Hi all,
I just had a quick look at the first few thousand data records coming into OBIS for my region (Australia). Just about every supplier who includes authority as dwc:scientificNameAuthor has used dwc:scientificName "incorrectly" i.e., for the canonical name not the canonical name + author. This data then flows into GBIF, ALA, etc. and circulates in this form. So "users" are already ignoring the definition of dwc:scientificName in practice, it would seem, with no apparent ill effects (?) - not sure whether this is good or bad, hence the title of my original question which prompted this thread...
- Tony
-----Original Message----- From: Chuck Miller [mailto:Chuck.Miller@mobot.org] Sent: Wednesday, 24 November 2010 9:56 AM To: Richard Pyle Cc: Rees, Tony (CMAR, Hobart); dremsen@gbif.org; tdwg- content@lists.tdwg.org; dmozzherin@eol.org Subject: Re: [tdwg-content] [tdwg-tag] Inclusion of authorship in DwCscientificName: good or bad?
Rich, I gather your reason would be because it's unclear if anyone would actually use a canonicalName element? That is, it's unneeded. So, following on, who says they need a dwc:canonicalName element?
You said you worry about feature creep. I suppose I worry about semantic creep. Extending the meaning of a term makes it more universal, but in a data world it increases the variability of the data that may be found attached to the term in some dataset. Imprecision in terms can create a lot of data quality headaches. Is that acceptable?
Chuck
On Nov 23, 2010, at 3:52 PM, "Richard Pyle" deepreef@bishopmuseum.org wrote:
What is the specific objection to adding canonicalName to DwC as an optional element, other than the fact it makes DwC one thing larger?
I don't have an objection to it per se, but I'd like to feel more
certain
that I understand exactly what it is, and what it is intended to
achieve,
that is not already achievable with existing terms and/or couldn't be
more
achievable with an alternative solution. I think there is value in
avoiding
feature-creep with DwC, except when we can solve a real problem with the existing terms. I agree there is a problem there, but I'm still
struggling
to understand exactly what specific problem that something like canonicalName will solve.
There are databases which do not have their names parsed and provide whatever they have recorded as ScientificName. But, there are also databases which do have parsed names and could provide this more narrowly defined element, in addition to the ScientificName. Those databases could make use of a dwc:canonicalName element in their data exchange or query response.
Right -- but the point is this: if the data are already parsed, where is
the
failure of the existing DwC terms in providing the desired service?
We've
already identified one of those: i.e., that "intermediate" uninomial
ranks
not supported by existing DwC terms don't have a place to put the
canonical
form of the name (other than scientificName, which isn't currently
intended
or required to be canonical). So yes, that's a clear problem in need of
a
soultion. But is a generic canaonicalName term really going to solve
that
efficiently/effectively? What other problems might canonicalName solve?
What we don't have and I think never will have is perfectly consistent names data from every database in the world. One reason is a mountain of inconsistently recorded legacy data from decades past that stands in the way of perfection. Another is variation in convention or tradition for a variety of reasons that have been explored in these recent threads. So, I think the pragmatic approach is to accept the inconsistencies and work around them.
Agreed! And my questions are:
- What specific problems with existing DwC do we wish to solve?
- How best to solve them?
I'll list two examples for #1:
A) Representing the canonical (sans-authorship) form of a uninomial name
at
a rank not already represented by existing rank-specific DwC terms
(kingdom,
phylum, class, order, family, genus) Because the current definition of dwc:scientificName allows (optionally)
the
inclusion of authorship information, there is no clean way to represent
a
uninomial name in a way that expressly excludes authorship -- except if
the
uninomial name happens to be represented at the rank of kingdom, phylum, class, order, family, or genus.
B) Content providers who have authorship data in a separate field from
taxon
name data, but who have not parsed the bits of a taxon name string In this case, the provider cannot provide the parsed bits of the name,
but
can provide a (sort of) canonicalName string separately from an
authorship
string. If they concatenate the authorship string with the taxon name string when populating dwc:scientificName, then the consumer has no easy
way
of extracting the name bits from the authorship bits (unless the
provider
also provides dwc:scientificNameAuthorship, wich could be exactly
removed
from the dwc:scientificName valu, yielding what the provider would have otherwised provided as canonicalName. Or, as David suggested, in this
case
the Authorship text would not be concatenated with scientificName.
I would like to know some other problems that could be solved with the addition of a canonicalName term before I start commenting on #2.
Aloha, Rich