[Taxacom] iSpecies with Wikipedia

[Richard Pyle]

I didn't see Doug's message concerning the algorithm until Paul included it
in his reply to Doug.  So, I hereby withdraw my previous suggestion that
nobody had yet questioned the ability to create an algorithm of the sort I
described. 

> How is an 
> algorithm going to know that a new generic synonymy proposed 
> by author X in some obscure publication is an objective 
> synonymy based on an examination of the type specimens - 
> which therefore requires an IMMEDIATE change to the consensus 
> classification (even if no other "experts" have submitted an 
> opinion regarding the change, or published revised 
> classifications of their own to reflect the change) - 
> compared to, say, a generic subjective synonymy proposed in a 
> self-published source by some crackpot whose opinion isn't 
> worth the paper it's printed on? Neither synonymy may be 
> cited by any other experts (either for support or refutation) 
> for several years, but the former requires an immediate 
> change, the latter does not, and an algorithm isn't going to 
> be an effective system for figuring this out.

I guess that in the world that I imagine (and, in case it hasn't been
abundantly clear from my previous posts, it is very-much an imaginary world
at this point in history -- I'm merely contenting that it's viable in the
near-term future), publications are not the only venue through which
taxonomic assertions are made.  Why not allow direect commentary online as
soon as information becomes available?  The algorithm could easily
incorporate a buffer time-lag for new information that allows time for
feedback to be generated.  And, an assertion with no feedback at all could
be ranked low, and therefore not trump existing stable classifications until
critical mass of supportive usage has been established. And, of course, if
we as taxonomists "know" that a given individual is a crackpot whose opinion
isn't worth the paper it's printed on, then I don't see why an algorithm
couldn't "know" this as well (via the same track record that the taxonomists
use to make this assessment -- i.e., that nobody follows the crackpot in
subsequent treatments).

As for objective synonymies, it is my hope that this would be in the domain
of nomenclators, which would have their own mechanism for staying current
and staying accurate.  The algorithm would obviously incorporate the
nomenclators into it.  Also, there is a subtle difference between
classification "changes" due to objective synonymy (which really represent a
need for a name-swap in an otherwise unchanged classification), and true
classification changes that result from lumper/splitter issues and
phylogenetic hypotheses.

> The ordinary mortals who want and need a consensus 
> classification are not going to know or care about esoteric 
> debates over whether (e.g.) Aristolochia is in Piperales or 
> Aristolochiales; they're people like high school students 
> doing a class project where they have to submit the full 
> Linnaean rank hierarchy for their favorite plant, and they 
> are going to be confused and ANNOYED if what they get is some 
> listing of three or four alternative classifications. 

This is exactly the problem I'm trying to avoid -- so I don't know to whom
this comment is directed.

> I'm of a similar mind to Mary Barkworth: we need the process 
> by which consensus is arrived at to be OPEN and transparent. 
> If someone presents a clean and convincing argument for a 
> change, then let the community examine it, approve it, and go 
> ahead with it; if their proposal generates controversy, then 
> the community discussion can and SHOULD be a vigorous 
> no-holds-barred debate over its merits and its drawbacks, 
> until something definitive emerges.

You are describing *EXACTLY* the sort of system I have in mind.

But the point where I side with Mary is the concern of the authoritarian
decision by committee.  Even if the decision is made on the basis of loogic,
and in the open, it seems silly to me that we simply pretend the alternate
perspectives do not exist.  If, however, you are advocating that we
acknowledge (and keep track of) the dissenting opinions -- thereby allowing
us to maintain an open mind when we interpret new information -- then it
appears that you and I are advocating almost *exactly* the same approach.
The main reason I use the word "algorithm" to describe the mechanism for
distilling the "preferred" classification (as already pointed out, it's
technically not a "consenus"), is that it avoids the authoritarian aspects.


If I understand you correctly, in your view the "community" replaces my
"algorithm".  In my mind, there is no difference between them, because the
"algorithm" is simply the direct, collective reflection of the "community".

> It's been suggested before: a single website where every 
> taxonomist registers, and indicates what subject matter is of 
> interest to them, so that every time anyone, anywhere, makes 
> a comment that is pertinent to that specified list of topics, 
> then EVERYONE IN THE WORLD who is interested in that topic is 
> instantly notified, and invited to respond.
> I envision it as a hybrid between things like the ToL (with a 
>  giant classification tree), Wikispecies, and a chat room.

Amen, brother!  It would be great if such a site could be developed. It
would be based on standard data protocols, and could integrate a wide array
of content from many different sources (Museum collections, genetic data,
morphological character data, images, etc.).  It would be tightly integrated
with nomenclators, and have many well-developed, open-source applets and
APIs to allow external developers to access and build upon its content.  It
would help to develop clever, powerful, and simple-to-use software tools
that would make the jobs of taxonomists much easier and more efficient,
while simultaneously allowing their research and opinions to be made
available to the rest of the world via the same standard data protocols. It
would track all the different classifications gleaned from the literature
and other "meta-authorities" (ITIS, Species2000, etc.)  It would coordinate
with initiatives like the Biodiversity Heritage Library and uBio to scan and
index the historical literature. 

All we need is a catchy name (something that could be easily distilled to a
three-letter acronym), a charismatic champion who is respected both within
science and is familiar to the outside world, and a commitment from a few
major funders of about $50 million or so over the next ten years.

Aloha,
Rich