At the ontolog-summit this week I introduced a standards effort that is starting in the Object Management Group (www.omg.org) – “Semantic Information Modeling for Federation” (SIMF).
The intent if SIMF is to directly tackle the “data problem” where interoperability, federation, reuse and sharing of information is made difficult due to variation in terminology, viewpoint and representation. Our belief is that this is best tackled with a semantic approach, one which is tuned to the needs and viewpoints of those who create and federate data and messaging models. Currently none of the data modeling capabilities, from E/R to UML to XSD handle the integration of independently conceived models and even the structural mapping techniques that are used today are non-standard (With the exception of OMG-QVT). On the other hand, the semantic technologies are not providing standard conceptual frameworks and notations that resonate with these users. There seems to have been sufficient progress in existing tools and techniques in this area to suggest the need for standards. Note that as an OMG effort this is not research or an academic pursuit, but would bring some order to already developed approaches to foster the development of mainstream tools and industry support.
The opportunity exists to make substantial gains in solving the data problem – a problem that has a multi-trillion dollar impact on cost and lost opportunity. A 100% solution is not required or expected, providing incremental improvements will be a success.
SIMF can also be seen as a start to a more semantically based approach to architecture, one where multiple viewpoints of systems can co-exist yet share information and be mutually supportive. Models and modeling languages (including those from OMG) have not been sufficiently integrated, causing users problems when they want to look at more than just process, just objects or just information. etc. In that UML has many diagrams it has served that purpose to some extent, but was not designed for that role and the issues with more and more UML profiles suggest it is not the right foundation for a wide family of architectural languages at all levels. For this reason SIMF is expected to federate with UML, not be based on it.
More information on SIMF can be found here: http://www.omgwiki.org/architecture-ecosystem/doku.php?id=semantic_information_modeling_for_federation_rfp
The status of SIMF is that we are still writing the RFP, which will be presented at the next OMG meeting. Note that writing the RFP requires a good specification of what is being asked for, not of the solution – the solutions come in responses to the RFP. Those interested are invited to participate on the mail list and wiki. Instructions on participation can be found on the wiki here: http://www.omgwiki.org/architecture-ecosystem/doku.php