On Wed, Jul 6, 2011 at 10:42 AM, Paul J. Morris <mole@morris.net> wrote:

I'd concur with John except for that pesky little word in the
definition of dcterms:source, "derived"

"A related resource from which the described resource is derived."

The history of recordURL, relatedInformation, and occurrenceDetails all have exactly the opposite meaning in my mind. All three of those were places to find additional information that was derived from some set of primary objects (specimen, field notes, map, etc). A publication that synthesises the most detailed information about an occurrence is derived from these primary sources of information. Carrying such a publication as a dcterms:source seems exactly backwards.

I wouldn't mind using dcterms:source to indicate that a specimen label is derived from field notes or that a specimen label is derived from a ledger entry. However I would object to dcterms:source being used to indicate that a publication is the source of a specimen record, when the reality is almost certainly the other way around.

I don't concur that the intent of dcterms:source is similar in definition or intent to occurrenceDetails.

-Paul

On Wed, 6 Jul 2011 10:09:02 -0700
John Wieczorek <tuco@berkeley.edu> wrote:

> I agree that the meaning of dcterms:source is different from the
> stated meaning of dwc:occurrenceDetails. The latter term originated
> with OBIS as
>
> recordURL - Gives the web address of the page where more information
> on this particular record (not on the whole dataset) can be found.
>
> and was re-cast in Darwin Core 1.4 as
>
> relatedInformation - Free text references to information not
> delivered via the conceptual schema, including URLs to specimen
> details, publications, bibliographic references, etc.
>
> and is currently defined as
>
> occurrenceDetails - A reference (publication, URI) to the most
> detailed information
> available about the Occurrence.
>
> So, one could argue that the dcterms:source is not necessarily the
> most detailed information available, and it would certainly not be
> only about Occurrences.
>
> So, occurrenceDetails certainly is no substitute for dcterms:source,
> and dcterms:source doesn't exactly circumscribe occurrenceDetails as
> defined. Nevertheless, the concepts are so nearly the same in
> definition, and certainly in intent, that I propose that adding
> dcterms:source obviates the need for occurrenceDetails as it will
> sufficiently cover the intended use of occurrenceDetails while
> allowing the same for all record types.
>
> On Tue, Jul 5, 2011 at 1:12 AM, Gregor Hagedorn
> <g.m.hagedorn@gmail.com>wrote:
>
> > I think the two terms are different.
> >
> > my understanding is that dcterms source expresses a relation
> > between a (usually digital) record and another record, that is
> > derived from the first.
> >
> > I think this is very useful in DwC in cases, where records are taken
> > directly from a publication.
> >
> > However, it does not replace the fact that the source of the record
> > is the digitization project of collection X and that the record is
> > also cited in a recent publication.
> >
> > I have doubts whether occurrenceDetail is a good label for the
> > latter concept, however.
> >
> > Gregor
> >

--
Paul J. Morris
Biodiversity Informatics Manager
Harvard University Herbaria/Museum of Comparative Zoölogy
mole@morris.net AA3SD PGP public key available