[tcs-lc] concepts of Higher taxa

Richard Pyle deepreef at bishopmuseum.org
Wed Apr 6 10:36:24 PDT 2005


> OK, it is verbose,

Agreed -- but I think most datasets would be proportionally smaller, because
of the re-use of Nominal instances.

> but ultimately it
> a) allows name-based reasoning about the implied concepts, and
> b) it allows to filter them out when presenting concepts for user
> selection
>
> It works but I still don't quite like it. To me the implied concepts are
> somewhat too similar to taxon name reference with undefined
> circumscription concept as used in almost every specimen identification.

Yes, but ultimately there is a complete spectrum of real-world cases, from
the most perfectly-defined concepts, to the name-only usage (and every
imaginable circumstance in-between).  TCS currently splits these up into
three categories:

- Defined Concepts
("Original", "Revision", "Incomplete" -- I assume the latter means the
concept was defined, but not fully entered into the data proivder)

- Completely Undefined Concepts
("Nominal")

- Name Usages
(Not treated as concepts, but rather datasets linked to concepts)

I've inserted a fourth category ("Implied") between Defined and Completely
Undefined; which effectively amounts to "Partially Defined".

Perhaps this fourth category doesn't serve a valid purpose, but especially
for legacy data, I actually think it will be the most common situation.

> It seems that the
> paradigm established here may cause a similar handling on the
> identifications,
> i.e. as implied concepts - which they certainly are.

Again, I ask: other than the consumption of GUIDs, what's the harm?  Local
users can establish all the Implied concepts they want for their own local
needs. If they don't take the trouble to map them to defined concepts, then
the rest of the world can easily ignore them (i.e., they are effectively
invisible to the rest of the world).  Even if they are mapped to defined
Concepts, they can easily be filtered.  Moreover, they could be transparent
to the data consumer, with some very simple software logic. For example,
provider software could very easily recognize this pattern:

SpecimenObject -->[Identified as]--> Implied Concept -->[Mapped as Congruent
to]--> Defined Concept

...and reduce the returned value to this:

SpecimenObject -->[Assigned to]--> Defined Concept

If the data provider failed to map the Implied Concept to a Defined Concept,
then the returned value could default to this:

SpecimenObject -->[Assigned to]--> Nominal Concept

You can replace "SpecimenObject" with any data unit you want (e.g.,
EcologicalObservation, etc.)

How is this any different from current TCS?

>From my perspective, allowing "Implied" Concepts not only accurately
reflects taxonomic practice (and historical reality -- note Paul Kirk's
earlier comment), but also provides flexibility of implementation without
significantly increasing risk of inconsistency of transferred data. The only
cost is GUIDs, and as said before, GUIDs are cheap.

Rich




More information about the Tcs-lc mailing list