Open Access Archivangelism

[This comment was written before I read Richard Poynder's Interview of Tim Gowers. In part 2 I comment after having read the posting.]

---

Part 1

I don't know about Richard, but I have not despaired of green, or green mandates; I've just grown tired of waiting.

I don't see pre-emptive gold (i.e., pre-green "fool's gold") as an alternative but as just another delay factor, the principal delay factor being human sluggishness. 

And I think the notion of a "flip" to fool's gold is incoherent -- an "evolutionary unstable strategy," bound to undo itself: not only because it requires self-sacrificial double-payment locally as well as unrealistic collaboration among nations, institutions, funders, fields and publishers globally, but because the day after it was miraculously (and hypothetically) attained globally it would immediately invite defection (from nations, institutions, funders, and fields) to save money (invasion by the "cheater strategy"). Subscriptions and gold OA "memberships" are simply incommensurable, let alone transformable from one into the other. (Memberships are absurd, and only sell -- a bit, locally -- while subscriptions still prevail, via local Big Deals. 

The only evolutionarily stable strategy is offloading onto green OA repositories all but one of the things that publishers traditionally do, leaving only the service of peer review to be paid for as fair-gold OA. 

But that requires universal green OA first, not flipped pre-emptive fool's gold. 

It will all eventually sort itself out that way after a huge series of false-starts. My loss of patience is not just with the needless loss of time but with the boringly repetitious nature of the recurrent false starts. I'd say my last five years, at the very least, have been spent just repeating myself in the face of the very same naive bright-eyed, bushy-tailed and non-viable non-starters. Locally in space and time, some people sometimes listened to my objections and my alternative strategy, but globally the very same non-starters kept popping up, one after the other, independently.

So (with an occasional exception like this) I've stopped preaching. Time will either show that I was wrong or, like evolution, it will undo the maladaptive strategies and stumble blindly, but inevitably toward the stable strategy (which also happens to be the optimal one): universal green first, then a rapid downsizing and transition to scalable, affordable, sustainable fair-gold. Amen.

---

Part 2

1. Publisher green OA embargoes are ineffectual against the right green OA mandate: immediate deposit plus the almost-OA Button

2. That a “self-styled archivangelist” has left the arena is neither news nor an OA development. It is indeed just symbolic.

3. The fool's gold "flip" is an evolutionarily unstable strategy, fated to flop, despite the fond hopes RCUK, Wellcome, VSNU or MPG.

4. The "impact factor" is, as ever, utterly irrelevant to OA, one way or the other. Metrics will only be diversified and enriched by OA.

5. An immediate-deposit requirement is not an "onerous bureaucratic rule" but a few extra keystrokes per paper published: a no-brainer. Researchers are not "foot-soldiers" but finger-soldiers, and the immediate-deposit mandate is just intended to set those last few digits into motion (the publish-or-perish mandate having already mobilized the legions ahead of it).

6. Leaders are welcome (if not Wellcome), but boycotts are busts (and there have been plenty).

7. Exposés of publisher profiteering are welcome, but not solutions. In any case, the root problem is not affordability but accessibility, and providing access (via green OA) is also the solution, first to accessibility and then, as a natural matter of course, to affordability (post-green fair-gold).

8. Founding a new gold OA journal is hardly new. Offloading everything but peer review onto green repositories is also not new (in fact it will be part of the post-green end-game: fair-gold). But making it scalable and sustainable pre-emptiively would be new...

9. Subsidizing fair-gold costs would be fine, if someone had the resources to subsidize at least 30,000 journals across all disciplines. But while journals are being sustained by subscriptions, and there is no alternative way to access the contents, there is unlikely to be enough subsidy money to do the job. (Universally mandated green, in contrast, would allow journal subscriptions to be cancelled, releasing the money to pay for fair-gold out of just a fraction of the windfall savings.)

10. The impact factor, it cannot be repeated often enough, is absolutely irrelevant to (green) OA. The known track-record of journals, in contrast, will always be a factor.

11. Open "peer" review, or crowd-sourced quality control, likewise a notion aired many times, is, IMHO, likewise a non-starter. Suitable for peddling products and blog postings, but not for cancer cures and serious science or scholarship. (That said, anyone is everyone is already free to post their unrefereed work for all comers; that's what blogs and open commentary are for...)

12. Open online collaboration is very welcome (and more and more widespread) but it is a supplement, not a substitute, for publishing peer-reviewed findings.

13. Mathematics and, to a lesser extent, physics, are manifestly atypical fields in that their practitioners are (1) more willing than others to make their own unrefereed findings public and (2) eager to see and use the unrefereed findings of others. If this had been true of other fields, Arxiv would long ago have become the global unrefereed preprints and refereed postrprint repository for all fields, universal (central) green OA would already have been reached long ago, and the transition to fair gold would already have taken place. (Arxiv has been held up -- including, for a while, by me -- as the way to go since 1991. But things have not gone that way. That's why I switched to promoting distributed institutional repositories.)

14. What if the "P" in APCs -- for those who are "imPlacably opposed" to article processing charges -- stood instead for Peer-Review, and paid only for the editorial expertise in the refereeing (the peers review for free): selecting referees, selecting which referee recommendations need to be followed, selecting which revisions have done so and are hence accepted. These are the sole costs of fair gold -- but they are predicated on universal green to "overlay" on...

15. The two crucial features of peer review are expertise and answerability. This is what is provided by a qualified editor and established journal and absent in self-selected, crowd-sourced, take-it-or-leave-it vetting (already proposed many times, including by another distinguished mathematician). "Fair OA" is synonymous with fair gold, but universal green is the only viable way to get there.

16. Open peer commentary is a fine idea (if I do say so myself) but it is a supplement to peer review, not a substitute for it.

---

...And let's get our figures straight

Rick Anderson posted the following comment on Richard Poynder's posting in google+: “Institutional Green OA mandates (as distinct from non-mandatory OA policies) are effectively nonexistent in the US, and it's difficult to see how they could ever become widespread at the institutional level. That's just the US, of course, but the US produces an awful lot of research publication.”

According to ROARMAP, which was recently upgraded to expand, classify and verify the entries, although it is probably not yet exhaustive (some mandates may not yet be registered) there are 764 OA policies worlwide, at least 629 of them Green (i.e., they either or request deposit)

The following are the total(subset) figures broken down by country for
total policies and the subset requiring - not requesting - deposit
for Institional and Funder policies.

Worldwide:
Inst 632 (390)
Fund 132 (82)

US
Inst 96 (69)
Fund 34 (11)

UK
Inst: 93 (79)
Fund 24 (23)

Germany
Inst 26 (2)
Fund 1 (0)

Netherlands
Inst 11 (6)
Fund 0

France
Inst 17 (3)
Fund 3 (3)

Canada
Inst 15 (7)
Fund 12 (9)

Australia
Inst 31 (15)
Fund 2 (2)

---

Rick Anderson: Happy to provide examples.

Duke University -- ROARMAP says deposit is required and that the policy does not specify whether or not a waiver is available. According to Duke's policy, "The Provost or Provost's designate will waive application of the license for a particular article... upon written request by a Faculty member."

Harvard (Faculty of Arts & Sciences) -- ROARMAP says that deposit is required and may not be waived. The Harvard policy says that the institution "will waive application of the policy for a particular article upon written request by a Faculty member."

Oberlin College -- ROARMAP says that deposit is required. Oberlin's policy says that while the submission of bibliographic metadata is required, "the Scholarly Communications Officer will automatically waive application of the policy for a particular article upon written request by the author."

Bucknell University -- ROARMAP says that deposit is required and that the possibility of a waiver is "unspecified." The policy says that "the license granted to Bucknell University for an article will be automaticaly waived for any reason and without sanction at the sole discretion of the faculty member." (Deposit of bibliographic data is requested but not required.)

That's probably enough for now. I can provide more upon request. In fact, if you'd like, I'd be happy to examine all 96 of the US institutional policies identified in ROARMAP, analyze the accuracy of the ROARMAP characterization of each one, and publish what I find (whatever the result might be). Would you like me to do that?

---

Stevan Harnad: The Harvard FAS OA Policy model (which may or may not have been adopted by the other institutions you cite without their fully understanding its conditions) is that:

(1) Full-text deposit is required
but
(2) Rights-retention (and OA) may be waived on an individual article basis

The deposit requirement (1) cannot be waived, and is not waived if the author elects to waive (2).

This is the policy that Peter calls "dual deposit/release" (and I call immediate-deposit/optional-access, ID/OA):
http://users.ecs.soton.ac.uk/harnad/Hypermail/Amsci/7841.html
(soton site temporarily down today, apologies)

---

Rick Anderson: Stevan, your characterisation of the Harvard policy seems to me to be simply inaccurate. The full text may be read at https://osc.hul.harvard.edu/policies/fas/. The relevant sentence reads as follows: "The Dean or the Dean's designate will waive application of the policy for a particular article upon written request by a Faculty member explaining the need." This language seems pretty clearly to me to refer to the policy as a whole, not just one component of it -- nor does the policy itself include an OA requirement; instead it provides the possibility that "the Provost's Office may make the article available to the public in an open-access repository" (note the word "may," not usually a prominent feature in mandatory instructions).

For the avoidance of doubt, I am contacting Harvard's Office for Scholarly Communication to either confirm or correct my reading of its policy. I'll let you know what I find out.

---

Peter Suber: Stevan's restatement of the Harvard policy is correct. Our waiver option only applies to the license, not to the deposit.

---

Rick Anderson: OK, thanks for clearing that up, Peter. (You guys might want to consider revising the wording of your policy to resolve the ambiguity.)

Stevan, do you disagree that ROARMAP's characterisations of the other policies I cited above are inaccurate?

---

Stevan Harnad: The other three policies you cited seem to have adopted the Harvard model policy. If they have diverged from it, they need to indicate that explicitly (and unambiguously). ROARMAP incorporates updates of corrections when it receives them. The ones you mention were either registered by the institutions themselves or derived from their documentation and sent to them for vetting.

I cannot vouch for 100% compliance or accuracy. But your assertion was not about that. Your assertion was “Institutional Green OA mandates (as distinct from non-mandatory OA policies) are effectively nonexistent in the US."

Do you think your four examples show that? One out of the four, Harvard FAS, would already disconfirm "nonexistent" ("effectively" being a weasel-word) even without the added fact that Harvard is not just any university, and the one whose model many US universities have adopted. And even if you could show (as you certainly have not done) that not one of the remaining 65 US institutional mandates (out of the total 96 US institutional OA policies in ROARMAP) was a mandate. Do you disagree?

---

Rick Anderson: All of the examples I provided (including the Harvard example) constitute evidence in support of my statement, since they are instances in which Green OA is not mandatory. They don't constitute the entire evidence base. I made my statement based on the fact that I have read many OA policies from US institutions, and I have not yet encountered (nor heard of) a single one that requires faculty to make their work available on an OA basis. A policy that requires deposit but does not require OA is not a mandatory OA policy.

If my assertion is wrong, it should be easy to refute. Can you cite a significant number of examples of US institutions that do require faculty to make their work available on an OA basis (as distinct from merely requiring deposit)?

---

Stevan Harnad: I would like to avoid empty semiological quibbling. The US has 96 institutional OA policies. That is uncontested. Of these, 69 are registered as deposit mandates, hence mandates.

There are many other conditions (such as whether and when it is mandatory to make the deposit OA), but it may be helpful to understand that the reason mandatory (full-text) deposit is the crucial requirement is that if (and only if) the full-text is deposited, the repository's automated copy-request Button (if and when implemented) can provide almost-immediate, almost-OA to any user who clicks it (if the author too chooses to comply, with a click).

The hypothesis (and it is indeed a hypothesis, not a certainty) is that this compromise mandate (DD/R, ID/OA), if universally adopted, will not only provide almost 100% Green OA, but will prove sufficient to eventually make subscriptions cancellable, thereby inducing journal publisher downsizing, the phasing out of obsolete products and services, and a transition to affordable, scalable and sustainable Fair-Gold OA, charging for peer-review alone, and paid for out of a fraction of the institutional subscription cancellation savings, instead of the over-priced, double-paid, and unnecessary Fool's Gold that is on offer now, paid for out of already over-stretched subscription as well as research funds.