[Taxacom] Data query
stephen_thorpe at yahoo.co.nz
Tue Jun 25 04:09:37 CDT 2013
Not quite 'nuff said! We have reached an interesting point in the identifier debate! I may be wrong, but in order to record mappings of the kind “ZB: 8BDC0735-FEA4-4298-83FA-D04F67C3FBEC maps to WoRMS:345844”, doesn't this have to be done manually and painstakingly?
From: Richard Pyle <deepreef at bishopmuseum.org>
To: 'Stephen Thorpe' <stephen_thorpe at yahoo.co.nz>; Tony.Rees at csiro.au; mesibov at southcom.com.au
Cc: taxacom at mailman.nhm.ku.edu
Sent: Tuesday, 25 June 2013 8:27 PM
Subject: RE: [Taxacom] Data query
Well, I didn’t really want to go down that path, but since you brought it up….
First of all, we have a service that already does that:
In other words, there’s your URL for the name. It takes you to a disambiguation page, rather than directly to the page. It would be fairly easy to ‘smarten’ this service up to catch cases where there is only one search result and auto re-direct to that page (ala Google’s “I’m Feeling Lucky” feature).
But frankly, I wouldn’t want to encourage people to link to a service as though it were an identifier. Yes, I know that in some circles, all URLs *are* identifiers; but some URLs are more optimally designed for identification than others. And as much as I am a HUGE defender of Linnean names as human-friendly identifiers (being an ICZN Commissioner and all), more than three decades of experience working with computer databases to track organism names has taught me nothing if not that scientific names make for really, really, really, REALLY bad identifiers for computer systems. The problem is not that text strings as such cannot be used by computers as identifiers. Rather, the problems mostly involve bad human behavior when confronted with a familiar-looking thing. The temptation to change that thing is just too powerful. Good Identifiers just don’t want to be changed.
I fear that if I launched into a more elaborated explanation than that, I would be typing this email for another hour. I’d rather add a half-dozen identifiers each to 60 different ZooBank records than spend another hour writing email….
For now, I’ll just leave it as this: I currently established WikiSpecies as an “Identifier Domain” (mostly for your benefit, in honor of the many hours of work you have personally put into adding content to ZooBank) – which means that I treat the URLs as “identifiers” rather than services. However, in a sense, it’s more of a service. I don’t really need to establish a “hard link” of this sort to WikiSpecies, because it’s much easier to establish a “soft link”. The idea of a “soft link” is that instead of explicitly recording in the database “ZB: 8BDC0735-FEA4-4298-83FA-D04F67C3FBEC maps to WS:Chromis_abyssus”, instead I can simply take the name-string (“Chromis abyssus”), replace the space with an underscore, and slap that at the end of “http://species.wikimedia.org/wiki/”. Before displaying the link to the user, I can ping that page, and only display the link if I get a real WS page. In other words,
there’s no reason for me to explicitly make that link, because I can make the link on the fly. This is the advantage of having a “soft” link. But there are other advantages to using opaque identifiers (e.g., DOIs … hint, hint, Rod) that are proper actionable GUIDs – and that’s where the “hard links” come into play. An example of this, that must be explicitly recorded in the database as a cross-link (instead of created on the fly as a soft link), is “ZB: 8BDC0735-FEA4-4298-83FA-D04F67C3FBEC maps to WoRMS:345844”.
OK, ‘nuff said.
From:Stephen Thorpe [mailto:stephen_thorpe at yahoo.co.nz]
Sent: Monday, June 24, 2013 9:35 PM
To: Richard Pyle; Tony.Rees at csiro.au; mesibov at southcom.com.au
Cc: taxacom at mailman.nhm.ku.edu
Subject: Re: [Taxacom] Data query
Actually Rich, you might think seriously about creating ZooBank URLs, in addition to, and which redirect to the current ones based on LSIDs, but based instead on the name, as Wikispecies does. This would greatly facilitate automatic link creation from WS to ZooBank ...
More information about the Taxacom