Use Cases
To help data migrate into a Semantic Web, I am expanding this web app to work as a Validating Generator
for metadata. Its particular benefit, tracing mostly to
SAID-based ontologies, is reduced complexity for a data provider's
staff.
SUBJECT MATTER EXPERT: With no training in
logic programming, such people can quickly create ontologies describing
domain-specific data. The process uses a fixed upper ontology to
help annotate, map and validate related triples:
Step 1: Spec a Namespace - SAIDs for all classes and properties need to be defined which cover
the expected problem domain, or the known schema of a relational data
base to be converted. Click on EDITOR and follow its instructions.
Step 2: Spec Constraints
- For validation, each property SAID needs
two more - its domain and its range. The web API for step 3 can
take empty or early "example" triples as defining the patterns which a
namespace deems legal.
SCRIPT WRITER: Given a SAID namespace
annotating a DB schema, any coder in IT can likely script the
conversion of a related data base into posted or returned triples with
little training beyond the requirements.
Step 3: Import SAID triples - Each triple relates a subject to its object
or a literal. The web API (see this example)
may be explored manually as a spec, but for bulk imports, batch
scripts driven by bulk data should pass it HTTP requests
Step 4: Export Result Sets
- A generator can dump these as CSV files,
or (eventually) run plug-ins to emit N3, RDF, OWL, XTM, RSS, CGIF, HL7, XHTML, etc.
Data providers may need such tools, but mostly they aid data consumers.
I'd like to sample community opinion on strengths and
weaknesses of
the above plan, which demands knowledge of RDFS and OWL only by those
who write or use related Step-4 plug-ins. Simpler uses for RDF
may need only SAIDs, valid triples and search logic..
Please try out this web app enough to
see how SAIDs might work in summarizing your own favorite data sets or
enriching your favorite domain ontology, then comment on potential use
cases, such as these:
- Q/A: Upper ontology (UO) choices help domain ontologies clarify what terms mean
-
Doc: Publishing SAIDs (as PSAIDs) can help end-users learn a domain vocabulary
- DBs: Converting from foreign DBs gets easier if PSAIDs exist as "targets" to match
- OWL: Constraints can exploit 10 new (broadly shared) UO semantic discriminants
- FAQs: Faceted queries across name spaces can exploit those same discriminants
- XRefs: Interacting with another ontology gets simpler once PSAIDs are exchanged
- Apps: A web service taking PSAID arguments can tap their embedded semantics
- URIs: Blank nodes for exchange files can be globally named by mutating PSAIDs
Your feedback on uses or problems would be much appreciated. Thanks
in advance for considering this new URI notation. Please let me know if
I can help in your evaluation of it, or in your data conversion projects.
Dan Corwin
Independent Architect/Consultant
Lexikos Corporation
|