Re-tooling Libraries for the 21st Century

The Integration of Resource Description Framework (RDF)
Object Linking into
Islandora XML-Form Functionality
Exemplified by The FBA Vocabulary and Object Linker Modules in Islandora
Nick Bywell: Freshwater Biological Association
9th May 2014
The FBA Vocabulary Module
The “FBA Vocabulary Module” was written by Simon Fox
of the FBA as part of the same project and also
benefitted from the oversight of Dr Michael Haft.
Specialist expertise in the use of metadata-schema was
provided by Richard Gartner and Dr Mark Hedges of
KCL. Resource Description Framework (RDF) expertise
was provided by Dr Anna Jordanous of KCL. The
analysis of the data requirements was performed by Ian
Pettman and Hardy Schwamm of the FBA.
The FBA Object Linker Module
The “FBA Object Linker Module” was written by Nicholas
Bywell of the Freshwater Biological Association (FBA) as
part of a joint project carried out by the FBA and King’s
College London (KCL) to archive the scientific data
produced by both the Demonstration Test Catchment
(DTC) Project and the Greenhouse Gas Project (GHGP),
projects that are funded by the UK’s Department of the
Environment, Fisheries and Rural Affairs (DEFRA). The
development of the module benefited from the oversight
of Dr Michael Haft of the FBA and draws on some of the
ideas and ontology processing developed by Giancarlo
Birello of CNR-Ceris and Rosemary Le Faive of McGill
University in the “Islandora Relationship Editor” module.
PHP 5.3.3
What does it look like from the perspective
of someone with Editor privileges?
Object Linker
One submit => Multiple records ingested
What happens if one of the records is
subsequently edited?
One edit => Multiple edits of symmetrically
linked datastreams and RDF links +
Multiple futher ingestions
What happens if you autocomplete on a
collection that, unlike a vocabulary, can
contain duplicate titles, such as a collection
of images?
The appended IDs distinguish duplicates
and are automatically removed on
A future enhancement could be a cursorhover function that retrieves specified fields
from the duplicate records to allow them to
be more easily distinguished.
What happens when records are purged?
Symmetric Datastream entries + RDF
statements are unwound on single and
collection purges, but not when individual
datastreams are purged or when purging is
done via the Fedora-Admin Utility.
How is a PDF record linked to a vocabulary
or geographic term?
The standard PDF XML Form has “Topic”
and “Geographic” fields that are Tag-type
fields so it is not possible to set up the sort
of autocomplete that required for object
Therefore they must be converted to Tabpanel-type fields. So there are implications
for existing data because the XML
produced by the two types are not
compatible with each other.
How is the object linking configured in the
Administrator functionality?
As with the other types of autocomplete, in
order to take effect, the name of the
autocomplete entry must be quoted in the
Advanced Options “Autocomplete” field
associated with a specific field on an XMLForm.
It can require the completion of 22 fields to
configure the object-linking behaviour of
one field on an XML form. 
How is the link from a PDF subject term to
a vocabulary term configured?
The autocomplete that controls the linking
of a broader subject term to a narrower
subject term in the vocabulary requires all
the fields to be completed because there
is a symmetrical relationship between the
fields in the “subject” and “object”
To cater for the scenario where a PDF has
been linked to a vocabulary term, and the
vocabulary term has been subsequently
purged, there would ideally be an option of
attaching a “free floating” autocomplete to
the vocabulary form. There is currently no
such facility, so the autocomplete has been
attached to a hidden “Creation Date” field,
which is a rather ugly solution.
If the autocomplete linking configuration is
changed after records have been added,
the RDF-links and datastream-entries
associated with any autocomplete entries
that have been removed will not be
unwound on subsequent edits and purges.
The functionality has implications for
collection design and the “Content
Model” to “XML-Form” relationship.
These issues and further detail on how to
configure the object-linking will be
covered in an explanatory video.
The Vocabulary Module is just one
possible application of the Object Linker
processing. It also lends itself to other
areas such as taxon processing, plus the
purpose it was originally designed for, the
publishing of scientific research data as
open linked data for the DEFRA DTC &
GHG Archiving Project.
In relation to the same DEFRA Project, Dr
Anna Jordanous of King’s College London
will shortly be describing her work on
ingesting and validating CSV files and
generating the relevant RELS-INT

similar documents