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 - vii
8/30/2004
Network Architecture
Since all members of GISIN are parts of larger networks, the architecture connecting Hub
sites will necessarily be heterogeneous, and will need to follow the developments of larger
initiatives such as the National Biological Information Infrastructure in the U.S., the North
American Commission on Environmental Cooperation (e.g. NABIN) in the trilateral region,
and global efforts such as the Global Biological Information Facility (GBIF). Nevertheless,
networking GISIN sites still requires some collaboration on network components.
Most important is a centralized access point, probably representing both a centralized index
database (UDDI registry) addressing a number of distributed metadata, content databases
and web services. The index may point to a number of data aggregators (e.g., regional data
nodes, such as country nodes under IABIN), but should be capable of providing discovery
services for lowest-common-denominator resources, such as metatags in HTML documents.
With current technology, a fully distributed system is likely to suffer severe performance
limitations, at least in part simply because servers down or bandwidth limitation can prevent
successful searches over the whole network at any one time. Addressing these performance
issues probably requires some level of caching and off-site replication, although
authoritative versions of most resources will still be held at the data provider’s site.
Technical Challenges
Languages
One of the most immediate challenges is how to construct GISIN to integrate data and
services for users speaking multiple official languages (much less a multitude of languages
used by immigrants and indigenous peoples).
Suggested approaches include:
• Examining the applicability of off-the-shelf multilingual tools, such as Systran’s software
• Adopting established multilingual controlled vocabularies, such as GEMET or those
developed by IUCN or UNEP
• Developing look-up tables and schemas for core multilingual content issues, such as
colloquial species names
There are important trade-offs in using established tools and vocabularies for multilingual
applications. The existing software and files can be unwieldy, and may not correspond well
to usages within any given language community. Whatever their virtues in automating
translation services and multilingual queries, they will not work unless they provide ease of
use in individual language settings.
Intellectual property
Any federation of databases that reference published literature must address the intellectual
property rights of information providers. Even where information is completely in the public
domain, it is important to identify and provide professional credit to the developers of each
information resource.
The issue of intellectual property rights is vast, and was not systematically addressed by the
workshop. However, immediate challenges to establishment of GISIN include how to access
or license some important material copyrighted by third parties (such as species descriptions
 |
 |
 |