On 2014-10-13 14:16, Sarven Capadisli wrote:
On 2014-10-13 13:54, Frans Knibbe | Geodan wrote:
Hello!
I wonder if a way of recording changes in properties of resources can be
recommended. Many resources in real life have properties that have a
time range of being valid. In some datasets, only the current (or most
recent) state of a resource is stored, but in many cases it is important
to keep track of the history of development of a resource.
An example:
:john_smith
a foaf:person ;
foaf:name "John Smith" ;
Let's say that on 2013-09-27 John Smith marries Betty Jones. John Smith
is still the same person, so it makes sense to extend the same resource,
not create a new version:
:john_smith
a foaf:person ;
foaf:name “John Smith” ;
ex:marriedTo :betty_jones ;
How could I efficiently express the fact that the statement :john_smith
ex:marriedTo :betty_jones is valid from 2013-09-27? And if the couple
divorces, that the property has expired after a certain date? It would
be nice if the way of modelling makes it easy to request the most recent
state of a resource, any historical state, or a list of changes during a
time period.
A quick web scan on the subject revealed some interesting research
papers, but as far as I can tell all solutions need extensions of RDF
and/or SPARQL to work.
Perhaps this question is really about the ability to make statements
about a triple? Which is a problem for which no satisfactory solution
has been found yet?
Regards,
Frans
Hi Frans,
This is not a comprehensive answer on this topic, but you might want
to take a look at PROV-O [1] (which can address validity and history
of entities) and maybe even employ OA [2].
I had a look at PROV, but I got the impression that it can be used to
describe the provenance of resources, not properties of resources. In
other words, PROV seems to assume that if a property of a resource
changes, a new version of the resource is created. For resources like
documents that is not a strange assumption. But for things like a person
or a city it makes less sense to create a new version if some property
changes. Also, creating new versions can lead to a lot of data replication.
Capturing temporal dimension of linked data by Jindřich Mynarz is an
excellent read [3].
Yes, it is! I came across it earlier and I was impressed by the quality
and thoroughness of the analysis. I must confess that I do find it
difficult to completely wrap my head around some issues.
The thing that stood out for me was that although there are several
approaches to solving the problem there is not a single one that stands
out as the clear winner. Other web sources could not make a firm
recommendation either. That is why I thought I would ask the question on
this list.
The author does seem to have the most faith in a solution that is based
on named graphs. I am starting to think about whether it is possible or
smart to relate a triple to multiple named graphs (meanwhile I am
starting to think that thinking about Linked Data is very close to the
subject itself: everything is interconnected and any closed world
assumption needs to be abandoned quickly).
Greetings,
Frans
[1] http://www.w3.org/TR/prov-o/
[2] http://www.openannotation.org/spec/core/
[3]
http://blog.mynarz.net/2013/07/capturing-temporal-dimension-of-linked.html
-Sarven
http://csarven.ca/#i
------------------------------------------------------------------------
Frans Knibbe
Geodan
President Kennedylaan 1
1079 MB Amsterdam (NL)
T +31 (0)20 - 5711 347
E [email protected]
www.geodan.nl <http://www.geodan.nl> | disclaimer
<http://www.geodan.nl/disclaimer>
------------------------------------------------------------------------
