Last I was reading Mike Bergman’s blog on the seven pillars of the open semantic enterprise. The seven pillars are a mere summary of the European Semantic Technology research programme (in which I was active as well) of the last 10 years. Mike takes the tricky but brave step to push this a bit more to industry in terms of seven pillars.
I like the idea of the open semantic enterprise. The problem however (and true for many business-it problems) is that these pillars are too technology-religious (rather than agnostic). Ontologies (I still do not like the word) should not be built by techies. That usually results in models where business concepts are labelled in tech lingo (e.g., CustIDRel, or ProdLbl-ID), and the rules to not correspond to what regulations enforce. Therefore business experts themselves must be involved in the job. In Business Process Management, business experts are supposed to model the processes conceptually in fulfillment of certain business objectives and compliant to a set of regulations. Similarly, in Business Semantics Management, business experts decide on a shared base of business semantics: i.e. meaning of vocabulary and rules used to communicate and coordinate processes, data, and applications. This happens again according to certain rules and business goals.
Agreeing on business semantics brings me to another important social pillar that is missing here, or should I say first-class citizen: the human stakeholders in the extended enterprise. Business experts in the business zone of the “semantic enterprise” must be facilitated with the right methods and tools (see e.g., Collibra or VUB STARLab) to agree on the meaning of their vocabularies and business rules (i.e. business semantics) before even thinking about flattening a version of that agreement into an enterprise information model (whether in monolithic OWL or in industry more widely known software modelling technique UML that is dubiously abused for knowledge modelling) that can be passed on to “semantic application” builders in the tech zone of the enterprise.
During the business semantics unification process, all perspectives should be taken into account: the business context drives the scope of business semantics, but as business semantics ultimately will be used to align with ICT, the technies’ perspectives should be incorporated as well. If the techies discover flaws in enterprise information model (E.g., some terms are underspecified leading to misunderstandings or so) these new requirements for starting a new consensus process is fed back to the business zone. SBVR is a promising standard OMG based (hence MOF compliant) on the proven foundations of fact-oriented modelling to allow business experts to rerpesent business vocabulary and rule semantics in a natural manner. Full-cycle Business Semantics Management then takes into account both feedback from both Business in (e.g. in SBVR) as ICT perspectives (e.g., in UML also MOF-compliant) when reconciling business semantics. Of course governance in terms of roles and responsibilities of the activities is also important. Social performance indicators must monitor the construction and management of business semantics at all time.
I also like the emphasis of the open world assumption. I have been evangelising this idea for years (much inspired by my former lab director and PhD supervisor Robert Meersman). Today I emphasise more on the opposite notion: indeed ICT in industry suffers the closed world “syndrome”. It is inspired by the closed world assumption. That is the presumption that what is not currently known to be true, is false. On one hand, this is caused by a rather naive assumption that there would never be a need for large-scale data exchange. Such attitude is obviously hilarious in the Internet age and this quote should be eternalized in the Hall of Most Stupid Quotes next to Bill Gates’ and Bush Jr.’s historic quotes. On the other hand, the closed formats are deliberately caused by vendor-specific database formats that create lock-in this way.