The Simple User Interface
Demonstrator for ABCD providers is an experimental interface providing
specimen level access to distributed biological collection
information. Its main function is to provide a "proof of
concept" for some functionalities that at a later stage will be available implemented in a much more thorough way by
the workpackages involved (e.g. the BioCASE
User Interface development). The software used is
being developed in the BioCASE
project and will also be employed by the Botanical
Node of the German GBIF.
BioCASE Protocol - requests
and responses as XML documents
Based on the experiences of the ENHSIN
pilot implementation, queries and query results are propagated as
XML
documents, following the BioCASE
protocol. On the portal side, the UnitLoader
software provides the functionality necessary to
generate a valid request document, to send a parallel
request to all registered databases, and to receive and handle
the results returned. On the providers' (data
holders') side, a piece of software installed (the "Wrapper")
is responsible for translating this request into a format that
the individual database can "understand", and to
wrap the individual results back into a common data structure: a
representation of biological collection data following the ABCD
standard.
Query Extension
To demonstrate the possible use of thesauri or similar
tools in extending a user request to related (e.g. synonymous)
concepts, a query extension is included for taxon names. When
a user enters a taxon name into the search form, this input is
used to consult a taxonomic, concept oriented database based
on the Berlin
Taxonomic Information Model. If related concepts exist,
the query sent to the providers not only contains the term
entered by the user, but also the related names derived from
the database. Presently, this query extension is limited to
the concepts of included taxa and synonyms, using a database
of higher plants. The mechanism is
available for names of rank genus or below. To apply it to a
genus name, the search term should be entered trailing a blank
(white space) before the wildcard (e.g. "Pyrus
*") to activate the query extension.
Request Limitation
With a rising number of databases linked to the unit level
network, as well as with a rising number of search requests,
it will become increasingly important to limit the number of
providers that receive a specific query to only those who are
at least remotely probable to have relevant data. In the
demonstrator implementation, the central core metadatabase (CORM),
apart from storing the data necessary for the search interface to
connect to the registered providers, maintains an index to a limited set of query concepts. The
demonstrator implementation consults this index for queries
on taxon names only, and sends the request out only to those
providers that are supposed to have data for that name. If
query extension has been used (see above), the related
concepts are tested against the index as well.
XSLT
For the demonstrator implementation, the XML results are
transformed into html by a server-side XSL Transformation (XSLT).
This provides some roughly formatted output for the XML data
that can be viewed with all browsers, including those that are not yet
XML-enabled.
Point Mapping of georeferenced data
For datasets that contain point location data for the
collection site in decimal format (lat/lon), a live request
can be sent to the Generic
Point Mapper service of the Canadian
Biodiversity Information Facility by clicking on a button
labeled "Show Map!" that shows up underneath the result
set (try, e.g., a query on taxon name "Oliarus*").
The service creates a dynamic map from the data. If data
providers associated web addresses with the individual
collection objects, clicking on the dots could open a web page
displaying further information.
|