Palmer Station LTER Information Manager Site Byte - September 2008 Karen Baker

In addition to advancements in design and population of our cross-site
information management system DataZoo, there were three major accomplishments
this year: 1) the writing and acceptance of the site's fourth 6-year proposal in
which information management is recognized as one of ten components where both
science and information management summarized their objectives in an itemized
list, 2) the publication of an information management paper in the oceanography
journal Deep-Sea Research (Baker and Chandler, 2008, 'Enabling long-term
oceanographic research: Changing data practices, information management
strategies and informatics'), and 3) initiation of a division-level computer
infrastructure recharge facility as an organizationally situated support for
systems administration.

DataZoo has reached 'flagship' status as evidenced by a change in inquiry from
'Where is dataset X' to 'Why isn't dataset X in DataZoo?'. Local design efforts
focused in particular on improving the user interface to DataZoo, creating an
online help system, and shifting code practices to use of libraries and an
object oriented approach. Design and flow of the system as a coherent whole was
revisited in order to clarify the design as well as to add consistency to users'
interactions with the data and metadata models. The management interface was
improved in terms of useability. We continue to address design decisions such as
whether code sets are associated at the attribute or the column level (decision:
move to column level), how to deal with metadata that changes over the course of
a dataset (decision: handle at section level), and what keyword classification
approaches to take.

The PeopleZoo personnel application module has been redesigned as an API and is
being used to track project and cruise participation. Three additional
information system elements stabilized: 1) definition of relationships between
units, attributes, and qualifiers as an interdependent set of dictionaries, 2)
development of web services in a locally developed hybrid approach using both
SOAP and REST, and 3) launch of a multi-element cooperative dataspace composed
of a set of related applications that capture and make accessible multi-faceted
datasets.

Support for the outreach component included development of an education web
portal and investigation of communication options resulting in redevelopment of
the picture-of-the-day activity as a blog to support the new teacher-at-sea
activity. LTER network collaboration included returning to design of a
community-developed, network-hosted, service-oriented unit dictionary database.
The Ocean Informatics team worked closely with the LTER Information Management
Committee Dictionary Task Force to define and coordinate continuing development
of the unit dictionary. In addition, we worked to capture site datasets
submitted to EcoTrends for submission to DataZoo. After a third unsuccessful
proposal to obtain NSF funding for the Ocean Informatics conceptual approach to
local infrastructure building and to distributed networking dictionary efforts
as alternative models for development, we obtained auxiliary support for our
team approach from NOAA for a new database project and from Cal Fish and Game
for work on fisheries program data management that supported development of web
services and a new generation targeted application.

Collaborative science studies research continued with a best paper award for a
Digital Curation Conference contribution that will be published in the
International Journal of Digital Curation (Karasti and Baker, 'Digital Data
Practices and Long Term Ecological Research Program Growing Global'). In
addition, the concept of  'Community Design' as an approach to infrastructure
building was investigated and presented at the Participatory Design Conference
(Karasti and Baker, 2008).  As part of our focus on articulation we continued to
write for Databits: 3 articles in Fall 2007  ( 'Tools: Web-based data
visualization with JPGraph, 'YUI: An Open-source JavaScript Library',
'Professional Learning Opportunities: Conferences, Meetings, and Mindsets'; 4
articles in Spring 2008 ('Big Science and Local Meetings', 'Developing and Using
APIs in System Design', 'Preservation Metadata: Another Chapter in the Metadata
Story' and 'Data Quality: Yet Another Chapter in the Metadata Story'). Four
Ocean Informatics participants attended the 2008 IMC meeting using support from
this year's supplement grant. Three posters were presented at the EIMC
Environmental Informatics Conference: 'Abstracting Functionality and Access:
Facilitating Data System Manageability and Site Coordination'; 'LTER IMC
Community of Practice: A Learning Environment'; and 'Local Information
Management and Information Infrastructure: Roles, Responsibilities and
Practices'.

1. Major Cyber-Infrastructure (CI) Challenge Our major need is for increased
staff support in order to be able to make long-term plans as well as to
establish and maintain a minimum local infrastructure (MLI) capacity that would
allow us to participate actively in development of a wider variety of tasks in a
more timely manner, e.g. tasks related to networking, enactment of existing
standards, community prototyping, standards-making, web services development,
design work for new data types, and pre-federation activities.

1a) Science Issues Driving the Need for a Minimum Local Infrastructure (MLI)
Capacity Traditionally within the LTER, local science is supported by local
information management. In order to expand this arrangement to include support
for synthetic science as well as cross-project, cross-institution, and
cross-network informatics, a new way of doing information management is
required. We have grown from preserving well-defined local datasets for
immediate use locally to needing to plan for new sets of expectations, new types
of organizational arrangements, and new kinds of learning environments. The
multiple levels of work involved in developing long-term data stewardship and
networked data repositories involving diverse data types have only begun to be
recognized and incorporated into the information landscape. Maintaining
usefulness of local endeavors means being prepared to address data
interoperability. Data interoperability, a requirement for a growing number of
contemporary large-scale scientific undertakings, will require development of
capabilities that enable LTER site participants to contribute to community
standards-making, local infrastructure-building, and sustainable innovation.

Without support for MLI, a divide grows between top-down and bottom-up
approaches to information management. Cyberinfrastructure efforts often focus on
high end computing, massive storage, and grid capabilities while local
infrastructure efforts focus on data organization, description, analysis, and
capture at the source. Local information systems typically evolve through
development of modular functionality, informal prototyping and dataset-driven
design. Local information management tasks frequently involve rapid responses to
unplanned opportunities that occur during data capture, analysis, or
preservation. The work involves both data-handling and articulation work; both
are critical to contemporary cyberinfrastructure-building. Development of local
infrastructure will insure that sites are prepared to manage, innovate, and
engage locally as well as to design and participate in community collaborative
environments. Work involves -implementation of improved information systems able
to address contemporary data and metadata exchange requirements in support of a
web-of-repositories vision - data organization that includes real-time and new
types of datasets as well as their integration into our information system as
well as for basic system administration - data integration that would be
supported by implementation of the proposed living dictionary in partnership
with the LTER IMC - availability of high quality data through QA/QC services
addressed in an extensible manner at the information system level in partnership
with the LTER IMC -mechanisms for working with data through development of local
toolkits for matrix manipulation as a resource for local contributors and users
of the information management system -new types of data access that involve
design and development of web services as well as support for a half dozen
enactment scenarios involving diverse users, datatypes, and interface needs -new
types of data discovery and query through implementation of a geographic module
and framework datasets compliant with existing standards and thus able to
interface to existing GIS or KML applications -taking leadership role in
articulating, learning and teaching of informatics so local researchers and
students benefit from and engage with the local information environment

1b) Impact of Insufficient Minimum Local Infrastructure (MLI) Capacity on Site
Activities Without MLI support, network design is conceptualized and funded
top-down.

Lack of staff precludes a site from addressing the topics listed above (1a) in a
timely or a practice-based rather than theory-based manner. Due to the increased
volume of data and maintenance of existing systems, site participants are
increasingly limited in their ability to contribute time for communication,
design innovation, or practice-based experience that would enrich theoretical
understandings of data organization and information management. There is also no
time or support to participate in opportunities to articulate and/or teach what
has been learned in practice. Site information management work would benefit
from a more formal approach that would lead to further advancements in local
information management and innovative contributions to cross-site coordination.

1c) Impact of Lack of Minimum Local Infrastructure (MLI) Capacity on Network
Activities Without site-based MLI support, a top-down approach to information
infrastructure-building influences design of contemporary networks in terms of
centralization versus federation.

We have had to become highly selective in identifying a few tasks that synergize
with site tasks to work on at a community level. We have focused on the unit
dictionary since 2004. Several years ago design was initiated collaboratively;
this effort was developed, prototyped, reported verbally at a meeting and in
writing via Databits. The effort was discussed in conjunction with other IMC
work on controlled vocabularies and ontologies. Having completed the first
prototype, it proceeds at an unfunded pace rather than at the speed of a
critical NIS component. Progress was made at the 2008 IMC meeting on the
dictionary although no mechanisms to support this community work are currently
available. Discussions at the annual IMC meeting resulted in agreement to try a
new site-network design approach. The plan is to carry out unit dictionary
development in a site development arena and deploy it into an LNO production
area using SVN as a coordinating mechanism.

2) Site CI development projects pertinent to a general Service Oriented
Architecture (SOA) framework

The unit dictionary described above is being developed in a SOA framework. We
will investigate development using the site-network design approach mentioned in
1c.  With appropriate site support several Ocean Informatics (PAL & CCE) modules
could be used community wide, e.g. unit, attribute, and qualifier sets in
addition to personnel, bibliographic, and media gallery modules. These modules,
designed with cross-project sharing in mind, have been developed as APIs with
management interfaces.