[tdwg-content] Darwin Core review

John R. WIECZOREK tuco at berkeley.edu
Fri Mar 6 23:19:55 CET 2009

Dear Peter and others,

Though the public commentary period for the proposed standard has not
yet opened, I find these comments quite useful for cleaning up
remaining issues and inconsistencies as they are discovered. My
responses are given in line below...

With appreciation,


On Fri, Mar 6, 2009 at 7:21 AM, Peter Desmet
<peter.desmet.cubc at gmail.com> wrote:
> Hi all,
> I had a look at the Darwin Core terms
> (http://rs.tdwg.org/dwc/terms/index.htm) and I have some questions and
> remarks.
> 1. The "Replaces"-links for terms in the "Complete historical record"
> (http://rs.tdwg.org/dwc/terms/history/index.htm) always link to
> themselves.
> See for example: HigherGeography
> (http://rs.tdwg.org/dwc/terms/history/index.htm#HigherGeography-2009-01-21).
> The "Replaces"-link links to same 2009-01-21 instead of 2007-04-17
> (http://rs.tdwg.org/dwc/terms/history/index.htm#HigherGeography-2007-04-17).
> And this 2007-04-17 also links to 2007-04-17 instead of the older
> 2003-06-17.
> A link for "Is Replaced By" instead of text would also be useful.

I fixed the problematic Replaces links and added links for Is Replaced By.

> 2. As a result, I had some difficulties mapping the terms from DwC 1.4
> (http://wiki.tdwg.org/twiki/bin/view/DarwinCore/DarwinCoreDraftStandard)
> with the new ones. Update: Until I found the Darwin Core Versions
> document (http://rs.tdwg.org/dwc/terms/history/versions/index.htm).
> 3. Are the terms StartDayOfYear, EndDayOfYear, StartTimeOfDay,
> EndTimeOfDay, YearSampled, MonthOfYear, DayOfMonth necessary if you
> can provide this information in EarliestDateCollected,
> LatestDateCollected? Or is this to provide e.g. the time at which
> something was collected if you don't know the date (I don't know if
> you can express this in EarliestDateCollected)? If so, YearSampled can
> always be expressed in EarliestDateCollected/LatestDateCollected.

The terms are needed to maintain the requirements of previous versions
of the Darwin Core, namely to capture atomized temporal information in
the event that one or more parts of an actual date are missing, for
which in some some cases the (for example a month and day without a
year) could not be captured in the date-based terms. A separate
consideration, which applies especially to Years, is that an atomized
term greatly simplifies actual use in applications.

> 4. Is it possible to provide a different example for SampleRemarks and
> SamplingEventRemarks for a clearer difference between the two? Now
> both have "found dead on road".

I have changed the SamplingEventRemarks example.

> 5. Why is ValidDistributionFlag removed? This looked like an
> interesting field for all our botanical garden specimens. We now have
> to express this in SamplingProtocol, which will need a controlled
> vocabulary for sorting these out.

The mapping to SamplingProtocol was incorrect. It has been amended to
point to EstablishmentMeans, which does not preclude the original use
intended for this element.

> 6. Is Behavior an attribute of a SamplingEvent or Sample?

Sample. This error has been amended and a Decision has been created to
reflect this change of Domain

> 7. I'm happy with addition of the term "PreviousIdentifications". Is
> it possible to provide the previous identificators and dates there as
> well?

Though it is possible, I wouldn't recommend it. It is always best to
stick strictly to the semantics of the term. If you require the more
complete histories of identifications with IdentifiedBy and
IdentifiedDate, then you should consider using an application profile
having structure with more than one Identification per sample. The
Simple Darwin Core, being flat, does not support this except in the
limited way described above.
I would be possible to capture all of the information for multiple
identifications in PreviousIdentifications so that the Simple Darwin
Core could support it. All that would have to be done is change is the
description of the term. If that can be supported as desirable, make a
case and let's see what falls out in the commentary period.

> 8. Can you add multiple "ResourceRelationship", "SampleAttribute",
> "EventAttribute" to one sample? If yes, how does this work in a flat
> schema like DwC?

It doesn't, for the reasons described above for
PreviousIdentifications. Instead, you would have to use the terms
provided for these purposes (AssociatedMedia, AssociatedReferences,
AssociatedSamples, AssociatedSequnces, AssociatedTaxa,
SampleAttributes, EventAttributes).

> 9. The name "SamplingAttributeID"
> (http://rs.tdwg.org/dwc/terms/index.htm#SamplingAttributeID) should
> probably be changed to "SampleAttributeID".

Agreed and changed.

> 10. What terms are mandatory?

The mandatory status is not an attribute of the term itself, rather,
it depends on the application. In Simple Darwin Core
(http://rs.tdwg.org/dwc/xsd/tdwg_dwc_simple.xsd) the required terms
are modified, SampleID, InstitutionCode, CollectionCode,
BasisOfRecord, and ScientificName. For those familiar with previous
versions of Darwin Core, the first two of these have been known as
DateLastModified and GlobalUniqueIdentifier respectively.

> Best regards,
> Peter
> --
> Peter Desmet
> Biodiversity Informatics Manager / Directeur de l'informatique de la
> biodiversité
> Montréal Biodiversity Centre / Centre sur la biodiversité de Montréal
> Canadensys - www.canadensys.net
> Institut de recherche en biologie végétale
> Université de Montréal
> 4101, rue Sherbrooke est
> Montréal, Québec, H1X 2B2
> Canada
> Télé / Phone: 514-872-5249
> Télécopie / Fax: 514-872-9406
> Email: peter.desmet.cubc at gmail.com
> _______________________________________________
> tdwg-content mailing list
> tdwg-content at lists.tdwg.org
> http://lists.tdwg.org/mailman/listinfo/tdwg-content

More information about the tdwg-content mailing list