As outlined at the Beechworth WDA symposium John Trueman and I believe it should be possible to 'put the Odonata on the web' ... and that we could make this a first (one up for the dragonflies!). We are also of the opinion that within the Odonata tree of life the terminal leaves are a lot more stable than the 'branches', so a 'bottom-up' structure is likely to deliver faster than a 'top-down' analysis.
To this end we have started developing a database to include all the appropriate terminal information on every dragonfly species. The structure is developed in XML to ensure portability, scalability and extensibility. We have arrived at a first offering and are putting this out for comment. The present XML DTD is pretty cruddy (but it is my first attempt at one of these things). The current DTD could be re-expressed later as a more sophisticated DTD or in a Schema without making any difference to content (XML is wonderful in this way).
For the moment I have entered the NZ fauna 'by hand' (which is painful) so I can play with the usefulness and look at queries and how they can be filtered. While the fauna is small and largely endemic things like Pantala flavescens formed a good test of naive concepts on coding distributions!
At this point the concept has the database being searched by scripts written in the Perl language or XSLT, so it will work.
The present format allows the following information to be stored and retrieved:
<species by current name>
<synonyms>
<warnings on taxonomy>
<keys>
<descriptions>
<original_descriptions>
<habitat>
<distribution>
<conservation>
<links to other resources>
The sorts of queries that would be possible include 'which keys identify species X?', 'what keys identify Corduliidae from Africa?', 'where does species X occur?', 'what are the larval habitats of genus X?', 'where is a picture of species Y?' ... and then you can get imaginative.
'Authorities' link to a separate database with the full citation and a link
to an electronic facsimile of the paper where this exists. The page number in
the 'authority' field would allow direct linking to the point of interest in
the facsimile paper when these points are coded in. Searches for Family-group
or higher structures would be based on lists of included genera; essentially
the system uses recursion through lists of higher categories until we reach
species …
For Libellulidae might do as Tribes or others. These lists would need to be
hand-crafted … but creation would be faster than 'coding' and the method
is also more flexible as any change immediately propagates. Filtering is also
trivial by requiring values in any specified fields.
A back of the envelope calculation shows a file of about 600 000 lines and 18Mb ... which is big and will require many contributors. We now need a consensus on what should be included before we develop automatic validity checking data entry tools (rather than entering data into a template that looks like this!) and start seriously assembling data.
An example of the data for a single species is shown here.
Please
email
me with comments and suggestions for material to be included. Once the
design freezes new data fields can be added (XML is great!) but only as
options.
Design freeze is tentatively set for 1 March 2004.