[Taxacom] Data query

Richard Pyle deepreef at bishopmuseum.org
Tue Jun 25 05:41:11 CDT 2013


Here's another way to look at it:

If you use names as identifiers, then EVERY human who EVER tries to link to that name will ALWAYS need to manually disambiguate.

On the other hand, if you have hidden GUIDs working behind the scenes, then only ONE person needs to disambiguate it ONCE, after which NO ONE ever needs to disambiguate ever again.

For example, suppose ZooBank used a URL along the lines of http://zoobank.org/Chromis (doesn't work, but pretend that it does).
The result would be a disambiguation page along the lines of:

Chromis Cuvier, 1814  (http://zoobank.org/B8F9F80D-5798-4342-810D-D8274164F8F1)
Chromis Gronow in Gray, 1854 (http://zoobank.org/EBF02DF4-0D67-4865-AFA1-A71FCB56BBBA)
Chromis Günther, 1862 (http://zoobank.org/94E77D5F-BCAE-4B06-8656-C02D2FE04E97)
Chromis Oken & Cuvier in Oken, 1817 (http://zoobank.org/754166A3-E1E3-401E-9FF3-1D9EBA20F8C8)
Chromis Plumier in Lacépède, 1801 (http://zoobank.org/9AE8FD1E-2B56-499D-8F54-CBB4170EBFA8)

Then, suppose they wanted to link to the same record in the Catalog of Fishes.  If they only had a link for something like "http://researcharchive.calacademy.org/research/ichthyology/catalog/Chromis", then we'd get another disambiguation page for the five records in CoF.

That means that EVERY TIME someone wants to jump from ZooBank to CoF or back, they'd need to pass through a disambiguation page.  To a human, this is merely annoying.  To a computer service, this is effectively a show-stopper.

OK, so forget homonyms for a moment.

I just took a look on WikiSPecies for one of the groups I did my PhD work on (genus Centropyge).

It was really easy for me to find this page (+1 for names-as identifiers): 
http://species.wikimedia.org/wiki/Centropyge

So far, so good.

Then I took a closer look.  One of the species is Centropyge aurantonotus:
http://species.wikimedia.org/wiki/Centropyge_aurantonotus

Suppose a bit more research reveals that "aurantonotus" was used as an adjective.  Oops!  Now it needs to be changed to "Centropyge aurantonota".  How does WikiSpecies handle this?  Does it keep the identifier as http://species.wikimedia.org/wiki/Centropyge_aurantonotus and then display the name as "Centropyge aurantonota"?  If so, how will someone who knows the name via the correct spelling ever find the WikiSpecies page?  They just get this: http://species.wikimedia.org/wiki/Centropyge_aurantonota -- and assume they need to create a new page.  Or, if you change the identifier to match the correct spelling, then what happens to anyone who ever linked to the old spelling?  Do those links just break?  Or does WikiSPecies re-direct from the old spelling to the correct one?  And what about when new genetic research reveals that the genus Xiphypops is good, and the name changes to Xiphypops aurantonotus?  What happens to the WikiSpecies link?  What about the people who don't know Xiphypops is masculine, and try to find Xiphypops aurantonota?  Or what about the common misspelling of the genus: Xiphipops?

And what about my good old friend Centropyge vrolikii?  And all the various names and spellings (Centropyge vroliki, Centropyge vrolicki, Holacanthus vrolikii, etc.)  Do we need redirections from all of these?

Dirty buckets and human web pages work perfectly fine with names as identifiers.  Clean buckets need more precision to work effectively (and to do things that Dirty Buckets can only dream of doing).

Rich







More information about the Taxacom mailing list