
Hi all, Do we move to fast into the mechanism on how to resolve the GUID before we have defined how the GUID will look like? Sorry if I am to jump too fast on this train in my remarks below. I think that what Arthur and Roger was touching, is that we already have various production mechanisms for resolving metadata object relations implemented. Many organizations already have systems established and populated with metadata and logic to resolve information and relation for the objects that we will assign GUIDs to? I believe this is in part why we are interested to come up with an improved method to address this task? I know we in the European seed bank community have discussed even the option to receive and redirect requests on living biological material (seeds) from crop metadata portals. The portals themselves do not own (or control) either the metadata nor the physical seed samples. There are also services providing additional or corrections to the published metadata on the seed samples. All these activities struggle to achieve a stable link back to the original physical sample and would be significantly improved by a stable GUID mechanism to resolve this link. There are methods and logic implemented to address this problem. Often the link back is lost, and often severe manual work is done to currate the link between these objects. So I think there would be a transition period to migrate the logic to RDFs and to re-engineer the mechanisms to resolve the link between the metadata objects and between the metadata and the physical sample. I am not very well into the theory of the RDF and need to read up before I would be able to start to migrate the logic we have established to resolve relationsships in the seed bank community to RDF. I would however be happy to do the reading if this is the standard GUID representation we will go for. I hope however we will not build the GUID resolver so complex that knowledge about how to use the GUID will be too much of a limitation...? For example if the GUID itself would be something you could paste directly into a browser and get information about the object, would less user threshold than if you would need to install some kind of RDF client to be able to "use" the GUID? We should perhaps start with a very basic GUID solution and remember to think about how this can be integrated with existing mechanisms. Not all existing solutions involve schemas described in XML. And I also believe that migrating even logic described in XML to RDFs would provide us with some challenges? This is why the call for use cases by Donald is important? Cheers Dag Terje Quoting Roderic Page <r.page@BIO.GLA.AC.UK>:
------------------------------------------------------------------------
participants (1)
-
Dag Terje Endresen