On Thu, Nov 25, 2010 at 3:13 PM, Gregor Hagedorn
<g.m.hagedorn@gmail.com> wrote:
> If this is so trivial then why does GBIF have one map for Felis concolor and
> one map for Puma concolor?
> also why does the Barcode of life not have all their Aedes triseriatus and
> Ochlerotatus triseriatus mapped to one id rather than one for each name and
> misspelling?
I think it shows that the real problem is not URLs versus strings, the
problem is the knowledge behind these strings.
Actually I think the problem lies with the co-mingled use of a scientific name. It is serving as both a stable identifier and as a phylogenetic hypothesis.
The concept of a "species" should have been separated from the hypothesis about where in the phylogenetic tree that concept fits.
The mosquito example above also demonstrates the conflict between those who want to use names as stable identifiers and those who want them to reflect the latest phylogenetic thinking.
The mosquito community seems split on this, so I would say that their is either no accepted name or two accepted names.
It certainly happens that bugs appear even after 20 years of using
URLs, but it is a class of highly generic bugs in any web-software. It
does not appear a good argument to me.
Yes, it would have been better if things like & or () did not need to be encoded for use in URL's but as a community we could simply decide to use some other characters at least for URL's etc. If we wanted to use full names in URLs.
I am not aware that people have problems linking to Wikipedia or
DBpedia, which happens to use exactly these human-proofreadable URIs.
The URLs for different Wikipedia and DBpedia entries change all the time.
I have had problems, the URLs for several mosquitoes have changed along with the URL for Carl Linnaeus since I started this.
> Also aren't TDWG URI's supposed to be opaque?
Why do we use dwc:scientificName instead of
dwc:entity013030d4a93abdd6206234b683c51b31 ?
I think there are LSID'ers who are doing something like that right now.
I am sure the semantic vocabulary management system for DarwinCore
would show the proper label for the opaque URI... :-)
This is a feature of the TripleStore (actually a Quadstore) you just have to tweak the settings to get it to work.
Basically, programmers demand human readability for their own domain,
but deny it to the biodiversity domain itself...
I think you might be mischaracterizing me. I started down this road while trying to solve biological problems.
I argued years ago on TDWG for some identifier for the species concept as opposed to an identifier for a name.
When that failed I started creating my own.
Intuitively you know that records for Felis concolor and records for Puma concolor should appear on the same map, but without some concept id, it can't happen.
In an ideal world, the name should be for the concept and the changing phylogeny should simply be new assertions about that concept.
Then you would not have millions of collection labels with the wrong name on them.
I would argue that the current nomenclatural system with all its codes and rules has little to do with real biology.
Many of the problems we are now dealing with would go away if a new system was created that was based on the biology we know today rather than trying to work around a system that was based on what people thought 250 years ago.
I don't see that happening except in a few areas like Bacteriology, so for now we have to devise complex informatics systems to work around a fundamentally flawed system.
*(My opinion)
> I just believe that those taking these decisions have a specific perspective and use case scenarios, that involves biologists only after the perfect software user interface system is finished. I challenge the last assumption
I also wish that the TDWG standards would include specific semantic uses cases and test data so that people could actually see if it works for them.
That is what I am trying to do with my examples, which are constantly being revised; but I am clearly not a member of the mysterious TDWG Illuminati.
Respectfully,
- Pete
I fully believe you and all who are doing it do it with careful
consideration of the needs as they see it. I just believe that those
taking these decisions have a specific perspective and use case
scenarios, that involves biologists only after the perfect software
user interface system is finished. I challenge the last assumption ...
Redesign tdwg vocabularies and Darwincore with opaque
dwc:concept013030d4a93abdd6206234b683c51b31 URIs instead of
dwc:commonName (where I really prefer the synonym vernacularName - or
is it the other way round?) and proof that it works well for
communication and discussion.
I believe Opaque IDs work OK if they can be systematically and
unambiguously assigned. Taxon names and concepts can not, they need to
be discussed and "debugged" probably over decades. Just like tdwg
vocabularies -- just 6 orders of magnitude greater scope.
Gregor