Business Decision Models Inc.

Yesterday's FOIS 2010 Workshop participants enjoyed scheduled morning presentations by Antony Galton, Fabian Neuhaus, Barry Smith and Michael Grüninger. This blog item is a report of those proceedings, along with a few editorial comments. A companion blog item will offer some analysis of the intersection of ontological engineering and practical opportunity, from the perspective of sociology and business.

Ontological engineering may be reaching a tipping point in capability and acceptance, and if that tipping point is reached, the construction of dramatically more powerful and effective software would be expected in the foreseeable future (although how steep the post-inflection-point curve is, remains to be seen.) But this ideal vision will only be reached if the market for ontology sciences, services and products takes off.

There are reasons to be sceptical about how quickly we will reach an inflection point in adoption of ontological engineering. For almost 30 years computer scientists and potential beneficiaries of software innovation have endured successive hype cycles in AI, expert systems and ontology, three closely related disciplines. It is perhaps unfair to use the word “hype”, pace Gartner, but it is indisputable that the human imagination of “the possibilities for the new” regularly outruns the difficulties of actual discovery, engineering and productization. The avatars of ontology and AI then are to be congratulated for their persistence of vision and labour.

At a high level, the morning's proceedings could be characterized as covering the developing demand, supply and market clearing for ontological knowledge. For his part, your host also found the event very stimulating, especially given that he is on the sales team for Intalio (www.intalioworks.com), the open source BPM provider, for which training and coaching is a significant activity. Intalio's skills transfer and strategic coaching business is all about “evangelization for BPM technology adoption”. With this in mind, yesterday's FOIS 2010 workshop on education was similarly "all about evangelization for ontological technology". And like all good evangelists, ontological engineers can certainly be enthusiastic about their subject, and with some justification: they are modelling what is perhaps the most important of things to model, which is reality. That might be justification enough for the theoretically-minded, but the practical applications of ontology, already deployed in some cases, includes some of the most significant real applications of computing, especially in medicine. 

Here is a summarized list of key themes from the morning's discussions, including comments by presenters and audience, mixed in with a few comments by the blogger:

1. Demand for Ontology Increasing: From government and industry, as well as academia itself, the demand for ontological engineers is increasing. (This information apparently comes from interviews with practising ontologists, perhaps not the most independent of sources, nevertheless the observation is likely a true reflection of the current situation.)

2. Supply of Ontological Engineers Not Increasing In Step: The production of a cadre of well-trained ontologists is a lengthy and expensive process – the supply is therefore inelastic. For a variety of reasons, including the next item, the supply inelasticity problem is not likely to disappear overnight, with the result that deployment of ontology-enabled applications will be retarded. 

3. Definition of Field Not Mature: Interestingly, FOIS ontologists commented on the lack of maturity in the ontology field. There is no consensus at this time as to a commonly accepted “body of knowledge” or “shared terminology”. This state of affairs is felt to be very unsatisfactory, and again retarding the development of the field. It was observed by several participants that the creation of a text book on ontological engineering would be a huge milestone, and encouragement to the field. Comment: Paul Samuelson's famous text book on economics, apparently influencing a generation of leaders in favour of Keynesian economics, is a good example of a domain-changing textbook. The sense that a textbook would help drive the field, for instance to common terminology and a common body of knowledge, is probably true. But the fact that a textbook has not appeared to date may be a symptom of momentum in the field, and that a textbook might not be feasible until the field itself is more mature, almost a chicken or egg problem. The answer to the chicken or egg problem might be “outside the system”, that is to say, momentum in the field of ontological engineering is to a great extent influenced by external factors such as demand for better software (unfortunately not assured) and budgets for software investment.

4. Success Stories Are Specific, Not Widespread: The availability of “wins” (using sales nomenclature) or “success stories” was felt to be very important for the increasing acceptance of the field of ontological engineering. However, while there are notable success stories (the genome ontology for example), these examples are themselves rather daunting and not necessarily models for easy replication. 

5. Ontology Production Varies Widely In Quality: Perhaps as an artifact of the lack of maturity in the field, found ontologies-in-the field vary widely in quality; there is even a suggestion that the distribution is not only “wide”, but skewed to the inferior. Mere use of the term ontology and a few ontology-oriented software tools is no guarantee of good quality results – with disappointment (hopefully) for project sponsors and negative feedback on a perception of the field. Further, the production of sub-optimal ontologies almost by definition reduces re-usability, and thus increases the isolation of organizational functions from each other – which is a primary motivation to undertake ontological engineering in the first place. 

6. Re-Use, Documentation & Complexity: These three topics were discussed; it seems that re-use, document and complexity are closely related. Ontologies-in-the-raw, without documentation are almost unreadable, even by ontologists. And this phenomenon is exacerbated the more significant the ontology. Together these three aspects of ontology highlight the fact that ontology is difficult and probably expensive, thus suggesting that organizational commitment to the project and to standards, are both essential to the success of any ontological engineering project. Embarking on a first ontological engineering project involves a signficant hurdle, or step function – another barrier to adoption.

7. Ontology-First Or Domain First? There was a significant debate on whether ontological engineering projects are best led by domain experts who have been trained in ontology, or ontological experts who learn a domain. The first option seemed to carry the day.

8. Vision, Hype & Cynicism: Any new field requires enthusiasm, visionaries, insight, competence and persistence, and the field of ontological engineering has multiple exemplars of individuals embodying all these qualities. There would have been no progress otherwise. Nevertheless, beyond the core of domain champions and the up-and-coming cadre of newly trained ontological engineers, there are two categories of the “ontologically aware” for which ontology is not an unqualified benefit.  First, there are those who expended their enthusiasm earlier on AI and expert systems, with disappointing results -- an experience that even the "ontologically hopeful" would not want to repeat, but increasingly which is perceived as a possibility, especially given the appearance of ontology-related hype phenomenon.  This first group of people could be described as having lower expectations than are justified.  A second group of people have no particular hopes for ontology, but again are not in touch with reality. This second group is comprised of the cynical, for whom ontology represents the latest “programming trick” and flavour of the day.  This group could be described as having inflated and unjustified expectations.  Given the challenges of doing ontology, the chances that either group, those with too many expectations and those with too few, will contribute anything other than noise in the market for ontological services, is dim.

The Tuesday morning FOIS session seemed to be almost a case study in the sociology of knowledge, as if one was in a laboratory viewing the construction of a new communities of understanding. The literature on the economics and sociology of technology diffusion, the history of innovation, theories of innovation, and the history of science, ranging from Thomas Kuhn to Clayton Christensen, provides a suite of tools for contextualizing the development of the discipline of ontological engineering. That will be, within the limits of the format of blogging, the topic of a planned companion blog item.

Draft Version 0.3