Palmer Station LTER Information Manager Site Byte - September 2009

Karen Baker

1. Ocean Informatics Environment

The development of Palmer Information Management continued within the broader
context of a growing understanding of diverse data types and data repositories.
The local Ocean Informatics environment expanded from a one-component
information system to a three-component architecture involving not only highly
structured data but also highly complex data organized in DataSpaces and very
large, homogeneous file collections. A focus on design and development of web
interfaces enhanced community and public data access. Application management
interfaces have been implemented to date for personnel, bibliography, metadata,
data upload, and file export to Metacat. Population of Datazoo with datasets is
ongoing; data originally submitted to Ecotrends via principle investigators is
being incorporated into Datazoo with emphasis on providing a unique identifier
for dataset. In terms of physical infrastructure, upgrades to operating system
and a suite of related software (e.g. mySQL, perl, php, apache, python) occurred
in October 2008.

2. Activities, Tools and Design Practices

Tools to support design practices are important for long-term information
environments.

With an increasing number of design efforts, patterns in design are beginning to
emerge and influence our approach. At the code-level, as distinct from
project-level design or system administration design, a single design pattern
emerged in working on applications such as the unit registry, controlled
vocabulary, plotting, and cruise registry. Discussions have moved from a focus
on the architectural level of the schema to code reviews involving classes and
methods.

In relational database work. visual modeling software became more important for
schema development and collaborative efforts. We have moved from use of
dbDesigner and cocoMysSQL which were hard to extend, improve or share to using
mySQL workbench with its improved in terms of having more generality, an active
product lifecycle, and running on multiple platform types.  Development has
included code structure redesigned to be library-based and object-oriented, APIs
created for internal consumption such as dbConnection, peopleZoo and Media
Gallery, and XML database use was initiated with input forms and data storage as
well as for indexing file systems in the CTD file finder application. Web
services, functioning as an online API or a developer tool that allows a script
or program to get info, are largely consumed locally. The Unit Registry is the
first web service to be used publically. Due to limitations in plots types,
response speed, and mapping elements, visualization software use moved from
jpgraph (apache based) to matplotlib (python based).

3. Partnering and leadership

Partnership activities included support for Palmer LTER outreach component
expanded efforts with field blogs.  Work with the zooplankton community involved
co-convening a NOAA sponsored zooplankton workshop that provided new insights
into organization and management of species-specific data. At an LTER network
level, we have contributed through leadership and active participation on a
number of Information Management Committee working groups (e.g. Unit Registry,
Unit Best Practices, Governance) and as co-editor of the Databits Newsletter.
Three REU students were mentored over the summer. Four information management
posters were presented at the LTER All Scientist Meeting: Growing Information
Infrastructure: Data Lifecycles and Subcycles (Baker, Millerand, Yarmey);  LTER
Unit Registry (Kortz, Conners, Yarmey, Todd, Baker), A Web-of- Data Repositories
(Yarmey and Baker), LTER IM History DataBase - HistoryDB (Petersen, Baker,
Kaplan, Melendez), and Site-Based Information Architecture (Conners, Kortz,
Yarmey, Baker).

4. Learning & Capacity-Building

With training and learning central to our information management strategy, an
informatics reading group was held weekly in the summer. Additional learning
opportunities were created by having informatics team members serve as
raconteurs for a zooplankton workshop and four participants attending the LTER
All Scientist Meeting.. Special informatics events were created around the
visits of Deb Peters, Nicole Kaplan, and Inigo San Gil as well as international
visitors Sonia Palfner, Stephane Couture, Helena Karasti and Florence Millerand.
Collaboration within the team and with Science Studies participants involved
articulation work and publications on information environments (Karasti and
Baker, 2008), ocean informatics (Baker and Chandler, 2008), information systems
(Millerand and Baker, 2009), and repository federation (Baker and Yarmey, 2009).

5) Looking forward

Planning has begun to replace old hardware, to improve the Datazoo metadata
model, to develop an interface for the very large data collections such as CTD
data, to design a dataset access log interface, and to create applications to
track dataset submission and changes. The trajectory of information management
efforts and sustainability of the team approach in terms of institutional
placement and LTER network interest are open questions discussed in a chapter on
growth of local information infrastructure (Baker and Millerand, submitted).