[tdwg-content] Treatise on Occurrence, tokens, and basisOfRecord

Richard Pyle deepreef at bishopmuseum.org
Mon Oct 25 00:43:07 CEST 2010

> What is a bit frustrating to me is that ideas like these 
> aren't laid out in an easy-to-understand fashion and 
> placed in easy-to-find places.  I have spent much of 
> that last year and a half trying to understand how 
> the whole TDWG/DwC universe is supposed to fit together.  

Understood, and agreed.  Part of the problem is that a lot of this stuff is
driven by passionate individuals, who also happen to be highly
over-committed.  There's barely enough time available to do the interesting
bits (conceptualizing, experimenting with implementations), let alone the
less-interesting bits (documentation).  Having said that, there are some
early documents that go into a lot of this in great detail.  One is Stan
Blum's description of the ASC model.  Another are a series of publications
from Walter Berendsohn on "potential taxa".  A lot of other stuff is
floating around the Specify project, and there are some other earlier
sources.  But I agree, it's not easy to find, and it doesn't always cover
the details we need it to in today's context.

> The point that I was trying to get at (eventually) was that it 
> was inconsistent to say that images need to be referenced as 
> associatedMedia and sequences needed to be referenced as 
> associatedSequences, and yet not say that specimens needed 
> to be referenced as "associatedSpecimens".  

Hmmmm...not sure I agree.  If it is so that Occurrence=Individual+Event,
then a Specimen can be said to *be* the Individual, whereas images, DNA
sequences, and the like are the tokens.  In other words, Individual "is a"
Specimen; but Individual "has a" image.  Now that I think about it, perhaps
Specimens should not be treated on an equal par with other tokens; and
indeed, maybe specimens aren't tokens (per your definition) at all.  This is
not consistent with how I've always thought about it (see my previous
email), but if the elusive "Individual" is key to this relationship, then
perhaps Specimens serve as bot "evidence" of an occurrence, and the "stuff"
of the Individual represneted by the Occurrence.

My brain hurts.

> I guess I'm thinking about this in terms of a token being 
> something to which we can assign an identifier and retrieve 
> a representation (a la representational state transfer).  
> Although I don't deny the existence of memory patterns in 
> neurons that are associated with a HumanObservation, 
> there isn't any way that we can receive a representation 
> of that memory directly.  

I guess it depends on what you mean by "representation".  We can't retrieve
a specimen directly either -- but we can retrieve a database record that
represents the specimen, and metadata associated with it.  I think the same
can be said about a human mmory (as the foundation of an observation).  That
is, there is a species identification, number of individuals, etc.,
associated with an observation that is based on the memory of the person who
made the observation, and that memory is represented by a database record
with associated metadata.

This conversation could go very weird, very quickly -- and maybe I'm just
being difficult (in which case I apologize).  But now that I see that a
specimen may, in fact, be fundamentally different from other kinds of
evidence supporting an occurrence, I'm not longer sure what I believe
anymore (especially after the 11-hr flight from Berlin I just got off of).

> > Maybe the answer to this is to treat different versions of DwC as
> > concurrent, rather than serial. 

> Yes, I agree about this concept.  I think that what I'm really 
> advocating for is that we agree on what the most normalized 
> model is that will connect all of the existing Darwin Core 
> classes and terms.  In that sense, when I'm asking for 
> Individual to be accepted as a class, I'm not arguing for 
> a "new" thing, I'm arguing for a clarification of what 
> we mean when we use the existing term dwc:individualID.  

Makes sense to me.


More information about the tdwg-content mailing list