My confusion with linked data is how it works at a logistical level. I
understand the idea of URIs and SPARQL and all that. My problems are at
the planning level.

1. How do you find/publish/update taxonomies? The creation of a linked
data application assumes you know where to find these things, but by the
very nature of linked data, they are decentralized and spread apart. I
know there are a few curated listings, but it seems like that would be
nearly impossible to keep up to date.

2. What do you do about overlaps between two taxonomies? I understand that
from a technical perspective there is no overlap as long as there are two
separate URIs, but in human terms, there is a lot of overlap and little
way of knowing which taxonomies are really best for which purposes. Again,
this seems inherent to the decentralized nature of linked data.

Josh Welker

-----Original Message-----
From: Code for Libraries [mailto:[log in to unmask]] On Behalf Of
Akerman, Laura
Sent: Wednesday, September 04, 2013 9:14 AM
To: [log in to unmask]
Subject: Re: [CODE4LIB] What do you want to learn about linked data?

Karen,

It's hard to say what "basics" are.  We had a learning group at Emory that
covered a lot of the "what is it", including mostly what you've listed but
also the environment (library and cultural heritage, and larger
environment), but we had a harder time getting to the "what do you do with
it".... which is what would really motivate and empower people to go ahead
and get beyond basics.

Maybe add:

How do you embed linked data in web pages using RDFa (Difference between
RDFa and schema.org/other microdata) How do you harvest linked data from
web pages, endpoints, or other modes of delivery?
Different serializations and how to convert How do you establish relations
between different "vocabularies" (classes and properties) using RDFS and
OWL?
(Demo) New answers to your questions enabled by combining and querying
linked data!

Maybe a step toward "what can you do with it" would be to show (or have an
exercise):

How can a web application interface with linked data?

I suspect there are a lot of people who've read about it and/or have had
tutorials here and there, and who really want to get their hands in it.
That's where there's a real dearth of training.

An "intermediate level" workshop addressing (but not necessarily
answering!) questions like:

Do you need a triplestore or will a relational database do?
Do you need to store your data as RDF or can you do everything you need
with XML or some other format, converting on the way out or in?
Should you query external endpoints in real time in your application, or
cache the data?
Other than SPARQL, how do you "search" linked data?  Indexing
strategies...  tools...
If asserting  OWL "sameAs" is too dangerous in your context, what other
strategies for expressing "close to it" relationships between resources
(concepts) might work for you?
Advanced SPARQL using regular expressions, CREATE, etc.
Care and feeding of triplestores (persistence, memory, ....) Costing out
linked data applications:
   How much additional server space and bandwidth will I (my institution)
need to provision in order to work with this stuff?
   Open source, "free", vs. commercial management systems?
Backward conversion -transformations from linked data to other data
serializations (e.g. metadata standards in XML).
What else?

Unfortunately (or maybe just, how it is) no one has built an interface
that hides all the programming and technical details from people but lets
them experience/experiment with this stuff (have they?).  So some
knowledge is necessary.  What are prerequisites and how could we make the
burden of knowing them not so onerous to people who don't have much
experience in web programming or system administration, so they could get
value from a tutorial,?

Laura

Laura Akerman
Technology and Metadata Librarian
Room 208, Robert W. Woodruff Library
Emory University, Atlanta, Ga. 30322
(404) 727-6888
[log in to unmask]



-----Original Message-----
From: Code for Libraries [mailto:[log in to unmask]] On Behalf Of
Karen Coyle
Sent: Wednesday, September 04, 2013 4:59 AM
To: [log in to unmask]
Subject: Re: [CODE4LIB] What do you want to learn about linked data?

All,

I had a few off-list requests for basics - what are the basic things that
librarians need to know about linked data? I have a site where I am
putting up a somewhat crudely designed tutorial (with exercises):

http://kcoyle.net/metadata/

As you can see, it is incomplete, but I work away on it when so inspired.
It includes what I consider to be the basic knowledge:

1. What is metadata?
2. Data vs. text
3. Identifiers (esp. URIs)
4. Statements (not records) (read: triples) 5. Semantic Web basics 6. URIs
(more in depth) 7. Ontologies 8. Vocabularies

I intend to link various slide sets to this, and anyone is welcome to make
use of the content there. It would be GREAT for it to become an actual
tutorial, perhaps using better software, but I haven't found anything yet
that I like working with.

If you have basics to add, please let me know!

kc



On 9/1/13 5:37 PM, Karen Coyle wrote:
> I'm thinking about training needs around linked data -- yes, that
> includes basic concepts, but at the moment I'm wondering what specific
> technologies or tasks people would like to learn about? Some obvious
> examples are: how to do SPARQL queries; how to use triples in
> databases; maybe how to use Protege (free software) [1] to create an
> ontology. Those are just a quick shot across the bow, and from my
> basically non-techie point of view. Please add your own.
>
> If you can't say it in terms of technology, it would be as good (if
> not maybe better) to say it in terms of what you'd like to be able to
> do (do searches, create data... )
>
> This is very unscientific, but I think it's a worthwhile conversation
> to have, and maybe can help get some ideas for training.
>
> kc
> [1] http://protege.stanford.edu/
>

--
Karen Coyle
[log in to unmask] http://kcoyle.net
ph: 1-510-540-7596
m: 1-510-435-8234
skype: kcoylenet

________________________________

This e-mail message (including any attachments) is for the sole use of the
intended recipient(s) and may contain confidential and privileged
information. If the reader of this message is not the intended recipient,
you are hereby notified that any dissemination, distribution or copying of
this message (including any attachments) is strictly prohibited.

If you have received this message in error, please contact the sender by
reply e-mail message and destroy all copies of the original message
(including attachments).