Consider SOA interfaces, “canonical data models” and similar structural approaches as viable when it is practical to assert a single standard structure for information exchange or access. You can actually get quite a bit of mileage out of such structural standards. The idea with a standard structural approach is that it can be sufficiently general to cover your business needs and it provides a single standard to adapt existing systems (usually manually). This works particularly well where the structural (SOA or information) model is used with a model driven architecture (MDA) pattern so that it is technology neutral and expressed in business terms.
Where structural approaches start having trouble is where there is a lot of variation – perhaps because of independently conceived models/interfaces/vocabularies that need to come together or due to different authorities, user viewpoints or perspectives. In these cases the standard structures tend to be brittle and difficult to federate or evolve. Ontologies can provide as conceptual hub between different ways to represent or analyze the same or related information. Since the ontologies are built around the domain, not the technology or a specific use case, they tend to be more flexible. In addition some of the axioms available in ontologies allow for better federation of independently conceived models or integration of diverse viewpoints.
There is not a “hard line” between these approaches and the technologies are frequently confused with the techniques. You can do ontological modeling in “structured” languages (like UML) and structured applications in OWL – it is the practice as much as the technology that will impact your results.
Hope this helps a bit (and the opinions expressed are my own)!
Model Driven Solutions
From: ontology-summit-bounces@xxxxxxxxxxxxxxxx [mailto:ontology-summit-bounces@xxxxxxxxxxxxxxxx] On Behalf Of MacPherson, Deborah
Sent: Monday, April 25, 2011 10:38 AM
To: Ontology Summit 2011 discussion
Subject: [ontology-summit] Legacy Software to Web Services
Hi John, Everyone
I am seeking a brief explanation for how ontologies can help bridge the gap from legacy software to web services. For example - the role of ontologies versus APIs, middleware, and relational databases - a process model or short narrative of all the parts that would need to work together.
For facility and geospatial data, most likely a large set of RDF assertions and small set of Tiny Ontologies All Stitched Together (TOAST) could solve many of our current problems. However to pitch this better, a concise explanation is needed for exactly where an ontology should live in relation to projects at firms behind firewalls versus accessible across an enterprise/industry. And what the effect is on legacy software and databases i.e. not losing or recreating data, instead bringing them up to speed for more precise information sharing and requirements tracking.
DEBORAH MACPHERSON, CSI CCS, AIA
Specifications and Research
1100 Wilson Boulevard, Suite 2900
Arlington, Virginia 22209
Direct Line 703 907 2353
4 Digit Dial 6353
ü Please consider the environment before printing this email.