[tcs-lc] Names as Objects

[Gregor Hagedorn]

> > can not separate issues. There is no way in w3c schema to ONLY use TCS
> > with type attribute having value = "nominal".
> 
> Why is that a bad thing?

Because it makes TCS non-reusable, except as a totality. SDD has similar 
problems, and we are thinking hard (pushing things out into UBIF, searching for 
new design patterns see the Objectxxx wiki topics) on how to restructure it.

> > You MUST use the entire type and since that the is a
> > complex type containing most TCS types (or containing elements
> > that depend on
> > TCS literature and voucher) almost the entire TCS schema.
> 
> Well...how else *would* you pass voucher data and literature data?  I don't
> believe that the "v090b-RLP" versions of the schema that I sent last night
> require any of those elements -- unless you specifically need to pass details
> about those elements.  There could always be whatever minimal set of
> "XxxxSimple" elements embedded within the TCS/LC instance that you need. But
> even if we did treat names as top-level objects, wouldn't it be preferable to
> keep voucher and literature details external to this, in their own top-level
> element containers (with internal reference pointers)?  Or do you want to embed
> an entire literature schema and an entire specimen data schema within TCS/LC
> "Name" container?

No, I exactly argue for the the freedom to only use an appropriate level of 
name knowledge. I argue for an object that can be re-used in online monographs 
to express nomenclatural knowledge. And as argued for SDD separately, SDD needs 
an ID on such name objects.

I see nothing wrong for many use cases to exchange taxonomic nomenclatural 
knowledge in nominal TCS records, except that people like Bob and me kind of 
like to have at least some syntactical validation (which is why we use xml 
schema at all). My argument there are many more use cases, and if TCS cannot 
provide something reusable, for example SDD and ABCD will have to stick with 
its own types for names as is currently already for years the case. I don't 
think that is desirable.

> > You cannot have a new publication already indexed in a database
> > that indexes
> > already published publications. Putting it first in the indexing
> > database would
> > be illegal under the nomenclatural codes, since then the name
> > would first have
> > been invalidly published.
> 
> Hmmm....not in ICZN.  An indexing database does not constitute a
> "publication", and therefore adding a name to an indexing database prior to the
> publication of the new name does not render it invalidated (unless the index is
> somehow "published", as defined by ICZN).

At the momement, *no* electronic publication is possible under ICBN at all, so 
the scenario is hypothetical. If such publication would become possible, an 
explicit exception for indexing databases could be made to avoid flooding 
nomenclature with nomen nudum cases.

However, I still think it would become a hindrance to taxonomic progress if 
anybody who want to publish a new species in digital form in the future would 
first have to go to a concept indexing service, "pre-register" the name, then 
use an GUID token obtained from there to insert into the publication that 
actually describes the new taxon, provides the character circumscription, 
ecological, distribution, synonymy and other information.

Only part of this information can be captured in current TCS. I see TCS, SDD, 
as building blocks in that future big thing, not as the germ to grow into it. 
Each of these and others could become everything...

Gregor----------------------------------------------------------
Gregor Hagedorn (G.Hagedorn at bba.de)
Institute for Plant Virology, Microbiology, and Biosafety
Federal Research Center for Agriculture and Forestry (BBA)
Königin-Luise-Str. 19           Tel: +49-30-8304-2220
14195 Berlin, Germany           Fax: +49-30-8304-2203