NADM Data Model Design Team (DMDT) Meeting Notes

November 6, 2001, Boston, MA

Executive Summary

Attendees: Éric Boisvert (GSC), Boyan Brodaric (GSC), Bruce Johnson (USGS), Jim McDonald (Ohio GS), Steve Richard (Arizona GS), Peter Schweitzer (USGS), Ron Wahl (USGS), Jerry Weisenfluh (Kentucky GS).

(1) Confirmed: Agenda

(2) Revised: Previous meeting notes

(3) Finalized: Requirements document

· A draft document listing some requirements for geologic (map) information database systems was developed by DMDT and circulated prior to the meeting.

· Minor changes to the requirements document were suggested and accepted.

· NADMSC will be solicited for their interest in reviewing the requirements document prior to its posting to the DMDT web site.

(4) Defined: Variant data models document format

· Several data models and implementations derived from NADM 4.3 (i.e. ‘variants’) were documented and circulated to DMDT prior to the meeting.

· It was determined that a consistent format was needed for these documents, and one was developed by DMDT at the meeting.

· Authors of variants will be contacted and asked to revise their documents according to the standard template by Jan. 1, 2002, when they will be posted to the DMDT web site.

(5) Future Work: Data model upgrade

· A new data model version prototype will be generated for the DMDT meeting to be held during the May 19-22’02 DMT workshop (Salt Lake City, Utah).

· The prototype will be produced by Boyan Brodaric, Bruce Johnson and Steve Richard, and will primarily build upon v4.3 while considering aspects of the variants.

· The new data model version will be assessed by DMDT at its next meeting.

NADM Data Model Design Team (DMDT) Meeting Notes November 6, 2001, Boston, MA

Meeting opened 1:45pm.

(1) Confirmed: Agenda

(2) Revised: Previous meeting notes

Previous meeting's notes were accepted, and Peter is asked to post them to the DMDT web site.

(3) Finalized: Requirements document

The requirements/scoping/use-case document is a distillation of the many queries that Jon Matti compiled. It is intended to convey to readers a sense of what questions people might ask of data represented in the data model, and to allow people to judge the effectiveness of various implementations of the data model.

The unfiltered, large list from Matti should be available on the DMDT web site as well; its purpose is to allow people to see how we got here from the unedited responses to Matti's call for queries. It was decided to ask Jon for a list of the agencies originally contributing to the queries as background for the queries.

There exist several earlier compilations of the queries, e.g. one was developed by Bruce Johnson, mainly with respect to the capabilities of 4.3. These original analyses by Jon, Bruce, Steve, and Jerry should be sent to Peter and made available on the DMDT site.

Boyan would like to close this exercise, and seeks approval of the document (‘as is’ with minor revisions), as opposed to undertaking another major analysis of the questions. The group agrees. Minor revisions would include moving into the relationships section those items currently under classification that deal with relationships. Other minor revisions should be sent to Boyan by email. The steering committee will be asked if it wants to approve the revised document.

(4) Defined: Variant data models document format

Descriptions of v4.3 data model variants have been received from Loudon Stanford, Steve Richard, and Eric Boisvert. We expect an additional report from the team of Boyan, Jordan, and Jerry, and request an additional description of USGS implementations of v4.3 to be coordinated by Bruce Johnson.

Part of the motivation for this exercise is to help answer the question "so what are people doing with the data model?". Another motivation is to help potential implementers select an appropriate variant to work with. However the most important motivation is to suggest changes to be incorporated into the upcoming revision of the data model.

Discussion of a standard data model variant document ensued. It was decided that each variant description should be provided in PDF format and should include the following:

o Abstract (maximum 2 letter-size pages, 12 pt font)

§ Metadata: Author, Agency, Date.

§ Intention: the reason for developing the variant.

§ Category: one of conceptual (a variant data model) or implementation (a specific implementation of 4.3 or a variant model).

§ A brief summary description of the variant.

o Diagram (maximum 1 sheet, any size)

o Comparison with 4.3 (maximum 2 letter-size pages, 12 pt)

o Example—a representative geologic unit description (maximum 1 letter-size page, 12 pt)

Variant documents will be circulated within DMDT for comment before being posted to the web site. The documentation will be complete and posted to the web site by Jan. 1.

Some comments:

One suggestion was to have all variants demonstrate how a representative geologic unit (to be selected by DMDT from a short list to be provided by Steve) is represented: e.g. show its traditional textual description alongside a list showing which database fields and values are used to store the information. Eric points out that this presentation would need to include some of the table linkages as well as field name and field values. The reports resulting from this exercise might take the form of plain text (table:field:value) or a diagram showing how the tables are linked, or both. We hope to use the same format for all of the reports so that they can be more easily compared by readers.

It was noted that it is important to put these documents in proper perspective, so as not to mislead people into thinking that the various model variants are more different than they are; this will be achieved by appropriate placement of the documents within the DMDT web site by Peter.

(6) Future Work: Data model upgrade

It was decided that a new version of the data model is to be developed, beginning Jan. 1, after the variant documentation has been completed. How to accomplish this was discussed. Expected sources for the upgrade include: the requirements/scooping/use case document, flaws identified in 4.3 during its inception, and various new needs that motivated the development of variants.

This unification will be explicitly developed as a major revision of version 4.3, rather than as a complete redesign starting from first principles. The redesign will be expressed using UML, although a corresponding relational E-R diagram will also be provided, eventually. It is understood that the data model will be implementable in relational database management systems as well as by using object-oriented means.

It was decided that revisions to the data model would initially be generated by a small group, likely one to three people, who would then present the draft to the whole team. Boyan, Bruce, and Steve will be the revision team. Their goal will be to submit a proposal for major revision of the data model to the DMDT at the meeting we will have when the next DMT meeting occurs, at Salt Lake City in May of 2002. It will likely require two days of DMDT meeting.

Meeting adjourned at 5:45pm