Talis Forums Archive

A place to discuss all your library issues
It is currently Sat Sep 23, 2017 10:48 pm

All times are UTC [ DST ]




Forum locked This topic is locked, you cannot edit posts or make further replies.  [ 6 posts ] 
Author Message
PostPosted: Mon Jun 16, 2008 10:21 am 

Joined: Thu Aug 19, 2004 10:40 am
Posts: 46
Location: University of Salford
If I wanted to store some data in the Talis Platform, and if I wanted to start with something basic such as data about the organisation I work for, how do I go about choosing URIs to represent objects in the store? Are there any conventions?

For example, to represent "The University of Salford", do I just use "http://www.salford.ac.uk", which happens to be the url of our home page?

Also, how would I create a statement that asserts the fairly obvious fact that "the University of Salford is a university"? Is there some url that represents a "standard" definition of a university?

And is it then possible to assert that "http://www.salford.ac.uk/library" "is the library of" "http://www.salford.ac.uk"?

Or if that's not possible, is there a standard way of saying something like
"http://www.salford.ac.uk/library is a library" and
"http://www.salford.ac.uk/library" "is a department of" "http://www.salford.ac.uk"?

_________________
Andy Bourne
University of Salford


Report this post
Top
 Profile  
Reply with quote  
 Post subject:
PostPosted: Mon Jun 16, 2008 2:22 pm 

Joined: Mon Oct 18, 2004 11:07 am
Posts: 188
Location: Talis
Hi Andy,
You might wish to have a look at http://directory.talis.com and perhaps more specifically this url which is an entry for the University of Salford.

You will see from the XML RDF returned by that link, that there are several namespaces used to allow definition of the data contained within each element.

To answer your question about the URI to represent the University of Salford, you could think of this a bit like the concept of a bib id in alto (as being a unique number applying only to the local collection) which in the case of the directory, is represented by the <rdf:about> element. There is no reason why you could not just use that.

Incidently, if you wanted to, using the uri used by the directory would mean that you could then query the directory for all other information related to that uri in the directory from your own application - that could then give you useful data for your app.

I would also point you to the n2 wiki where you will find other info around this area.

foaf (friend of a friend) could possibly be used to specify relationships between departments, although there may be other ways :)

Hope this helps somewhat?

_________________
Cheers,
Tim
-------
Tim Hodson
Consultant
Talis Information
www.talis.com
www.timhodson.com
informationtakesover.co.uk (blog)


Report this post
Top
 Profile  
Reply with quote  
 Post subject:
PostPosted: Mon Jun 16, 2008 3:07 pm 

Joined: Thu Jul 01, 2004 4:14 pm
Posts: 72
Location: Talis
Tim was quick off the mark and gave you most of the reply I was composing, but here it is anyway:

You would describe your organisation using a combination of ontologies and/or RDF schema.

Which you use will depend on what you are trying to achieve. You could simply use something like FOAF:
Code:
<foaf:Organization rdf:about="#me" xmlns:foaf="http://xmlns.com/foaf/0.1/">
  <foaf:name>The University of Salford</foaf:name>
    <foaf:homepage rdf:resource="http://www.salford.ac.uk" />
</foaf:Organization>


You could use or combine this with others such as the Library Schema used in the Silkworm directory and/or the Academic Institution Internal Structure Ontology - work by several Talis people to help the description of academic institutions.

As Tim says, the N2 Wiki is a good place to get more information about developing with the Platform - come back to me if you want any further information.

_________________
Richard Wallis
Technology Evangelist, Talis


Report this post
Top
 Profile  
Reply with quote  
 Post subject:
PostPosted: Mon Jun 16, 2008 3:36 pm 

Joined: Thu Aug 19, 2004 10:40 am
Posts: 46
Location: University of Salford
Tim, Richard,

Thanks for the info. Posting the forum item was my tentative first step towards trying to understand what the Platform and RDF is all about. When you try to understand a new programming language these days, finding your way round the class libraries that come with it is usually more of challenge than understanding the core language. Although it's not a programming language I feel the same about RDF - it's got a simple core, but it's not easy to get a grasp of what ontologies [a bit like class libraries in my analogy] are appropriate to what circumstances. I'll have a think about your replies and will get back to you later [possibly much later!].

Andy

_________________
Andy Bourne
University of Salford


Report this post
Top
 Profile  
Reply with quote  
 Post subject:
PostPosted: Tue Jun 17, 2008 10:15 am 

Joined: Thu Aug 19, 2004 10:40 am
Posts: 46
Location: University of Salford
Tim, Richard,

I didn't have any particular application in mind when I wrote the original post but I was thinking that, as the semantic web and RDF pick up a bit more momentum, I could find that RDF statements about the University of Salford library [UoS] start turning up all over the place. Anyone can devise any unique identifier they like to represent UoS and anyone could make any kind of RDF statement about UoS. These RDF statements could occur in data stores like the Talis Platform (including the Silkworm directory) or they could occur as RDFa attributes embedded in XHTML documents on various different web servers.

The implication is that, if we want to influence what unique identifiers people use to represent UoS-related entities and if we want to influence what facts are asserted about them then we need our own "presence" in semantic web terms just as we have a presence in the "normal" web.

For example, in the Talis directory entry you mentioned above, the unique identifier for the UoS library is "http://directory.talis.com/res/dddconv-papid413789#self" - not what I would have chosen myself!

Also, the foaf:homepage attribute has the value "http://www.salford.ac.uk". I might regard the "correct" value to be "http://www.salford.ac.uk/library". So your version the homepage fact might contradict my version of the same fact; how would I get the Talis directory updated to be consistent with my version? And are there any other directories out there that also need updating? - I only know about yours because you told me about it!

But then, if we were to set up our own repository of RDF data or if we started using space on the Talis Platform for the same purpose, how do we publicise its existence? And in the event that there's a contradiction between our version of a fact and someone else's version, how do we persuade people to use data about the UoS library taken from our repository rather than somewhere else?

_________________
Andy Bourne
University of Salford


Report this post
Top
 Profile  
Reply with quote  
 Post subject:
PostPosted: Tue Jun 17, 2008 11:18 am 

Joined: Mon Oct 18, 2004 11:07 am
Posts: 188
Location: Talis
Andy,
You may find this FAQ on the semantic web to be of use. at the very least it will introduce you to some terms and concepts for further reading.
http://www.w3.org/RDF/FAQ

By the way, you can register for a username for the silkworm directory (if you haven't already) and update any of the information there, to keep it up-to-date.

As the statements in rdf are assertions of fact, it is necessary to use what some have described as 'radical trust'.

You raise some interesting points about the control that you may or maynot have over the data describing UoS. what I would suggest is that people will trust information about an entity that is described on a source that is controlled by that entity, over and above a source that is not.

As an example, I would trust rdf embedded within http://salford.ac.uk/library.html over and above rdf embedded in http://studentx.com/salfordlibrary.html

In terms of publishing it's existence (to machines rather than people), there are several ways of doing so. a GRDDL profile could tell an 'aware' web crawler that other information was to be found and how to find it. also a request to a server for a resource with a html mimetype might return an html page, whereas the same resource request to the same server with an application/rdf+xml might return an xml doc of rdf describing the resource.

_________________
Cheers,
Tim
-------
Tim Hodson
Consultant
Talis Information
www.talis.com
www.timhodson.com
informationtakesover.co.uk (blog)


Report this post
Top
 Profile  
Reply with quote  
Display posts from previous:  Sort by  
Forum locked This topic is locked, you cannot edit posts or make further replies.  [ 6 posts ] 

All times are UTC [ DST ]


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum

Search for:
Jump to:  
cron
Powered by phpBB® Forum Software © phpBB Group