Re: When is a Journal Open Access?

From: peter murray-rust <pm286_at_CAM.AC.UK>
Date: Sat, 2 Dec 2006 16:13:30 +0000

Thanks to all,
As you will see there is substantial agreement between Stevan and
me. I think some of the agreement raises further practical
difficulties but these are probably outside the scope of the list. So
far there have only been 4 voices (SH, Les Carr, Sally Morris, and
PM-R). As this list has a large informed membership of OA experts who
are not shy I shall assume that silence means general agreement with
what follows. I am primarily searching for a definition (or
definitions) of OA that are agreed and which can be used to shape
(further) action. Some of this is old ground to many, but what is
clear is that OA is a complex subject, at least in so far as there
have been many different topics discussed. I make it clear that I
have no intention of trying to change or widen the definition of OA -
simply to understand it.

First to thank Sally for sending me her (Closed Access) paper. My
concern had been about the use of licenses by "OA journals" in the
DOAJ and whether they were "BOAI-compliant". The article didn't seem
to cover that (certainly not to give figures on these topics and was
more about the viability, vigour and longevity of such journals.

I shall reply briefly to a few points.

At 14:07 02/12/2006, Stevan Harnad wrote:
>This is an online reply to Peter Murray-Rust, followed by an online exchange,
>added with permission.
>
>On Sat, 2 Dec 2006, peter murray-rust wrote:
>
> > Corollaries:
> > 1. It is an irrelevance or nonsense to talk of OA journals.
>
>OA Journals (Gold) are of course relevant to OA and the growth of OA (and
>perhaps eventually to the future of journal publishing). But for now, they
>have far less potential for immediate OA growth than OA self-archiving
>(Green) does.
>
>The definition of OA is simple and functional (free for all online),
>and hence so is the definition of an OA journal-article. The definition
>of an OA journal, in contrast, is somewhat arbitrary. If it were put to
>a vote, I would vote that an OA journal should be defined as a journal
>all of whose articles are made into OA journal-articles by the journal
>itself. I am sure Murray [Murray-Rust?] would vote for this too. But
>alas there is no
>consensus on this.

Understood. Effectively "OA journal" is not an operationally useful term.


> > 2. Whatever contractual conditions are attached to an "OA article"
> > are meaningless or irrelevant, or overridden by the fact of the
> > article having been posted openly (which has achieved by being OA)..
>
>Yes; once an *author* makes his own journal article OA by self-archiving it,
>that trumps all other contractual conditions, insofar as OA is concerned:
>The *user* has free online access to the full-text, along with all
>the other capabilities that come with that online territory. But that
>excludes offline capabilities, such as re-publishing and distributing it
>on paper, and even online mass-re-use capabilities, such as incorporating
>the contents (rather than links) in a commercial (or noncommercial)
>database without permission.
>
>Before self-archiving, however, some *authors* may feel contractually
>constrained about making their own journal articles OA, if the journal has
>not yet explicitly endorsed the practice of self-archiving (as 69% of
>journals have already done for the postprint and a further 25% for the
>preprint).


>For this 31% of articles/authors, the immediate solution
>is to deposit the full-text of the postprint in their OAI-compliant
>Institutional Repository immediately upon acceptance for publication,
>exactly as the other 69% do, but to set access provisionally as Closed
>Access (CA) rather than Open Access (OA). Only the metadata will then be
>accessible webwide, but that will be sufficient to allow all individual
>user-needs to be fulfilled with the help of the IR's EMAIL EPRINT button
>(allowing users to request a single email copy semi-automatically via
>the IR's web interface and email).

This is extremely useful. In fact I spent time discussing this very
point with my group yesterday. We agreed that we would deposit the
final version of all manuscripts before publication in our
archive/repository. These would then be made public at the
appropriate stage (probably final acceptance). We note, however, that
this is the pre-review manuscript in most of our discipline.

It was interesting to note that at the recent Digital Curation
Conference the reasons for loss of pre-review manuscripts (quite a
high percentage) were reviewed (I'll expand on this in my blog).
These included (a) crash of machine (b) forgetting where it was (c)
another author took responsibility (d) author moved institution (e)
another institution managed the submission and so on. All of these
have happened to me. So from a purely utilitarian point of view
self-archiving will be very valuable to us.

>CA is definitely not OA, however, and does not provide what Peter
>Murray-Rust needs in order to data-mine the full-texts via robot-trawlers
>and data-crunchers! That will have to wait a bit for this subset of the
>articles (31%), making do for now with the rest (69%). But once 100%
>of authors are depositing immediately (as the mandates will ensure they
>do), the day when the last 31% are made OA too will not be far off. (Just
>let human nature, and the benefits of OA, take their natural course)...

I have a lot of empathy with this last sentence. OA may, indeed,
carry the world and the rest comes. It may, however, provide a useful
but not complete part of the solution that the world is too weary to
carry through completely. It is the second prognosis I am actively
trying to prevent.


> > 3. The various declarations of OA are overly complex and should be
> > replaced by the first SH sentence above and commonsense.
>
>Yes!
>
> > > SH:
> > > "distractions that have been slowing our progress in reaching the
> > > optimal and inevitable. (I have no theory of why we tarry so) "
> >
> > I apologize for the distraction - it came from an over-literal
> > interpretation of the BOAI.
>
>It is not at all a distraction to point out that the definition of
>"OA Journal" is problematic. The risk of distraction came from the
>implication that explicit robot-trawling licenses (or anything else) are a
>prerequisite for making a journal-article OA (via author self-archiving),
>rather than that they are merely a prerequisite for ensuring that all
>OA journals make all of their articles robot-trawl-able.
>
>(My guess, though, is that all OA journals [Gold] will also be Green
>on author self-archiving; so the blanket remedy is that all authors
>should self-archive all their articles, whether or not they happen to
>be published in an OA journal!)
>
>Here is a further offline exchange, reproduced with permission. Peter's
>responses follow too. I have no further reply. I agree with Peter's summary.

Thank you.

My intentions are to summarise what has been written and pick up any
additional comments on this list (or other well-known forums( and try
to write a simple guide for scientists, especially in data-rich
subjects. It will, necessarily, highlight the fact that OA only
provides part of the solution, but it should be possible to decide
exactly what part it does support.

Finally I shall be retrospectively active in self-archiving of Closed
access articles. To be fair it has not been easy to self-archive CA
up till now as the escrow facility was not originally in place. It
mean that you had to keep the final pre-review manuscript somewhere
and then, at the time the paper appeared, self-archive. Since in some
cases not only did I not know when the article appeared but was
actually forbidden to read it (as the University had no online
subscription) it was sufficiently tedious that it didn't get done.

P.


>On Sat, 2 Dec 2006, Leslie Carr wrote:
> > > I believe that Peter's enquiry comes from the fact that he does a lot
> > > of data-mining of chemistry journal articles, trying to pick up
> > > references in the text to particular kinds of chemical. I think that
> > > his concern is whether he is allowed to do this in any given "so-
> > > called" OA Journal.
>
>SH:
>I could more or less infer that, but what I am trying to bring out is that
>this is, in an important way, putting the shoe rather on the wrong foot:
>
>The basic OA question is not whether a *journal* is OA (and hence
>its contents can be harvested and data-mined) but whether an *article*
>is OA (and hence its contents can be harvested and data-mined).
>
>The definition of an OA journal is obviously rather arbitrary (though
>we would all obviously prefer that it meant that 100% of its articles
>are OA).
>
>But in expressing his very valid and justified frustration with
>self-styled "OA journals" not all of whose articles are in fact OA,
>Peter inadvertently conflates (as many others inadvertently do) the
>"definition" of OA with the definition of "OA Journal." Those are the
>two things that have to be very explicitly disentangled.
>
>Otherwise Peter, in his valuable and important efforts on behalf of the
>data-mining of OA journal-article content, inadvertently joins the ranks
>of those who think licensing certain rights is one of the preconditions
>for making a journal-article OA: It is not, and if it were, it would
>greatly retard the progress of OA (via self-archiving).
>
>(I may be wrong that Peter is conflating journal-article data-mining with
>empirical-data data-mining, but then he is conflating the requirements for
>being an OA journal-article with the requirements for being an OA journal.)
>
> > Date: Sat, 02 Dec 2006 13:00:33 +0000
> > From: peter murray-rust
> >
> > Thanks both,
> >
> > I am seriously trying to reach points of agreement.
> >
> > The following may be helpful:
> > - I come from disciplines where formal definitions often relate to
> > *procedures* rather than *principles*
> > - there are very many people who take a roughly similar approach to
> > me. That is a fact - perhaps regrettable - but it will help to
> acknowledge it.
> > - I suspect that some scientific funding bodies may see OA as
> > containing operational components (other than the process of free
> > online indefinitely)
> > - If we can clarify OA - its extent , its use - then we can help to
> > define those principles and procedures which lie outside OA but
> > which, for certain communities, are necessary. I have used the term
> > "Open Data", which I separate from OA.
> > - It is possible that in some cases the precise practice of OA leads
> > to OD and that in some cases does not. This may not be a problem
> > for OA (and if it is separated is not), but it is a problem for some
> > of the disciplines.
> >
> > The BOAI reads partly like (or can be read by some as) a political or
> > philosophical statement (e.g. a constitution) and partly as an
> > operational procedure. This has been, I suspect, the root of confusion.
> >
> > > > LC:
> > > > I believe that Peter's enquiry comes from the fact that he does a lot
> > > > of data-mining of chemistry journal articles, trying to pick up
> > > > references in the text to particular kinds of chemical. I think that
> > > > his concern is whether he is allowed to do this in any given "so-
> > > > called" OA Journal.
> >
> > Yes
> >
> > > SH:
> > > I could more or less infer that, but what I am trying to bring
> out is that
> > > this is, in an important way, putting the shoe rather on the wrong foot:
> >
> > I believe I now understand Stevan's position on this.
> >
> > > SH:
> > > The basic OA question is not whether a *journal* is OA (and hence
> > > its contents can be harvested and data-mined) but whether an *article*
> > > is OA (and hence its contents can be harvested and data-mined).
> > >
> > > The definition of an OA journal is obviously rather arbitrary (though
> > > we would all obviously prefer that it meant that 100% of its articles
> > > are OA).
> >
> > An "OA journal" is an operational definition. It is easily (and
> > perhaps mistakenly) taken to mean "all articles in this journal
> > inherit (in the computer sense) the following principles...
> > (statement of purpose and licenses).
> >
> > > SH:
> > > But in expressing his very valid and justified frustration with
> > > self-styled "OA journals" not all of whose articles are in fact OA,
> > > Peter inadvertently conflates (as many others inadvertently do) the
> > > "definition" of OA with the definition of "OA Journal." Those are the
> > > two things that have to be very explicitly disentangled.
> >
> > I think the term "OA journal" conflates it, not me.
> >
> > >Otherwise Peter, in his valuable and important efforts on behalf of the
> > >data-mining of OA journal-article content, inadvertently joins the ranks
> > >of those who think licensing certain rights is one of the preconditions
> > >for making a journal-article OA: It is not, and it it were, it would
> > >greatly retard the progress of OA (via self-archiving).
> >
> > I now understand this. The BOAI and (if I am correct) my
> > interpretation of SH's definitions are statements of principle, not
> > practice. They rely on an (undefined) body of quasi-legislative
> > practice for their operation. Stevan believes that no such body is
> > required, unfortunately others do.
> >
> > > SH:
> > > (I may be wrong that Peter is conflating journal-article data-mining with
> > > empirical-data data-mining, but then he is conflating the
> requirements for
> > > being an OA journal-article with the requirements for being an
> OA journal.)
> >
> > Not really. The journal concept is a convenience but not a necessity.
> > I can mine single articles (which operation I take to depend on
> > license, not openness of availability). The journal confers the
> > following advantages:
> > * a single statement by a journal can cover a very large number of articles
> > * a number of bodies (e.g. DOAJ) can potentially help us to make
> > decisions effectively and cheaply
> > * (navigation) it is often convenient, though not necessary, to use
> > the structure of a journal (or repository) to locate articles.
> >
> > If there is substantial agreement within the OA community on this
> > issue then I have a clear idea of what phrases I can use and in
> what context.
> >
> > Please feel free to re-use anything here.
> >
> > Peter Murray-Rust
> > Unilever Centre for Molecular Sciences Informatics
> > University of Cambridge,
> > Lensfield Road, Cambridge CB2 1EW, UK
> > +44-1223-763069

Peter Murray-Rust
Unilever Centre for Molecular Sciences Informatics
University of Cambridge,
Lensfield Road, Cambridge CB2 1EW, UK
+44-1223-763069
Received on Sat Dec 02 2006 - 16:58:49 GMT