North American Data Model Steering Committee

Notes of meeting November 7, 2001, Boston, Massachusetts

Attending:

Notes:

  1. Update on science language activities in Canada, provided by Boyan:

    To support the needs of the new national bedrock geology project, a science language is under development. This activity intends to create a hierarchical classification for the interpretive and observable attributes that form the basis of rock names -- genesis, composition, texture, and fabric. Historically, rock names have been based on various individual attributes or combinations thereof, and a unique hierarchy of names has not developed. The present strategy will consider these attributes to be the vocabulary for describing rocks, to be managed as principal descriptive fields that, in various combinations, will be related to typical rock names. Unlike the rock names, attributes can and will be hierarchically classified. Boyan will send the NADMSC a preliminary version that contains rock names directly associated with the four attributes.

    A national surficial geologic mapping project also has begun this year. It is likely that a similar approach to science language will be devised, with additional attributes (e.g., morphology). Andy Moore is the contact person for this activity. Dave will ask Ron Dilabio (SLTT Surficial subgroup co-leader) for any preliminary documents that Andy might have.

  2. Report of Science Language Technical Team (SLTT), provided by Steve:

    The group met on the preceding day (Tuesday). General status of subgroup activities: the Metamorphic subgroup met this week; the Sedimentary subgroup is circulating for comment a draft classification; and the Igneous subgroups (volcanic, plutonic) are continuing to make progress toward a draft standard classification. The leaders of the Surficial subgroup have decided to meet once, to identify one or more strawman classifications; they then will circulate the classification(s) to the subgroup for comment. The subgroup leader's plan to meet in early September was aborted; they now plan to meet in January. Clearly, this group needs to make substantial progress soon, in order to integrate its draft classification with the other subgroups.

    It was suggested that geologic feature names contained in the pending FGDC geologic map symbolization standard should, if possible, be closely related to the SLTT nomenclature. When the map symbol document is prepared for final FGDC review, the coordinator of that effort (Dave) will ask the SLTT if they would like to reflect on the science language contained in the FGDC document. Document preparation is expected in early 2002.

    The NADMSC then considered issues raised by the SLTT Chair (Jon Matti), in a recent email:

    Issue #1 -- should the SLTT draft classification be well-coordinated with activities of the data model design team (DMDT) and tool development team?

    The DMDT already has used SLTT work (i.e., responses to the "20-questions" exercise) to develop a set of use-cases or functional requirements for the data model design. Continued coordination is important, but may not yield significant additional benefit.

    Issue #2 -- what process is envisioned for public review of SLTT documents, and for official, public release of revised, peer-reviewed documents?

    Upon receipt from the subgroups of the draft classifications, SLTT leadership (Jon and others...) will, to the extent possible, integrate them into a coherent classification for rock and sediment lithology. This revised document will be submitted to the Committee for their review. Following revision and Committee approval, the document will be publicly released for comment; mechanisms could include release in the USGS Open-file Report series and, if appropriate, also through the GSC. Advertisements would be placed in appropriate journals, to encourage this public review, and other entities would be notified (e.g., the Geoinformatics consortium). Following a period of review, the document would be revised and published by the USGS (and GSC?), and/or in a journal.

    Issue #3 -- should the documents be authorless (i.e., released by the Committee, with an alphabetic list of contributors including the Team chair?).

    Yes.

    Issue #4 -- in the document head notes, should the sponsoring entities (AASG, GSC, USGS, Provincial GS's, etc.) be recognized?

    Yes.

    Issue #5 -- Should the FGDC be a/the sponsoring entity?

    No, but they play a role in supporting acceptance of the standard. If considered appropriate, after the public review process, the document might be submitted to the FGDC for consideration as a Federal standard.

    Regarding public access to SLTT draft classifications and related activities, before or during the public review period, should the Committee sponsor a poster session at the annual GSA meeting? We decided to consider this in the coming weeks, and decide before the January deadline whether to propose a GSA poster session, to focus on rock classifications and SLTT results. Further, we might offer the draft classification at the next IAMG meeting (September, 2002, Berlin).

  3. Report of Data Model Design Technical Team (DMDT), provided by Boyan:

    The group met on the preceding day (Tuesday). Four items were on their agenda, and these were reported to the Committee as follows:

    Item #1 -- the minutes of the previous meeting (May, in Tuscaloosa).

    These were approved and will be posted to the Web site and distributed to the Committee.

    Item #2 -- the document containing a catalog of the scope of questions that the data model can/should address.

    This "use-case" or "requirements" document was developed from the SLTT's "20-questions" exercise of last year. The document's purpose is twofold: 1) to provide external communication to data model users, so they know what we're doing, and 2) to provide criteria for our further development and evaluation of the data model.
    The DMDT is almost ready to submit this document to the Committee for comment; the Committee will be given a deadline for submission of comments, then the document will be revised and posted to the Web site, along with the SLTT's original list of questions.

    Item #3 -- documenting the data model variants.

    The DMDT is attempting to identify and evaluate the pertinent variants, in order to propose a revised standard data model. The intention is to encourage users to converge (or re-converge) on a standard approach. As a first step, all variants will be summarized according to the DMDT's newly-revised template. These summaries then will be posted to the Committee Web site. At the Committee meeting, minor discussion ensued on how and where these would be posted, in order to make their presence (and our data model revision strategy) most visible to site visitors. Therefore, Peter has agreed to propose how these variants should be posted with supporting text that explains our strategy; his proposal will be circulated to the Committee for comment.
    The variants to be posted include: Arizona GS, Idaho GS, Cordlink, v.4.3, and the Kentucky-National Geologic Map Database prototype. The revised template includes: an abstract less than 2 pages in length, a 1-sheet diagram, a comparison with v.4.3, and an example implementation. The document will be 5 pages + 1 sheet, in PDF format.
    Deadlines -- Web page design and all variant documentation will be completed and posted before Dec. 25, 2001.

    Item #4 -- evaluate the variants and propose a revised standard data model.

    After item #3 is completed, a DMDT subgroup (Boyan, Bruce, and Steve) will evaluate the variants and propose revisions to the standard data model. Their recommendations will be presented for discussion at a 2-day DMDT meeting, just prior to the next Committee meeting, in late May, 2002. Two or three meetings will be needed to develop the recommendations; these meetings may require some travel support from Dave.
    After evaluating the variants, the DMDT subgroup may decide simply to revise v.4.3, in a "bottom-up" approach that focuses on modification of concepts and tables. Alternatively, a "top-down" approach that identifies the core assemblage of concepts needed in a data model will be pursued by Jerry, and may develop into a discussion item at the next DMDT meeting.
    The proposed revised data model will be diagrammed in UML. Upon acceptance of the data model, an E-R representation will be devised.
    How will the new version be designated? This is yet to be determined, but might be something like "NADM_x_".
  4. Miscellaneous issues:

    Dave led a short, introspective discussion on the value of the Committee, and whether it is fulfilling its proper role. It was decided that, all things considered, satisfactory progress has been made. Also, the body provides authority necessary to help facilitate the international cooperation that has allowed progress to occur.

    Regarding the data model Web site, it was decided to sign up all Committee members for all threads of the data model discussions; if this becomes burdensome, each member can opt out of discussions as desired. Presently however, most activity occurs in the Technical Teams, not in the public dicussions.

    Do we need two meetings per year? We will discuss this at the May meeting, when more Committee members are present, and after we hear significant results from both the SLTT and DMDT.

    When is the next meeting? It will be coordinated with the DMT'02 meeting, which will be held Sunday, May 19 through Wednesday, May 22, in Salt Lake City. Plans for the Technical Team and Committee meetings will be determined in early 2002.