between the steps in the analysis are described using RDF
expressions. At a higher level the predicates and vocabularies used
to describe the analysis process are defined in the ontology. The
process, which is referred to as the OntoProcess, is implemented
using OWL (W3C Web ontology Language), with a hierarchy structure as
shown in Figure 1.
1. Class hierarchy of OntoAnalysis ontology used to describe the
specified steps in the linear data analysis chain. The grey
background classes are essential to describe the analysis process.
As described above the
user can follow a chain of analysis pulling back data from the
relevant and distributed databases. While producing the data models
it became clear that while a traditional relational database was a
useful concept for systems that produce similar data over a long
period but it was not ideal for rapidly evolving experiments; the
effort to continually change the relational database structure
overwhelmed the advantages. Never-the-less, we did not wish to lose
the advantages of ensuring data was maintained and accessible provide
by a database (and associated managements systems). This led to the
development of the RDF based approach outlined above. The current
system uses a relational database to store the raw data out of the
laser experiments, together with considerable metadata about the
experiment. All the subsequent analysis and tracking is driven from
a RDF description.