[Zoobank-list] Code fear

[Doug Yanega]

Rich Pyle wrote:

>  > When I (and probably you) first joined the Internet, there was almost
>>  no spam; Nigerian Scams, phony bank notices, identity theft engines,
>>  and all that other crap COULD have been there from day one, but it
>>  wasn't.
>
>True, but I would argue that it could not have been prevented from day one
>either -- at least not without substantial costs.  For example, one solution
>to avoid these things would be to have incorporated a mandatory peer-review
>process for every electronic document sent via the internet (email, web
>pages, files, etc.). Such a review process would have virtually eliminated
>spam, scams, identity theft, etc.  But do you REALLY think that would have
>been a wise thing to implement at the dawn of the internet?  A "moderated
>web"?  Do you think that the internet would have thrived as well as it has
>if such a mandatory peer-review system was incorported?  Would the benefits
>have exceded the costs?
>
>Hey -- you introduced the analogy -- I just ran with it!

As you said, though - the analogy is a good one, and I'll note that 
spam *could* have been stopped dead in the water if the powers that 
be had designed internet protocols to prohibit anonymity (i.e., make 
it impossible to conceal the source of a message - this IS 
technically feasible). Declaring a source as BAD was actually 
something we did in the early days of anti-spamming, and it 
continues, in a fashion, in present-day filters (it also, however, 
prevents legitimate messages from those same sources from seeing the 
light of day, which is itself a *bad* side-effect). The reason spam 
got so completely out of hand is that the spammers found out how to 
forge IDs and guarantee their messages could not be traced; i.e., 
there was no way to hold the source of a message accountable for it 
and flag that source as BAD (and then block it). If we accept the 
analogy, you claim that someone who is determined to publish bogus 
names can be stopped because we know who they are: i.e., we can 
declare the source bad and block it. But this assumes that there is 
no way that anyone can submit a name for registration without first 
having their identity checked and confirmed; a security measure that 
represents a novel burden for those submitting names, and a burden 
for those running the registry (and something not included in your 
descriptions of models 1 or 2). You also need to decide at what point 
to cut off a person's ability to submit new names; it's a rather 
subjective threshold, almost as complex to implement as the process 
of open peer review in the first place. As you said:

>and
>respond with automated warnings and/or revocation of registration privledges
>for the offending individual.  You mentioned eBay -- it has an unbelievably
>simple, but REMARKABLY effective system for rating buyers and sellers, such
>that every transaction that occurs on eBay does not need to be individually
>peer-reviewed.  I can imagine an analagous system for ZooBank, such that
>when an algorithm spots a potential fraudulent scientist, it is brought to
>the attention of a peer-review board.  That way, the peer review would focus
>only on suspected fraud taxonomists -- a much smaller task that mandatory
>peer-review of every single name (the vast majority of which would be
>legitimate).

True enough, but you would still need an arbitrary judgment as to 
when someone had committed enough offenses - or an egregious enough 
offense - to merit the suspension of registration privileges (and 
then be able to prevent them from re-registering under a new phony 
identity; the use of pseudonyms to facilitate publication of bogus 
taxa is an actual trick used by at least one publishing insect 
taxonomist). That may not be a bigger task, but it's a lot harder to 
carry it out without it being criticized as a form of censorship.

>  > Basically, if we
>>  let frauds start to auction off bogus names, then we'll lose the
>>  ability to auction off real ones
>
>In order for this to be true, it means that the potential buyers of names
>would have to feel "burned" in some way by the scammers.  That would only
>happen if the buyers found out that the names thay paid for were bogus
>(defined as immediately sunk by other taxonmists).  Thus for the scammers to
>ruin it for the rest of us, there would need to be conscious feedback to the
>buyers -- which would then self-regulate the scammers.

Actually, the other side of the problem is that scammers would focus 
on ultraglamorous taxa, meaning that a real taxonomist trying to 
auction the name for, say, a new polychaete worm, is not going to get 
any bids in competition with the thousands of birds and butterflies 
that are available to bid on. Who would bid on a genuinely new but 
uncharismatic taxon when they could have a butterfly instead?

>  > Giving
>>  people a quick way to name taxa without requiring that they be NEW
>>  taxa would be a big mistake. The only way to ensure that they're NEW
>>  taxa is to have open review - so, for example, the instant Joe Fraud
>>  submits a new butterfly name for approval, every butterfly taxonomist
>>  in the world gets an automated message alerting them that a new
>>  butterfly is being proposed, and within hours it can be confirmed
>>  that the proposed taxon already has a name.
>
>How is this not equally available in Scenario 2?

Because under Scenario 2 a name is available from the moment it is 
registered, and registration is automatic upon submission. There is 
no screening. By the time the email notice reaches the taxonomists, 
it's a fait accompli - the name is already registered, available, and 
someone will have to publish a paper in order to synonymize it. Under 
my model, it is nearly impossible for a new proto-synonym to become 
available, because of the screening process. Yes, fraudulent names 
are perhaps only a major problem in insects - more so in butterflies 
than anywhere else - but I don't think it's fair to shrug it off as 
an insignificant concern. If all a butterfly dealer has to do to turn 
a 5 dollar butterfly into a 5000 dollar butterfly is describe it as a 
new species, then they will do so: they ALREADY do this routinely, 
and butterfly taxonomy is utter chaos these days (some beetle groups, 
such as Carabus [s.l.] are almost as bad; insect dealers publish most 
of the new "taxa"). How will Scenario 1 or 2 prevent this abuse?

>  > Under models 1 and 2, he
>>  could describe and self-publish 500 taxa a week, and there would be
>>  nothing anyone could do except synonymize them after the fact
>
>I gather you haven't been reading my earlier descriptions of Scenario 2
>carefully -- further proof that we do not yet fully understand each other's
>vision, and therefore are CERTAINLY not ready to present those visions to
>the broader taxonomic community.

???? I have your earlier summaries right here, and I did read them 
quite carefully:

>Scenario 1 (Mandatory registration only)
>This scenario establishes ZooBank as the universal registry of all new
>zoological names [and nomenclatural "acts" -- however those end up being
>defined], and mandates (after the next version of the ICZN Code) that all
>new names be registered in order to be "available" according to ICZN rules.
>All other aspects of the existing (4th Edition) Code would remain in
>effect -- *including* the requirement that the new names be published
>according to existing rules (e.g., "durable media", etc.). Thus, names still
>would need to be published as they currently are, but they *also* must be
>registered within some pre-defined time window in order to become available.
>This scenario includes provisions to manage the temporally dissociated
>"publication" and "registration" events (both of which are required for name
>availability).
>
>Scenario 2 (Registration fulfills Publication)
>This scenario includes the same mandatory registration requirement as
>Scenario 1, but defines the act of registration as, itself, minimally
>fulfilling the act of publication, thereby making names "available"
>(sensu-ICZN) the moment they are registered.  Secondary publication of
>descriptions by traditional (or any other) means would certainly be
>encouraged, but would no longer be an additional *requirement* for
>nomenclatural availability over and above the act of registration. This
>scenario decouples the "science" side of taxonomy from the "nomenclature"
>side of taxonomy. The registration process would include the same minimal
>bits of information that are currently required for a name to be validly
>"published", except for the "durable media" requirement.

I see nothing in either of those summaries that indicates any review 
process or other step that would prohibit publishing/registering 
without peer review. Are you implying that there are unwritten 
refinements missing from these summaries (but that actually should 
have been there) that would act to *prohibit* it? Simply 
*discouraging* people by threatening to revoke their registration 
privileges if they're repeat offenders isn't going to have much 
effect. Or am I still missing something?

Peace,
-- 

Doug Yanega        Dept. of Entomology         Entomology Research Museum
Univ. of California - Riverside, Riverside, CA 92521-0314
phone: (951) 827-4315 (standard disclaimer: opinions are mine, not UCR's)
              http://cache.ucr.edu/~heraty/yanega.html
   "There are some enterprises in which a careful disorderliness
         is the true method" - Herman Melville, Moby Dick, Chap. 82