Publications and the usability of scientific data

Peter N. Schweitzer

1. Who am I
    Office of the Chief Geologist, Information Resources Group
    Coordinate scientific data management divisionwide (4 years)
    Previous work:
        Data management in GD Global Change Program (5 years)
        Software developer, Microway Inc. (2 years)
        Science background, Ph.D. Oceanography (micropaleontology)

2. Purpose of talk:

    Describe and interpret changes I have seen in the interaction
    of publication processes and scientific data management during
    the past 10 years, emphasizing USGS.

    Main points:

    (a) Rising technological capability of both scientists and
        the public has increased the usefulness of digital
        scientific data and diverted the interest of USGS research
        staff from the printed publication towards the release of
        operable digital data products.

        Operable digital data products means data that people can
        process using software that understands the scientific
        content of the information, not merely graphics.

        Print is not going away, but print is no longer the only
        game in town.

    (b) The use of digital data products by the public requires
        changes in the way we do our work; these changes are
        partly organizational, partly philosophical, and are
        nowhere completely effective.

        Positions in science teams explicitly designated as data
        managers have had their roles expanded to subsume more
            Documentation directed towards outside users
            Design and deployment of data access systems
            Direct user support

        Positions in the publications groups have been explicitly
        redirected towards digital processes, and also include
        some of these same roles.

        Emerging needs now widely recognized as crucial tend to
        appear between the boundaries of existing organizational
        purviews:
            Reengineering of data catalog systems
            Enhancement of data documentation
                increasing consistency
                widening audience
            Support of data users
        These functions stress existing personnel in data management
        and publication roles.

    (c) Institutional resistance to these changes stems in part
        from the diversity of functional roles that USGS plays,
        each of which is supported by a rather different, but
        valid philosophy of interaction with the public.

        Historically, USGS has had 3 different ways of dealing
        with the users of our information:

            (1) Basic researcher
                Anticipate questions nobody has asked yet
                Limited contact with users, mostly core professionals
                Credibility established through publication

            (2) Consultancy
                Address specific questions from clients
                Intimate interaction with client organizations
                Data not generally available to the public
                Credibility established through clientele

            (3) Data warehouse
                Questions answered are longstanding and well known
                Data transfer (order processing) support
                Redirects customers to consultants for help
                Credibility established by quality of data,
                which is governed by procedures

        These represent a continuum along which the types and
        roles of the data and information users varies.  The
        increasing usefulness of digital data by the public
        and non-core professional strains these roles because
        operable digital data have different properties than
        printed publications or client-specific services.

            Data products permit (and require) rapid revision

            Data improve as people use them if you're paying
            attention to the users' experiences and feedback

            Changes in users' tools tend to change real or
            perceived needs in data structure and format

        Concepts and methods from the software industry can
        usefully inform our handling of digital data products

            As with software, review of data cannot detect all
            errors and usability problems

            Early release of products with promise of technical
            support and working feeback mechanisms
                Flexible, multi-tier technical support
                Teachers and learners vs wise hermits
                    Organizations that rely on people who know what
                    nobody else knows tend to be brittle, weak, and
                    vulnerable to staff turnover.  A more robust
                    philosophy is for people to strive to know what
                    everybody else knows.  The result is a community
                    in which everyone is a teacher and everyone is
                    a learner as well.  This effect is most obvious
                    in the field of desktop technical support.
                Reputation for responsiveness counts, but takes time

            Explicit versioning of products, increasing the
            awareness of changes among end-users

            End-users become partners in review process and
            consequently subsequent development efforts

            (4) Software developer
                Questions arise in feedback from users
                Work with users to understand their needs
                Modifies data, develops tools for users
                Credibility established by usability of data
                and quality of technical support

(end)