Interaction Extension Concept List

This document lists the proposed concepts of a new Interaction Extension for version 1.4 of DarwinCore. This extension is being proposed as a means of recording the interaction between two organisms: one represented through DarwinCore? and the other represented through this extension. This list of concepts basically consists of a concept indicating the type of interaction and then the same Taxonomic, Identification and Biological concepts that are present in DarwinCore?. Although the names for these concepts are the same as in DarwinCore?, they will be under a different namespace. This extension can be further extended by other lists of concepts that are specific to certain types of interactions (for example the PollinationExtension).

Status:

• First version (0.1) of the XML schema finished. Ready for testing.

Submitted by: Renato De Giovanni, Etienne Cartolano, Tereza Giannini, Antonio Saraiva and Pedro Pizzigatti.

Submitted: 16 May 2007

Development of this extension was supported in part by the Pollinators Thematic Network of the Inter-American Biodiversity Information Network (IABIN) and by the Pollination Information Management System of the Food and Agriculture Organization of the United Nations (FAO).

Open Questions:

• 16 May 2007 - Should we allow opposite relationship types? For instance VisitedBy? and Visited (R. De Giovanni).

• 8 June 2007 - How should we handle interaction attributes to avoid a terminology explosion like: feeds on flowers of, feeds on nectar of, feeds on pollen of, feeds on seeds of, feeds on fruits of, feeds on anthers of, etc, etc. (R. De Giovanni).

• 14 June 2007 - InstitutionCode and CollectionCode are null if both species are in the same institution or collection (E. Cartolano)?

Comments & suggestions:

• 16 May 2007 - This extension might include a RelatedGlobalUniqueIdentifier? (J. Wieczorek).

comment: thanks, it was included. (R. De Giovanni)

The American Heritage(r) Dictionary of the English Language: Fourth Edition. 2000. (http://www.bartleby.com/61/45/P0514500.html)

...Inflected forms: predated, predating, predates

1. To mark or designate with a date earlier than the actual one: predated the check.

2. To precede in time; antedate....

I'd prefer to see "PreyedUponBy" (D. Inouye)

comment: thanks, I made the change. (R. De Giovanni)

I would support the inclusion of opposite relationships, because/if the direction of the relationship will depend on the placement of the data regarding two the organisms either in the DarwinCore? or this extension.

I am aware of several examples where a related organism has a different catalog number e.g., plant vouchers deposited at one institute and the insect vouchers at another. This also may apply to vouchers in different collections at the same institute.

-- AaronWilton - 13 Jun 2007

comment: thanks, we added the possibility of having opposite relationships and we also included concepts to fully locate the record of the related organism: institution code, collection code and catalog number. (R. De Giovanni, T. Giannini and E. Cartolano)

You should be aware of the TaxonOccurrenceInteraction? object being developed as part of the LSID Vocabularies for use at Herb IMI to indicate interactions between fungi and hosts.

http://rs.tdwg.org/ontology/voc/TaxonOccurrenceInteraction

The TaxonOccurrence? object is a very close match to DarwinCore? 1.4 and the TaxonOccurrenceInteraction? class allows multiple arbitrary links to be made between these objects using an extensible mechanism.

I strongly recommend not continuing to extended DwC? in this manner. Especially if it involves linking objects together like this. It will only store up problems for the future.

-- RogerHyam - 2007-09-06

comments:

Roger, I may agree with you that an RDF approach can better formalize links between objects, and I also like the TAG effort to produce a standard ontology that we can all use. I just disagree that extending DarwinCore? "will only store up problems for the future". You know that TAPIR (the protocol intended to exchange this data) is very flexible in terms of output, so, if necessary, it should be possible to get a valid RDF output (according to the new TDWG ontology) from providers that mapped DarwinCore? + extensions. It should also be easy in the very near future (I'm working on this right now) to migrate local TAPIR mappings to equivalent ones (eg. from DarwinCore? + extensions concepts to TDWG ontology concepts). This means that what was proposed here, if adopted, will in no way prevent interoperability and compatibility with initiatives like the TDWG ontology. DarwinCore? is being used now to serve more than 100 million records to GBIF, and extending it was the simplest and most effective way to move forward in the projects that were interested in exchanging interaction data.

-- RenatoDeGiovanni - 2007-09-07