[Open-access] @ccess versus open access

[Stian Håklev]

Hi everyone,

I've been following this discussion for a while, and I wanted to jump in.
First I wanted to say that I _really_ appreciate your hard work and
advocacy for making Open Access more "free and open". While it is very
encouraging to see the many advances that the OA movement has made, it's
been discouraging to see how little understanding of licenses, file
formats, metadata etc in the Open Access formats. It's great if everyone in
the world are able to download PFDs from a website - but we can do _so much
more_!

What seems to come out of some of these discussions, however, and what I
feel from my own experimentation and thinking, is that there is a whole
array of different "levels of openness" or "attributes of openness" that we
would be looking for. Personally some of the things I would love to see
happen are:

* using an open license, ideally CC BY
* making the machine-readable metadata easily available for scraping on the
website where the abstract is hosted
* embedding the license and download URL in the machine readable metadata
(I've almost never seen a BibTeX citation for an OA journal that has a URL
directly to the downloadable file, or that has any kind of markup
indicating that it is OA, license etc)
* embedding the machine readable metadata in the downloadable file itself
(for example using the XMPP field in PDFs, like JabRef does)
* having the journal (or repository/preprint archive) automatically ping
some public server with machine readable metadata whenever a new
publication is published
* for the journal/repository/preprint archive to offer standardized APIs
for accessing metadata etc, and also ability to download a dump of
machine-readable metadata for every single article, as well as a dump of
all the actual article files (PDF or otherwise)
* moving from PDFs to some kind of standardized semantic file format, where
for example all in-article citations also have machine-readable metadata,
use ORCID tags for authors, DOIs or other unique identifiers for cited
articles etc
* making all data relevant to the article openly licensed (CC-0) and linked
from the article, same with source code
* moving towards "executable" papers, where all the graphs and stats and
calculations are generated dynamically using embedded code (for example R
with Knitr) and the linked open data
* etc

I am sure people could even add to this list. Now this is quite a
"maximalist" list, and I don't expect there to be almost a single
publication today that fulfills all or even most of these expectations, yet
I really hope we can work towards a future where much of this is possible -
which requires work on a multitude of levels, both better software support,
better infrastructure and standards, more conscientization of authors,
policy work with institutions and publishers, changes in funding, etc...

The question seems to be, which of these "demands" / "ideals" are "crucial"
and need to be bundled together in a statement about @access or BBB-OA or
whatever it is called? And what is the value of having one such
"milestone", as opposed to maybe a spectrum of "opennesses"? If you aim too
low, saying "We just require publications to be CC-BY licensed", that would
of course still be a huge improvement from today - but without much of the
other things mentioned, it wouldn't  necessarily have a huge practical
significance... (Hard to modify locked down PDFs...). And if you maximalize
the demand ("You're only open if you fulfill the following 20
requirements"), the movement might be seen as impractical and not attract
much interest.

So I'm not really sure what's the best way forward, tactically, to achieve
our long term goals. But I really commend you again on at least pushing to
expand the notion of Open Access, and I'll be excited to discuss how to
reach / or even begin experimenting with / many of the points I listed
above.

Much of my thinking on this comes from building my own "research workflow" (
http://reganmian.net/wiki/researchr:start), which integrates a bunch of
tools both to speed-up my own harvesting of citations, PDFs, note taking
etc, but also enables me to share as much as possible of this with others,
as an "open scholar". Here's a recent blog entry about how I try to
determine whether a publication I import is "open access" (in my case,
simply defined as "PDF is directly downloadable from a public URL"), and
how I can then add a bunch of functionality for collaboration based on this
fact. To my mind, this greater ease of collaboration and sharing - EVEN for
people at institutions which pay for TA journals - is a powerful argument
for OA which we haven't exploited enough:
http://reganmian.net/blog/2012/04/17/api-to-check-if-a-publication-is-open-access/

Stian
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.okfn.org/pipermail/open-access/attachments/20120601/3cb891a3/attachment.html>