On 4/30/14, 9:19 PM, Chad Nelson wrote:
If libraries aren't willing to put in the the effort to make their own data
more useful and connected, then I don't think they are going do much of
anything useful very with "linked data cake" served on a silver platter.
Are you really suggesting that we cede linked data creation, management and
curation to vendors.
Gee, that's pretty sarcastic. No, I am suggesting that there is a needed
service to help folks with textual data take that first step: adding the
identifiers for those strings, like adding $0 fields to their MARC
records. Perhaps you weren't around for the previous transitions, but
such services jump-started both the conversion of cards to MARC and AACR
to AACR2. You may not be aware but OCLC and other vendors provide
conversion services of this nature on a continuing basis. It's much more
efficient than having every library do the same coding for themselves.
Oh, and remember that we share cataloging through copy cataloging
services. There are lots of things that it just doesn't make sense to
"do it yourself."
kc
Chad
On Apr 30, 2014 10:28 PM, "Karen Coyle" <[email protected]> wrote:
On 4/30/14, 6:37 PM, Roy Tennant wrote:
In the end there may need to be reconciliation services just like we had
similar services in the card-catalog-to-digital years.
Roy
Roy, yes, that's what I'm assuming. I think we are indeed in the same
leaky boat we were in in the 1970's when all of a sudden we realized that
in the future we wanted our data to be "digital" but most of what we had
was definitely analog. In the early days, we thought it was an impossible
task to convert our cards to MARC, but it turned out to be possible.
I believe that linking our heading strings (the ones that hopefully
resemble the "prefLabel" on someone's authority file) to identifiers is not
as hard as people assume, especially if we have systems that can learn --
that is, that can build up cases of synonyms (e.g. "Smith, John" with title
"Here's my book" == "Smith, John J." with title "Here's my book"). This is
what the AACR->AACR2 services did. OCLC surely does a lot of this when
merging manifestations, and undoubtedly did so when determining what are
works, and when bringing authority entries together for VIAF. No, you don't
get 100% perfection, but we don't get that now with any of our services.
And for all of those who keep suggesting Open Refine -- it's like you
walk into bakery to buy a cake and they hand you flour, eggs, milk and show
you where the oven is. Yes, it can be done. But you want the cake -- if you
could do and wanted to *make* a cake you wouldn't be in the bakery, you'd
be home in your kitchen. So in case it isn't clear, I'm talking cake, not
cake making. How are we going to provide cake to the library and archives
masses? And, if you are feeling entrepreneurial, wouldn't this be a good
time to open a bakery?
kc
On Wed, Apr 30, 2014 at 5:53 PM, Karen Coyle <[email protected]> wrote:
Roy, the question that I have is, as I say below, about DISCOVERABILITY
of
URIs, not intellectual property issues. It's great that there are lots
of
URIs for useful things out in the world, but they don't jump into your
data
store on their own through some kind of magic. To me, the big problem
today
is that of populating legacy data with useful identifiers. I know that
some
folks have worked at making connections between subject headings in
their
catalog and the URIs available through id.loc.gov - and as I recall, it
turns out to be fairly frustrating. It seems to be that the solution to
this is that providers of URIs and users of URIs have to both make an
effort to meet half-way, or at a mutally convenient location. It simply
is
not enough to say: "Hey, look! I've got all of these URIs. Good luck!"
So
let's talk about how we make that connection.
kc
On 4/30/14, 1:17 PM, Roy Tennant wrote:
Also, this:
"OCLC identifiers, and Linked Data URIs, are always in the public
domain.
Independent of the data and/or information content (which may be
subject
to
individual licensing terms open or otherwise) that they identify, or
link
to, OCLC identifiers (e.g. OCLC Numbers, VIAF IDs, or WorldCat Work
URIs)
can be treated as if they are in the public domain and can be included
in
any data exposure mechanism or activity as public domain data."
http://www.oclc.org/developer/develop/linked-data.en.html
Roy
On Wed, Apr 30, 2014 at 9:59 AM, Richard Wallis <
[email protected]> wrote:
To unpack the several questions lurking in Karen’s question.
As to being able to use the WorldCat Works data/identifiers there is
no
difference between a or b - it is ODC-BY licensed data.
Getting a Work URI may be easier for a) as they should be able to
identify
the OCLC Number and hence use the linked data from it’s URI <
http://worldcat.org/oclc/{ocn}> to pick up the link to it’s work.
Tools such as xISBN <http://xisbn.worldcat.org/xisbnadmin/doc/api.htm>
can
step you towards identifier lookups and are openly available for low
volume
usage.
Citation lookup is more a bib lookup feature, that you could get an
OCLC
Number from. One of colleagues may be helpful on the particulars of
this.
Apologies for being WorldCat specific, but Karen did ask.
~Richard.
On 30 April 2014 17:15, Karen Coyle <[email protected]> wrote:
My question has to do with discoverability. Let's say that I have a
bibliographic database and I want to add the OCLC work identifiers to
it.
Obviously I don't want to do it by hand. I might have ISBNs, but in
some
cases I will have a regular author/title-type citation.
and let's say that I am asking this for two different institutions:
a) is an OCLC member institution
b) is not
Thanks,
kc
On 4/30/14, 8:47 AM, Dan Scott wrote:
On Tue, Apr 29, 2014 at 11:37 PM, Roy Tennant <[email protected]
wrote:
This has now instead become a reasonable recommendation
concerning ODC-BY licensing [3] but the confusion and uncertainty
about which records an OCLC member may redistribute remains.
[3] http://www.oclc.org/news/releases/2012/201248.en.html
Allow me to try to put this confusion and uncertainty to rest
once
and
for
all:
ALL THE THINGS. ALL.
At least as far as we are concerned. I think it's well past time to
put
the
past in the past.
That's great, Roy. That's a *lot* simpler than parsing the
recommendations, WCRR, community norms, and such at [A, B] :)
Meanwhile, we have just put nearly 200 million works records up
as
linked
open data. [1], [2], [3]. If that doesn't rock the library open
linked
data
world, then no one is paying attention.
Roy
[1] http://oclc.org/en-US/news/releases/2014/201414dublin.html
[2]
http://dataliberate.com/2014/04/worldcat-works-197-million-
nuggets-of-linked-data/
[3] http://hangingtogether.org/?p=3811
Yes, that is really awesome. But Laura was asking about barriers
to
open metadata, so damn you for going off-topic with PR around a lack
of barriers to some metadata (which, for those who have not looked
yet, have a nice ODC-BY licensing statement at the bottom of a given
Works page) :)
A. http://oclc.org/worldcat/community/record-use.en.html
B. http://oclc.org/worldcat/community/record-use/data-
licensing/questions.en.html
--
Karen Coyle
[email protected] http://kcoyle.net
m: 1-510-435-8234
skype: kcoylenet
--
Richard Wallis
Founder, Data Liberate
http://dataliberate.com
Tel: +44 (0)7767 886 005
Linkedin: http://www.linkedin.com/in/richardwallis
Skype: richard.wallis1
Twitter: @rjw
--
Karen Coyle
[email protected] http://kcoyle.net
m: 1-510-435-8234
skype: kcoylenet
--
Karen Coyle
[email protected] http://kcoyle.net
m: 1-510-435-8234
skype: kcoylenet
--
Karen Coyle
[email protected] http://kcoyle.net
m: 1-510-435-8234
skype: kcoylenet