From: "Jean-Claude Guédon" <[log in to unmask]>
Date: Tue, 28 Mar 2017 10:30:06 -0400

Ann Okerson's question is important and actually reaches well beyond
its ostensible target. The development of digital publishing leads to
a deep transformation in the nature of documents.

Software writing was the first area where this issue arose.
Interestingly, and probably because the field of software writing was
unencumbered by previous, legacy habits, programmers developed
techniques that could track versions in a flexible and agile manner.
We are now all used to grappling with version 7.3 (or whatever) of
software X, and think nothing of it. "Requests for comments" (RFC), as
developed and used by the IETF in the context of the technical
developments of the Internet, followed a similar philosophy, while
applying it to straight texts, not software.

Print forced us, a long time ago, to fix and harden what was being
published. It also foregrounded a vision of documents that depended on
techniques now embedded in critical editions: each critical edition
stands for eternity until a new critical edition challenges it, and
the result is a staccato mode of evolution that worked well with
print, but appears increasingly out of step with the potentials of
digital documents (especially with the inclusion of data sets and
software). Scientific and scholarly articles as they evolved through
print (and survive as PDF files, which is about as close to print as a
digital file can be - true digital incunabula, to use G. Crane's
clever image) also force the staccato mode of conversation and debate
that still dominates in scientific and scholarly circles. This leads
to a very inefficient way to feed the "Great Conversation" of science.

With digital documents, what we need is an orderly system of versions
similar to what is used in software. In so doing, we no longer have to
grapple with any version "of record"; instead, we have to deal with a
time-driven succession of documents that can be easily identified if
the version system is well designed. Normally, the penultimate version
should be the most reliable, and the latest should be the cutting edge
solution that is on the chopping block for the next round of
Popperian-style refutations. Sometimes, the latest version adds little
to the previous one, but corrects minor problems; sometimes, it is a
real advance; sometimes it is a huge step forward. The numbering
system can reflect all of this, while keeping the history of a certain
thread of thought. Contributions to any document, however small or
large, so long as it is accepted, can be attributed to various
individuals. Free software shows how such an approach can be extended
to a broadly distributed system of contributors (which science and
scholarship, in general terms, are).

If academic libraries, with their repositories, begin collaboratively
(and in a distributed  manner) to develop a peer-review system -
somewhat like the F1000 approach to reviewing - then, the issue of
filtering unwanted materials and noise also begins to find a solution.

Metadata can incorporate versions. Therefore, retrievability ,
discoverability, as well as visibility issues can be addressed as
well. In fact, unpaywall, which is a wonderful tool by the way, could
be tweaked to take charge of the version issue, if a good version
system can be developed. OpenAIRE could be a good place to develop
such a system.

Put all of this within the context of platforms, and not of journals,
and the digital communication system of science begins to take a
meaningful shape.

Jean-Claude Guidon


Le mardi 28 mars 2017 à 00:50 -0400, LIBLICENSE a écrit :
From: Anali Perry <[log in to unmask]>
Date: Tue, 28 Mar 2017 01:03:34 +0000

Hi Ann,

I'm sorry I wasn't more clear in my earlier response. Unpaywall only
searches institutional repositories and publisher sites, so it's about
as legal as one can get - assuming all articles on publisher's sites
are legal, as well as whatever posted on a repository being vetted by
repository management. Regarding which version, if it is on a
publisher's site, it will be the final published version. If it is on
a repository, the default assumption is likely author's final
manuscript, but most repositories have some indicator about which
version was archived - there are some cases where final published
version is allowed, for example, and the repository description would
probably reflect that.

Hope that helps,

Anali Maughan Perry
Associate Librarian - Collections & Scholarly Communication
Arizona State University Libraries
(480) 727-6301
[log in to unmask]
http://libguides.asu.edu/profile/amperry
http://orcid.org/0000-0001-7173-4827

My pronouns are she/her/hers.


-----Original Message-----
From: Ann Shumelda Okerson <[log in to unmask]>
Date: Wed, 22 Mar 2017 14:16:48 -0400

Hi, Anali, that still doesn't answer my question of "how does one know
which version?"  "OA version" isn't exactly an answer, or at least not
what I was wondering.  Assuming that what is pointed to is definitely
legal, what I meant is, how does one know if this is a final final
published version, an e-print version, an author's manuscript version,
etc.  Cheers, Ann Okerson

On Wed, Mar 22, 2017 at 4:52 AM, LIBLICENSE <[log in to unmask]> wrote:

From: Anali Perry <[log in to unmask]>
Date: Tue, 21 Mar 2017 15:41:08 +0000
Subject: RE: ImpactStory releases new Chrome Extension: Unpaywall Hi
Brian,

Google Scholar does not reliably index all institutional repositories
(there have been a few studies documenting this - there are issues on
both the repository & Google sides). Unpaywall specifically indexes
known OA locations, so it would be more reliable for finding OA
content. From the FAQ:
“We rely on some fantastic open data services, especially PubMed
Central, the DOAJ, Crossref (particularly their license info),
DataCite, and BASE.”

To answer Ann’s earlier question – Unpaywall searches both publishers’
sites as well as repositories. If an OA version of the article is
available via the publisher, then it will point to that. If a
postprint is available on a repository as the only OA option that
should be what it pulls up. There’s an option in the settings to have
Unpaywall signal whether the article is Gold or Green.

Anali Maughan Perry
Associate Librarian - Collections & Scholarly Communication ASU
Library
(480) 727-6301
[log in to unmask]