Minutes of UBL Library Content Subcommittee

Tuesday 5th February, 9-11 a.m California time

Attendees

By Phone: Joe Chiusano, Lisa Seaburg, Arofan Gregory, Sally Chan, Mike Adcock, Gunther Stuhec, Ray Seddigh, Sue Probert

Via NetMeeting: Kurt Conrad, Peter Yim, Phil Grifin, Ray Seddigh, Sue Probert, Tim McGrath, VNC User

Agenda

1. Welcome from Chair and appointment of Secretary to take minutes

2. Acceptance of previous minutes

3. Action arising from previous meeting

a. Lisa, Polly, Ray, Sally to model common/core xCBL elements

b. Arofan and Gunther to revise the XML Schema to align with new metadata and apply to 'Party' sample.

c. 'Black Jacket' team to meet via teleconference Thursday at 9:00 Pacific time

d. Peter to coordinate Netmeeting trials.

4. Presentation and discussion on modelling approach and results of initial analysis. (I hope to distribute the latest model before the meeting.. we shall also attempt to use the Netmeeting facility to share the view)

5. Work plan

6. Reports on status of other UBL subcommittees

a. Tools and technique sc (Arofan)

b. Naming and design rules sc (Mark)

c. Context drivers sc (Sue)

d. Context assembly sc (Arofan)

7. Technical Editor's report, web site updates, etc. (Lisa)

8. Other Business

9. Next meeting.

Minutes

1. Welcome from Chair and appointment of Secretary to take minutes

- Kurt to take minutes

2. Acceptance of previous minutes

- Previous minutes accepted

3. Action arising from previous meeting

a. Lisa, Polly, Ray, Sally to model common/core xCBL elements

b. Arofan and Gunther to revise the XML Schema to align with new metadata and apply to 'Party' sample.

- Schedule

- Get feedback today

- Discuss on Thursday

- Present revised version of the spreadsheet structure

next week

c. 'Black Jacket' team to meet via teleconference Thursday at 9:00 Pacific time

- Subgroup (Black Jacket Team) met on Thursday to discuss common aggregates and basic elements

- Goal to do automated cut and past from xCBLl to UBL analysis model

- Now have a hybrid solution

- Led to combined view that will be presented, today

d. Peter to coordinate Netmeeting trials.

- Also experimented with NetMeeting at Black Jacket Team telecon

- Ironed out most of bugs

- Expect to use to share a single file (as a supplement to previous file transfer-based approaches)

- A few people still having firewall problems

- 2 active worksheets

- Order: Provides entry point for this spreadsheet

- Reuse Type: Builds a library of reusable types

- Development since face-to-face

- When describing aggregates in BIE library, really defining reusable types

- So everything that see in analysis that looks like aggregate is likely a reusable type

- All aggregates noted with term "details"

- Focused on party to discuss/resolve aggregation and aggregation extension issues

- Expect 500-600 entries in BIE order catalog

- Issues

- Consider names that have been chosen, especially representation terms, for appropriateness (especially

'text' and 'quantities')

- Where would want to use other terms?

- Ideas about other representation types coming up as part of core component tech spec discussions in Seattle

- A call for issues is to be forwarded to Seattle group

- Don't expect show stoppers, only refinements

- Camel case rules not used consistently

- General discussion of naming rules

- Tension between visual pleasure and struggling to keep structure of names

- Structured into 2-3 components

- Only way to delimit is w/ upper case

- If deviate, have to introduce new thing into model

- These are just tag names

- The link back to structure will be handled by data dictionary

- Keep these names intuitive

- That should follow spirit of naming and design rules, but should pitch issue back to them

- Another key factor is readability outside of English language

- Core component spec assigned IDs for localization (ID allows other names to be associated)

- How build dictionary?

- Dictionary names become tag names

- Naming and design sees dictionary name as absolute path from root to data, absolute unique ID of data

- UBL seems to be following rules of tag names

- Don't need to spell out structural context in tag names

- No decision, yet, but strong possibility that representations will always be part of dictionary name

- Expect to use "/" (pure XPath) to describe steps in dictionary names

- Don't expect to resolve issue today...

- Contacts w/ names and other contacts: How capture that something is a list of things (how to handle repetition)

- Perhaps need a special naming rule for "list of" things

- Show in spreadsheet as "0..n"

- But can't get xsd from that

- Need to distinguish whether want a wrapper around group

- Question whether need to capture both cardinality and semantic relationship

- One approach is to automatically generate wrapper around these structures

- Except that may not always want the wrapper

- Should be capturing enough info to answer that question (still needs to be verified for each table entry)

- Issues list: Lisa started at last face-to-face

- Lisa will log issue on master list which will be posted before each meeting

- They can be submitted via email

5. Work plan

- Target: final draft presented to group next Tuesday...

- Realized that much more effective to do analysis in one place at one time

- Been working on order detail part of model

- Hope to be done by Thursday and present to group on Tuesday

- Ask group to sign off doc as deliverable

- Will we be able to generate XSD by Tuesday?

- Problem w/ reuse types spreadsheet

- Hard to figure out children and substructures within aggregates

- Could Gunther change?

- Subgroup to work this question offline

- Status

- Have defined enough types that mostly reusing, not defining new ones

- Could generate draft docs w/ scripts from xCBL to do some early testing

- Have succeeded in simplifying some of xCBL complexity

- Expect to have groups of people working in verticals on specific types of docs

- Group needs to do some strategic thinking of what it's role is w/ that shift

- Had another 6 people join discussion lists over last week

- Is base info coming through email or to be accessed from some site?

- Need a copy of library content starter kit w/ xCBL/HTML doc (web view) and spreadsheet that have been working at

- Spreadsheet available through Peter's FTP site and has also been emailed out

- All of core component stuff included in the spreadsheet (called "UBL-Library_v0.5_Order_a.xls"

- Not comfortable putting spreadsheet up on web page yet, so get from Peter's FTP site: ubl.cim3.org/~lcsc/CoreComponents

6. Reports on status of other UBL subcommittees

a. Tools and technique sc (Arofan)

- Nothing to report

- Arofan and Guther to meet in Seattle

- Not everone can get through NetMeeting

- There are some web-based alternatives

- Take that discussion offline w/ Peter

- We seem to be getting better at it

- Only 3 people still have problems

- Some question of practicality of workaround

b. Naming and design rules sc (Mark)

- Last meeting canceled

- Expect one tomorrow

- Please pass on that LCSC expects more formal input by end of week

- Sue has list of new numbers for NDRSC telecon (email her if you need copy)

- Would be good if NDR looks at naming conventions in spreadsheet to give any feedback prior to LCSC telecon on Thursday

- Also mail archive publically available through OASIS

c. Context drivers sc (Sue)

d. Context assembly sc (Arofan)

- Analysis conclusions:

- Everytime a type is reused, it appears to point to a new context

- Ties to rules for context drivers

- Close to point of having to consider how to capture that info (added columns to end of spreadsheet for 8 context, but not sure that such really makes sense)

- That should be a topic for discussion next week: How to capture right info for context

- Have a preliminary list of potential values for each context area

- Need to start getting serious about capturing context info

- Understand that role of context assembly group is to define rules for taking info and customizing order doc that UBL produce to be an oarder doc for specific set of organizations

- Have you discussed interoperability of contexts and rules for people generating contexts?

- Know that we have to have lists for context drivers

- Core components need that, as well

- May find that "8" is not the right number

- Who will be group within ubl that will provide articulation of context drivers?

- Will be a new version of core component doc coming out of seattle, which should be reviewed on this topic

- Lisa will send current draft of Core Component technical spec

- Core Component group of EWG

- 1.75 is current version

- Doc is also at ublcim3.org/~lcsc/CoreComponents

- Should also check out ubl home page (john has recently updated)

8. Other Business

- Notes for Phil: Be aware that when working onscreen, working on Peter's machine, not Phil's... saving to Phil's machine

9. Next meeting.

- Lisa to check on number for Thursday call and send out via email

- Kurt: to send minutes to Lisa for review and posting