[Taxacom] Propagation of bad sameAs statements

joel sachs jsachs at csee.umbc.edu
Fri Sep 10 09:17:14 CDT 2010


Stephen,

The invasive species definitions you complained about have nothing to do
with Linked Data. They are in dbpedia because they are in Wikipedia.
(Wikipedia introductions are mapped to dbpedia abstracts.) So you're free
to change them at their source.

That said, I'm kind of with you here. At last year's Phyloinformatics
VoCamp, Mark Wilkinson had a slide along the lines of "Because I said so!"
being an illegitimate answer to scientific questions. His point was
that we need to build description logics that give necessary and
sufficient conditions to classify all manner of instances.

But with invasive species, it's almost all "because I said so." That is,
things are defined as invasive by being put on a list, often through
treaty or legislation. So rather than give characteristics of invasive
species that would be very hard for a classifier to compute, we can 
create a class for each list of invaders, and then, for each
species on the list, assert that it is a subclass. For example,

<owl:Class rdf:about="http://spire.umbc.edu/ethan/Alectoris_chukar">
<rdfs:subClassOf rdf:resource="#CaliforniaInvasiveAnimalsThing" />
</owl:Class>

(Whether we should be modeling species as classes is, of course, another 
question.)

Since conservation status also comes down to a species being on a
list, we can deal with endangered, threatened, etc. species the same way. 
(We did this during the Spire project and I'd be happy to share examples 
with those interested.)

So if you say that precise definitions of "invasice species" aren't
helpful (or possible), I agree.

What about precise definitions of individual taxa? How should we define
the Chukar partrige? Morphologically? Genetically? Behaviourally? We can
give descriptions, with various levels of precision, but, ultimately,
"Some taxonomist said so" is the ultimate answer to why a population is
classified a particular way. I liked Jim's namespace example because, for
us to share a common semantics, it's easier to say 
"Stephen_Thorpe:Ptinus_tectus" than it is to express your definition
formally. In other words, I'd rather explore someone's
headspace by browsing their tag cloud than by translating their brain into
OWL.

*That* said, I'd love to be proved wrong. For example, there were 
papers a few years ago 
(e.g. http://dcpapers.dublincore.org/ojs/pubs/article/viewFile/808/804) about 
the possibility of using OWL-DL as the basis for a polyclave key. A user would enter some characteristics of the thing 
observed, and a classifier would decide what things it could be. I'm 
curious to know how these efforts turned out. Maybe we can have a 
"DL-based key" BoF at TDWG.

Joel.


On Thu, 9 Sep 2010, Stephen Thorpe wrote:

> well, I did seriously offer a definition of "invasive species", viz.

an invasive species is one which has a definite tendency to expand its range 
into areas where it is unwanted

Definitions which define them in terms of particular or general "negative 
effects" of the species are not good, because there are a great many 
possible particular effects and the general concept is not well-defined, unless 
you define it as an "unwanted effect", and then with a bit of simplification you 
get my definition. Basically, it is invasive if border control wants to keep it 
out ...

Adding extra bits to the definition just serves to confuddle (my word!) the 
issues ...



________________________________
From: Jim Croft <jim.croft at gmail.com>
To: Stephen Thorpe <stephen_thorpe at yahoo.co.nz>
Cc: joel sachs <jsachs at csee.umbc.edu>; taxacom at mailman.nhm.ku.edu
Sent: Fri, 10 September, 2010 5:07:35 PM
Subject: Re: [Taxacom] Propagation of bad sameAs statements

On Fri, Sep 10, 2010 at 2:49 PM, Stephen Thorpe
<stephen_thorpe at yahoo.co.nz> wrote:
> So, we define everything in terms of everything else, and end up with a huge
> circularity in which it is hard to see that we are any the wiser!

It only becomes circular if you let it.  If there are terms that tend
to be ambiguous, just don't let them travel without the namespace or
context in which they were defined.

For invasive species, find a definition you like and anchor it (I
would be really surprised if this term was not rigourously defined
somewhere in a scientific or semi official context).  Or, if the
definition does not cover what you want, make you own up and anchor it
to that - then, when you start your chatter, people will know they are
dealing with stephen.thorpe:invasive_species  which may or may not be
the 'same_as' or 'near_as_damnit' tosomebody else's definition of the
term.

> Actually, I'm all for pedantry...

it has been noticed... :)

jim
-- 
_________________
Jim Croft ~ jim.croft at gmail.com ~ +61-2-62509499 ~
http://www.google.com/profiles/jim.croft
'A civilized society is one which tolerates eccentricity to the point
of doubtful sanity.'
 - Robert Frost, poet (1874-1963)


More information about the Taxacom mailing list