2017-04-24-PIDs_Types_and_the_Semantic_Web-Schwardmann.pdf (608.65 kB)
PIDs, Types and the Semantic Web
PID Information Types are becoming a crucial role in scientific data
management because they can provide state (what) and binding (where)
information about digital objects as attributes of the PID. This is a
similar but much more flexible approach than the well known mime type
characterization, because both of these types concepts allow to decide
about preconditions for processes in advance and before touching the
data. One aspect of this is the need for standards and correctness of
the used types to ensure reliability for the processes operating on the
digital objects. This requires registries and schemas for PID InfoTypes
and suggests an automated schema generation process. Such a process in
combination with data type registries will be described in more detail
in the intended talk. Another aspect of PID InfoTypes is its intrinsic
grammar as subject-predicate-object triple, with the PID as subject, the
type as predicate and its value (often again a PID) as object in this
relation. Given the registration of types and the proposed syntactical
rigidness of the value, guaranteed by the schema, together with the use
of PIDs in subject and predicate, the type concept has the ability to
overcome the fuzziness and lack of reliability of semantic web
categories with its URL references and possibly changing locations and
content. The intended talk will also describe this approach in more
detail, discusses the differences to linked data and describes some
necessary technological developments for the type concept to keep up
with the possibilities currently provided by the semantic web.