[SEEK-Taxon] Thoughts on GUIDs

[Nico M. Franz]

Hi Rich, et al.:

    I find this discussion quite productive! The more I think about the 
concepts "definitions" and concept relations, the less I'm convinced it's 
something *we* ought to invest too much emotion in. Meaning: we need to 
look at practice and get a grasp of the RANGE of notions of stability and 
change perceived by taxonomists within and between taxonomic treatments. 
Then try to produce a system that can support (most of, at least) that 
range. Then let taxonomists play with it.

    Some will define "sameness" in terms of "subsumed constituents" (I 
believe that's an ostensive definition; "pointing at" something). Others 
may consider things like diagnoses of (evolving) properties as a more 
useful criterion for sameness (intentional, though the actual extension is 
subject to argument). Then there's already nomenclatural synonymy 
(referring to the identity of types) which does some but not a lot of 
cognitive work. Each way to look at it comes with strengths and weaknesses. 
Humans usually pick and choose what's best for their own understanding, 
which may be different from what a machine can understand. But a machine 
CAN spit back "congruent (=) in terms on properties [acc. to specialists 
X], yet overlapping (><) in terms of constituents [list of things now 
excluded yet then included; and vice-versa]".

    When computers handle this kind of information to make calls about 
concept synonymy, they may perform more reliably with ostensive 
definitions, especially at lower taxonomic levels. To capture what humans 
sense about the stability of meanings of names, particularly for 
higher-level taxa, properties connected to a small sample of exemplars 
(i.e. a sophisticated way to do typification) will probably take over.

    I believe I'm not just making this up, but rather, this seems to be 
what the German moss list authors - my personal and unfortunately very 
anonymous concept champs - ended up doing. They started with the 
constituent approach, and then abandoned it each time it "felt" wrong based 
on what they knew about properties. They never went beyond genus, I ASSUME 
because things started getting very awkward (not for computers, though!).

    So you're right, Bill, there are multiple ways to read an "=" sign. I 
think we need to make those alternatives explicit and available side by 
side, rather than put all our own money on just one of them. Apparently, 
that's not what practitioners (should) do.

Let's keep this going!

Cheers,

Nico



At 09:27 AM 5/27/2004 -1000, Richard Pyle wrote:
> > guid() is the function that, given a concept returns its guid. Whether
> > it comes from the concept instance document or the algorithm that
> > assigns it ought to be immaterial.
> >
> > S here was meant to be the relation "isTheSameConceptAs". So in natural
> > language, that expression would be the requirement that
> >
> >    c1 is the same concept as c2 if, and only if, their guids are identical
>
>I don't mean to sound like Bill Clinton in trying to define the word
>"is"....but in this case, I think it's important to define what we really
>mean by "same concept as".  From how you describe it, I see "same concept
>as" in the very strictest sense.  For example, if I have a dataset that
>reports Species X from Locality Y, that dataset would need to assign
>"Species X" to a particular concept GUID (of which there may be many for
>"Species X"), and then that concept GUID is matched indentically to some
>universal metadataset for the GUID, as well as to other datasets that
>reference the same GUID.
>
>This is a very different "same concept as" function of "S" than would be
>another core component of the SEEK model, which is to make the statement:
>
>"Concept GUID 1234 is congruent with concept GUID 4567"
>
>They maintain their separate GUID values because they are *potentially*
>different (that is, that congruency is an interpretation, rather than an
>objective statement).
>
> >   If c3 is a synonym of c2 and c2 is a synonym of c1, then c3 is a
> > synonym of c1
>
>I take it that your use of "is a synonym of" is different from "is the same
>concept as" (i.e., that c1, c2, and c3 would necessarily have different
>GUIDs)?
>
>Aloha,
>Rich
>_______________________________________________
>seek-taxon mailing list
>seek-taxon at ecoinformatics.org
>http://www.ecoinformatics.org/mailman/listinfo/seek-taxon