[open-bibliography] Fwd: FAST as "open" data

[Adrian Pohl]

Below Eric Childress' response to my answer about the licensing of the
FAST data that he kindly provided quickly and with much detail.

Adrian

---------- Forwarded message ----------
From: Childress,Eric <childree at oclc.org>
Date: 5 January 2012 22:00
Subject: RE: FAST as "open" data
To: Adrian Pohl <adrian.pohl at okfn.org>
Cc: "Bolander,Robert" <bolander at oclc.org>

Dear Adrian,

Thank you for the query with respect to some confusion that surfaced
in recent exchanges on the open-bibliography list about OCLC’s release
of FAST as linked data.  I was surprised to see a private email
response from OCLC distributed and mis-interpreted on the
open-bibliography list.  It would have saved some mis-understanding if
the parties involved had simply queried us for clarification if our
response to their original query did not seem to make sense. Indeed,
it should have been particularly simple given that my email address
was embedded in the excerpted text ;)

And thank you for asking for clarification.

The confusion seems to have surfaced in part due to  the
unavailability on the FAST landing pages of a link to download the
entire dataset. That the linked data release was available under an
Open Data Commons Attribution license led to the expectation that the
full dump would be made available under those same conditions. Not a
surprizing expectation, but in the interest of moving FAST out as
linked data sooner rather than later, we made FAST available as
decomposed linked data in mid-December and planned to address various
further-work-we-want-to-do in the New Year including addressing
suggestions for enhancements, making a dump available, etc.

Earlier this week a request for a download of the full dataset was
received by one of the system engineers in OCLC Research who responded
that complete dumps were at present only made available under a
research-use license and referred the user to me in OCLC Research as I
field external researcher requests. The system engineer actually
responded correctly as that had been OCLC’s policy up until the FAST
linked data release as access to the full database has required that
researchers and other parties request the FAST file through OCLC’s
external research support request process. Full files of FAST have
been made available for many years to a number of parties under this
process and are provided under our standard research-use license. And
it was *this standing approach* to offering FAST dumps that was
referenced in the email correspondence excerpted in Karen Coyle's post
to the open-bibliography list.  It was what we could offer immediately
to anyone wanting a dump. The intent was not to confuse, but to be of
service.

So to point: Karen Coyle’s email
[http://lists.okfn.org/pipermail/open-bibliography/2012-January/001272.html
] contains a statement, "So, the file is NOT available under ODC-By"
which is simply incorrect (FAST in decomposed form was released by
OCLC under an ODC-BY license; and there really is no ambiguity. The
data is published under a license; one need only read the license to
understand what term & conditions apply.). And there is no
contradiction – as I’ve explained the excerpt was referencing a
separate and distinct channel for gaining access to FAST.

All said, now that the holiday break is over, OCLC will be making FAST
available as a dump on our web site under the ODC-BY license in the
near future. I’d suggest it might be worth waiting for that rather
than crawling the data through the API. Perhaps in retrospect it would
have been more useful to have simply informed the person corresponding
with us that the dump was “coming soon,” but we were trying to be
helpful with a what's-available-now solution.

Ultimately, our interest is in furthering library linked data
experimentation. The full FAST dataset may be useful in some other
circumstances, but it doesn’t do as much to advance experimentation in
this area as the linked data service. To update those who are
interested OCLC will be glad to send an email to the listserv when we
have the full dataset available, including the URL.

OCLC appreciates the strong interest in FAST as linked data and is
pleased that its availability will support further work in the library
linked data community.

On a personal note, thank you for the invitation to prepare a blog
entry on openbiblio.net blog. If I may, I would like to consult
colleagues about this kind offer and get back to you.

I wish you a very Happy New Years.

Regards,

Eric


Eric Childress
Consulting Project Manager
OCLC Research
6565 Kilgour Place
Dublin, OH 43017 USA

-----Original Message-----
From: ad.pohl at googlemail.com [mailto:ad.pohl at googlemail.com] On Behalf
Of Adrian Pohl
Sent: Wednesday, January 04, 2012 4:16 PM
To: Childress,Eric; Childress,Eric; Bolander,Robert
Subject: FAST as "open" data

Dear Eric, dear Robert,

the announcement of OCLC releasing the FAST schema as Linked Open Data

was taken up in the open bibliographic data community with much

interest. But today's mail by Karen Coyle to the open bibliography

mailing list[1] was rather irritating. I quote from this mail:

On 4 January 2012 21:36, Karen Coyle <kcoyle at kcoyle.net> wrote:

> Someone I know wrote to OCLC to get a copy of FAST and received this reply:

>

> "Currently, complete dumps of the FAST file are not available under an

> ODC-By license.  OCLC does make dumps available on a case-by-case basis

> under a research-use license.  To apply, please contact Eric Childress

> (childree at oclc.org)."

This obvious looks like self-contradictary behaviour: On the one hand

licensing a dataset under ODC-BY which means one can re-use the whole

database or significant parts of it subject only ot the requirement of

attribution and on the other hand telling people they get the whole

dataset only under a "research-use" license.

I would be very happy if OCLC clarified its position and I invite you

to do this in a guest post on the openbiblio.net blog.

I am looking forward to your answer.

All the best

Adrian

[1] http://lists.okfn.org/pipermail/open-bibliography/2012-January/001272.html