Report prepared for the Experts Meeting Towards the Implementation of a Global Invasive Species
Information Network (GISIN), 6-8 April, 2004. Baltimore, Maryland, USA.
I - iv
8/30/2004
GISIN and related efforts develop metadata content standards and tools that satisfy the
intersection of these approaches (at the content level).
Acceptance and use of particular metadata formulations among the scattered and diverse
contributors of invasive species data undoubtedly will depend upon the availability of
straightforward metadata entry tools. Candidates for creating both Dublin Core and FGDC
records are available, and could be specifically configured for invasive species applications
(i.e., with pulldowns for application-specific information such as species names and vectors
and pathways.) It would be very useful to develop de novo tools to create RDF tags to help
tag and search web pages, scanned documents, images, and similar static information
types.
As the network matures, metadata should probably be developed in several “tiers”, with
minimal metadata for small and casual datasets (the frogs of Walden Puddle) nested under
full metadata for important and ongoing data collection programs (e.g., the Non-Indigenous
Aquatic Species system).
Any successful metadata tagging framework for the North American Hub should be
accessible to general-purpose search tools, such as Google, as well as to specific engines
and services maintained by GISIN partner websites themselves.
Content
Previous workshops have identified a series of information types that frequently occur in
invasive species applications, and have recommended that they be adopted as the initial
elements of a distributed invasive species information system. Products in development
include web tools for data entry, databases for information on species, organizations,
projects, experts, and bibliographic resources, shared among more than a dozen countries
and multiple sites within some countries. The approach has been to specify a relatively
small number of core fields or elements expected to be found in all nodes, but the ask that
they be populated with entries chosen from standardized sources. Interoperability should
arise from shared semantics (language). The list of core data types varies slightly from
initiative to initiative (IABIN, NBII, GISP, IUCN-ISSG, North American Aquatic Invasive
Species Program, NISbase, etc).
Some important data types held by GISIN data providers are summarized in Table 1.
In addition, the products include information services, such as interactive mapping,
modeling, automated change detection, and early warning systems, which extract targeted
information to “turn data into information.” Standards bodies and commercial vendors (e.g.,
IBM, Microsoft, Sun) are putting considerable resources into developing technologies to
implement such services on-line as modular “web services”. One standard, SOAP (formerly
Simple Object Access Protocol), is already widely implemented, and should probably
provide the basis for initial implementation of web services within the GISIN network. The
SOAP specification can be looked upon as metadata for a automated computational service
on another machine, and can share structure and vocabularies with more familiar metadata
for datasets and bibliographic resources.
Using such standardized approaches (especially with wide implementation in the
commercial software development community) we expect that our ability to implement
specialized, value-added applications (for example, early warning tools individually
 |
 |
 |