Rod

Bizarre is not the word I would use, having (because of my association with the Global Names project) some appreciation of the extent of the problem. But, I am certainly embarassed that we have made so little effective progress towards what is such an obvious goal.

Progress, I think, requires a more analytical perspective, and a willingness to work collaboratively.

Our sub-discipline contrasts massively with, for example, the molecular domain, where open sharing of content is the norm. As a community we have shown much less readiness to share names and taxonomic content. So, a 'social' change needed there. We can provide the tools, but attitudes will need to change. And change does not just relate to people and content, but also a change relating to development of services and software. New tools need to be designed so that they can be interconnected, creating a much better open toolkit. Money, and especially money not tied to short duration projects

Then we require an infrastructure that allows those who are willing to share content, annotate names as being valid now according to someone, declare synonymies, flag chresonyms, disambiguate homonyms, interconnect lexical variants, offer alternative taxonomic perspective, and provide means for integrating vernacular names and names for surrogates. But that's not too hard, is it?

That structure is in need of the means of capturing ALL new changes that occur out there.

Rod, you have been significant in showing us that it is feasible to make massive progress, but then you are equally aware of the spectrum of outlets and players. That diversity defies simple and quick solutions, and frustrates the universal fix. The result should be seen as a process that will improve with time. That process will need some kind of overview, but an overview that is aware that resources for developments that go well beyond 'proof of concept' are lacking.

Given the diversity of players and the wealth of expertise, the solution needs to enable crowd sourcing. Such crowd sourcing would include the capacity of anyone to comment on any element of information, for some players to have authority to make changes, for all initiatives to be interlinked, and for an alert system that keeps all interested parties aware of all changes as they happen. The expectation is that this evolve to synchrony of all parts. Crows sourcing needs to accept that there is more than one point of view as to what constitutes an entity, and how the entities should be arranged.

With such a structure, what progress might we get?

1. Is this a name? This answer needs access to all names and all variants. Global Names has about 22,000,000 strings that are purported to be names, but the contents of this are extremely dirty (intentionally so). To find a string in GNI certainly does not mean that it will be a name. So, that needs to be fixed by flagging entries as 'names' and 'not names'. Some of the other weaknesses in Global Names are those taxonomic territories without good coverage or where content sharing does not happen. A lot of namestrings from the older literature have yet to be added. Taxonomic sources tend not to be comprehensive with synonyms, certainly not with lexical variants, are often contaminated with chresonyms, and vary in terms of taxonomic currency. There is often too much distance between expert and point of contact with the product - making it difficult to correct errors. So, these are a few of the issues on this front. It would be useful if we could run some exercises to assess how close to the asymptote we are. It would also be of interest to get input on priorities, given that there are many tasks.

2. Is this the correct way to write it? I would suggest this question needs to be rephrased, given that there are many correct ways to write a scientific name - mostly with variations in the authority department. If we limit ourselves to thinking about the correct spelling of the scientific elements, then this remains the responsibility of the nomenclators. Making nomenclator content open and placing nomenclators within crowd sourcing will help to overcome issues of scale. Historically, nomenclators have defined their own taxonomic context; but this is no longer a feasible stance. Homonyms - as Tony Rees has pointed out abound, some more taxa become ambiregnal, some less so. Open-ness of nomenclatoral content is an issue.

3. Is this name currently in use? Google can probably provide an answer of sorts to that question, but again I suspect this is a question to be refined. Is the question: Is this name, currently, considered to be a nomenclaturally valid for a taxon? Given that species concepts are rarely universally accepted, and that understanding of relationships is improving such that binomials change, then I am also assuming that we accept that there will be more than one list of valid names for the same taxonomic area. Building that polytheism into the infrastructure is not too challenging, but will the experts accept that, and will the consumer accept it?

4. What other names are related to this name (e.g., synonyms, lexical variants)? Yes, this is what we called, in our TREE paper, 'reconciliation'. Lexical variants are probably the simplest to deal with, but algorithms that seek to scale to all name strings run into problems where the fuzziness of lexical variants of similar names overlap, and that leads to massive aggregates of names - many of which are not lexical variants of all others. There will never be a perfect algorithm for this, and this also needs to have a human interface that allows results to be refined. Homotypic synonyms (objective synonyms) can probably be found with fair success by algorithms. They are, like heterotypic synonyms, embedded in taxonomic treatments. Which leads us to the issue of access to taxonomic treatments. Can we make more openly available, and in a form that the content will flow to all of us, and can we create an infrastructure that any complementary data can flow back to reward the participating taxonomists. Taxonomic treatments differ in completeness, may take differing and equally valid perspectives on the same taxa, so we still have a long way to go to merge the compatible thoughts and separate the incompatible ones.

5. Where was this name published? Can I see that publication? The nomenclators should be our primary point of reference for the first point. Nomenclators for prokaryotes, fungi, plants, and viruses are reasonably good. The protists (especially the heterotrophs) and animals present us with massive problems. ZooBank is setting up the infrastructure to help make progress on animals, and with Index Animalium and (let's cross our fingers) Nomenclator Zoologicus included, there will be a reasonably good generic framework. What incentives will entice Zoological Record to make their content openly available. After that, we are back to the taxonomists, and helping them to make their content openly available and ensuring they are rewarded for doing so. A proximate task, one that you demonstrated very nicely, is to reconcile the alternative ways of pointing to a reference. You are probably in a better position to estimate what proportion of the relevant literature is in a digital format, is indexed, and is not behind a paywall or closed off by copyright issues.

Sorry for the length, but it is my way of saying that the task is not a small one, and I am sure I have missed out many many issues. This is a task that GBIF, TDWG and others have in mind, but if anything the most proximate problem is that the focus is still too diffuse.

David Patterson

On Wed, Feb 15, 2012 at 4:50 AM, Roderic Page <r.p...@bio.gla.ac.uk> wrote:

--

___________________________________ David J Patterson

Senior Scientist, Marine Biological Laboratory Life Sciences Lead, Data Conservancy globalnames.org

7 MBL Street, Woods Hole, MASS 02543, USA.

_______________________________________________

Taxacom Mailing List Taxa...@mailman.nhm.ku.edu http://mailman.nhm.ku.edu/mailman/listinfo/taxacom

The Taxacom archive going back to 1992 may be searched with either of these
methods:

(1) by visiting http://taxacom.markmail.org

(2) a Google search specified as: site:mailman.nhm.ku.edu/pipermail/taxacom
your search terms here