
discussed in an upcoming Atlantic TC call.
With regard to Altova's request for UBL namespaces to be used
in their XML catalog, we believe that "all relevant UBL
namespaces" means for all the schema modules, and "all URI
pointers" means references to those modules. We note that we
should revise the URLs for the UBL schemas to update the name
(it shouldn't be "cd" any more) in line with the latest OASIS
naming rules, change the internal references in line with this,
and get the revised files copied to a new location.
UBL TC to review the naming rules and determine what we should
do here.
Agreed: We have to keep the current URLs working regardless, so
there's no harm in recommending that Altova use those in their
catalog.
##################################################################
THURSDAY MORNING PLENARY 3 FEBRUARY 2005
##################################################################
UBL AND UN/CEFACT
JonB: At the November 2004 MoU/MG meeting, I reported to the MG
that a group consisting of myself, Jean Kubler, Francois
Vuilleumier, Sophie Clivio, Mark Palmer, Jean-Luc Champion, and
Jamie Clark were considering a convergence of UBL, UNeDocs, and
UNTDED to create a single CEFACT deliverable; that we needed to
investigate the technical feasability of doing this before
suggesting a concrete plan to the various groups involved; but
in advance of that investigation, I felt quite optimistic about
this idea as a direction forward. Jean-Luc added that there
would be a new TBG group for this and that NDR would have to be
split off into ATG. Jean Kubler added that this would be a
document-centric approach and that UBL and eDocs share a
similar approach. Francois added that TC 154 needed to
identify the workload involved in the mapping of TDED to UBL
CCs, first looking at the degree of correlation and then
estimating the workload. Francois suggested starting with UBL
but also keeping in touch with OAG and allowing OAG to comment
and perhaps offer to help.
SueP conveyed the invitation from UN/CEFACT FMG for UBL to join
UN/CEFACT:
http://lists.oasis-open.org/archives/ubl/200502/msg00004.html
JonB gave a summary of the history of UBL and UN/CEFACT and the
relationship between UBL and UNeDocs.
Agreed: We are in favor of moving UBL into a UN/CEFACT TBG with
UNeDocs after the release of UBL 1.1 if appropriate conditions
are met and stated in a public roadmap (MoU) agreed to by both
OASIS and CEFACT.
Agreed: A negotiating team consisting of JonB, MarkC, SueP,
TimM, and MichaelD is empowered to work out a draft agreement
with UN/CEFACT to be submitted to the TC for its approval,
taking the previous round of negotiations with the former
UN/CEFACT CSG as its starting point. This team expects to be
corresponding offline with a UN/CEFACT team consisting of
MarkC, SueP, MichaelD, Mike Doran, and Jean Kubler.
Agreed: JonB to draft a response to the FMG and have it
reviewed by the UBL negotiating team before sending it. A
72-hour response time is assumed for all discussions among
members of the UBL team. The response will include a
suggestion regarding NDR adoption.
##################################################################
FRIDAY MORNING (CLOSING) PLENARY 4 FEBRUARY 2005
##################################################################
MarionR to talk to SueP about liaison with TBG17.
NDR (MikeG): The group went through the comments and issues
list, addressed all NDR-related comments, made appropriate
changes to the NDR text and rules, and updated the issues list
to reflect this. The NDR-related issues are completed.
MavisC to issue a revised NDR issues list.
UBL TC to review the NDRs in committee.
UBL TC to decide on format and content of schema documentation.
SSC (AnneH): The SSC members present participated in the NDR
discussions. Still editing input about the issues list.
JonB: The Thursday afternoon EDIFIX training session exposed
some workflow issues; we should review these. We need EF
licenses for JonB, AnneH, TimM, KenS, BettyH, and JessicaG.
Agreed: TC and SSC not meeting next week, will resume the week
after.
SueP and MichaelD will probably not be able to attend the TC
meeting in Hangzhou.
The UBL TC thanks LMI Government Consulting for their kind
hospitality and JessicaG in particular for her work in
organizing our visit.
Jon Bosak
Chair, OASIS UBL TC
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]