[M] Dealing with a mixture of discovery formats in the Discovery service

[selatham]

It seems unlikely we can avoid having to harvest a number of different discovery formats. DIF, ISO19115/139 at least. Need to decide how we are going to handle them in Discovery portal. Will we transform everything into one standard? Would this be too lossy? If we don't transform we will have to maintain a number of XQueries which will have to be in parallel then the results merged. (Unless there's a better way?) Decision required.

[lawrence]

Input Issues

• need to be able to find all the spatio-temporal data to input into postgres

Backend Issues

I think the number of xqueries required isn't too bad. Ticket:515 suggests we have

• doFullTextSearch - which needs to point at different collections, but that's all.
• doParameterSearch() - will need to know where the parameters are in each backend format.
• doAuthorSearch() - will need to know where authors are in each backend format.
• doSpatioTemporalSearch() - no problem, it's pushed into the input.
• getResultSummary() - no problem.
• doPresent() - no problem (simply returns an xml document)

Front End Issues

Essentially we have to decide what it is we want to show from each format. Currently I'm supporting abstract, title, authors, parameters, xml-full text ... and geospatial ... so that means I have to know how to get the same things that Matt needs to know, so we only really need the xqueries once ...

Supported Formats

• ISO19139 (but we may have multiple flavours!)
• DIF
• MDIP

[selatham]

The need to distinguish by originating Data Centre(DP) will probably run through all these queries. Linked to task #430 " How can we limit discovery searches to specific sites?" We won't be too popular if we leave it up to implementors to do the filter themselves.

[hsnaith]

Milestone ISO19115_PreBeta deleted

[selatham]

DECISION: ProjectBoard? Oct06 We will support core ISO19115(encoded in ISO19139) as well as DIF and something for EDMED/MDIP. Core means encoded as vanilla ISO19139 (so if people create different names in their own 'restricted/extended schema' then we won't handle that).