[BiodiversityCollectionsIndex:70] Using BCI
Peter Midford
As many of you know, the Phenoscape project has been building a knowledge base of annotations of character data from taxonomic publications focussing on Ostariophysian fishes. Part of our process for curating publications includes entering specimen data using Phenex, a GUI-based tool developed as part of the project. Up to now, we have maintained our own vocabulary of collections and abbreviation codes, but we are making a push to use BCI's URNs and drop our vocabulary.
Looking at the API page it seems that the preferred way to make the BCI names available to our tool is to grab the nightly .csv updates and process them locally rather than use the web API for, as a example, name completion with the accompanying hits on BCI's server. Is this correct?
We have also started adding of our collection information into BCI, and that has raised a question as well. Specifically, how does BCI handle synonymy in collection names, either due to collections being merged or due to institutions changing names. Ideally there is a place to record earlier names and codes to avoid generating new entries, but maybe another approach is more appropriate.
Thanks,
Peter
Peter E. Midford
Phenoscape Ontology Curator
Peter....@gmail.com
--
You received this message because you are subscribed to the Google Groups "Biodiversity Collections Index" group.
To post to this group, send email to biodiversityco...@googlegroups.com.
To unsubscribe from this group, send email to biodiversitycollecti...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/biodiversitycollectionsindex?hl=en.
rogerhyam
Hi Peter,
Sorry for the tardiness of this reply.
> Looking at the API page it seems that the preferred way to make the BCI names available to our tool is to grab the nightly .csv updates and process them locally rather than use the web API for, as a example, name completion with the accompanying hits on BCI's server. Is this correct?
Yes and no. The data in BCI changes relatively slowly and, in the
great scheme of things, there is relatively little of it. So the
easiest way to work is just to grab a snap shot occasionally but I
would probably not plan to do this every night. If you need data this
up to date you should probably download a snapshot and then keep it up
to date with the OAI-PMH service.
> We have also started adding of our collection information into BCI, and that has raised a question as well. Specifically, how does BCI handle synonymy in collection names, either due to collections being merged or due to institutions changing names. Ideally there is a place to record earlier names and codes to avoid generating new entries, but maybe another approach is more appropriate.
This is a good question both technically and philosophically. Once a
record is created it can't be deleted however just about all the data
about it can be changed- everything but the ID. Conceptually this is
like that friend in your address book who may move house, get married
and even have a sex change so that every single piece of data about
them has changed but they still have the same entry. It is therefore
important to first try and avoid creating duplicates and the user
interface attempts to do this. If duplicates are discovered then the
only mechanism currently available to deal with them is to make one of
the collections a parent of the other collection. A note is then added
to the child collection saying this is a duplicate id. Conceptually
this is an OK way of doing it as there will be a subset of resources
out in the world that use the id of the sunk collection and therefore
the sunk collection is in some sense a subset.
BCI has been in a holding pattern for nearly 2 years now but we have a
bunch of resources becoming available at the end of this year to
continue developments in 2011. I have just instigated a discussion
within the partners about what this might entail hope to open it out
for wider discussion soon.
If there is anything you would like BCI to do that it doesn't do
please shout about it on this list as it will effect how we develop
things going forward.
All the best,
Roger
Peter Desmet
There are plans to collect metadata about collections in Canada, and I think BCI would be a great platform for doing this. It hasn't been decided yet what information will be captured exactly, but it is along the lines of # specimens per family/genus, # specimens per region, # specimens digitized / georeferenced, etc. As this is rather structured information, I would prefer if they use a standard for this (EML?) and it would be great if the BCI database could: 1) Allow the input of these data 2) Express those data in the chosen standard.
It's all very tentative at the moment, but I hope it gives an idea of what direction BCI development could take.
Peter Desmet
--
You received this message because you are subscribed to the Google Groups "Biodiversity Collections Index" group.
To post to this group, send email to biodiversityco...@googlegroups.com.
To unsubscribe from this group, send email to biodiversitycollecti...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/biodiversitycollectionsindex?hl=en.
--
Peter Desmet
Biodiversity Informatics Manager / Directeur de l'informatique de la biodiversité
Montreal Biodiversity Centre / Centre sur la biodiversité
Canadensys - www.canadensys.net
Institut de recherche en biologie végétale
Université de Montréal
4101 rue Sherbrooke est
Montréal, Québec, H1X2B2
Canada
Télé / Phone: 514-872-5249
Télécopie / Fax: 514-872-9406
Email: peter....@umontreal.ca / peter.de...@gmail.com
Public profile: http://www.linkedin.com/in/peterdesmet