The Rise of Ontology: A Conversation with Felix Van de Maele

Originally published 29 September 2010

The term "ontology" is increasingly being heard in data management forums and elsewhere. Yet it is not easy to find out precisely what it means and what significance it might have for the data management community. In fact, there is more than a suspicion that it is yet another buzzword used to market products and services without the need to clearly explain what is being provided. So is ontology "real," or is it destined to disappear in a trough of disappointment after a few years? To get some answers to this and other questions I recently turned to someone who is making a big bet on ontology—Felix Van de Maele, CEO of the semantic start-up Collibra.

Malcolm D. Chisholm (MDC): "Ontology" seems like one of those words that is very academic, and nobody really knows what it means. Personally, I have never worked on an ontology project—at least as far as I know—and I think that many data management professionals would say the same. Yet there is a tremendous amount being written about it. How would you define ontology, and why is it important?

Felix Van de Maele (FVdM): Technically ontology is the study of how things exist—it's a branch of philosophy. But I don't think that helps anybody to understand how it applies to practical problems, and at Collibra we do not use it because it does not suggest anything of practical value. I prefer the term "business semantics." Enterprises need to know exactly what is meant by terms like "Customer" or "Product," how these terms relate to each other, and what rules there are for them. Problems exist because different communities or contexts within enterprises define these terms differently, and that impacts the ability turn data into information effectively—as well as a host of other difficulties. Today, businesses want to get value out of their data, but if they don't really understand what the data means, they are not going to succeed.

MDC: I agree there are a lot of problems in unlocking the value of data, but surely aren't you talking about conceptual data modeling, which we have been doing for decades now? Isn't this "business semantics" as you call it?

FVdM: No—you're wrong there. There is a big difference. All data modeling is what we call "closed world." That is, it occurs within a framework of assumptions and implicit understandings. Usually it is oriented to building a physical database, and that's what data modeling tools were originally designed to do. Let me give you an example. Legacy databases are "closed worlds." They function pretty well with few production issues. But when you extract the data from them to use it somewhere else, it is usually a bloodbath. All the assumptions and implicit understandings now explode into real problems. And these are often connected to the content of the data as well as the structure—and the structure is all you will see in a data model.

We have an "open world" perspective. We try to get enterprises to describe their information without reference to any applications or databases. There must not be anything implicit—everything has to be made explicit. Every assumption has to be documented. You can't have the attitude that the business users make decisions about data content that impacts knowledge representation, while data modelers only look after structure. It is the single whole that matters to the enterprise in the end.

MDC: But surely data modeling has its place, although I must confess that it is not seen as solving as many of the issues of data management in the way it used to be.

FVdM: Don't get me wrong—of course data modeling has its place. But it is an activity that is carried out usually by one, or just a few, technical specialists at a time. We strongly believe that this cannot work for "open world" business semantics. We see collaboration as the key for knowledge representation. Business semantics exists differently in different contexts and for different communities in the enterprise. This is what has to be captured. It cannot be filtered and transformed by a small cadre of technical experts who make the decisions about what they think things are supposed to mean.

MDC: Are you talking about a free-for-all wiki-like environment for collaboration?

FVdM: Certainly not. You have to have a governance model to make the collaboration effective. For instance, "owners" have to be assigned to definitions, and there has to be a process for adding to the content of a definition as contributions come in from across the enterprise. Getting the relationships and business rules right for important business concepts just cannot be done in what you are calling a "free-for-all wiki-like environment."

MDC: I think that most data management professionals would agree with that, and I wish we had more time to dive into the governance model. But I have to close out by asking you what practical value you get from doing "business semantics"? A lot of senior executives may nod their heads in agreement with the ideas you are expressing, but at the same time I am sure they wonder if it is all just some kind of academic exercise.

FVdM: Let me give you an example. It is generally agreed that data exchange—say in XML messages—has to be as close to business meaning as possible, and not some derivative of an arbitrary database structure. If you read books and articles on enterprise service buses (ESBs)—or look at the federal government's data sharing initiatives—you see this clearly laid out as necessary, and often it is called a "canonical data model." Well, just how do you get to it? The best way is from a true representation of all the business concepts, clearly defined within their contexts and communities, with all their relationships and with all of their rules. If you can get to that point, it is not so difficult to create a set of XSDs that give you the canonical data model and optimize information exchange. And this is the kind of practical thing that is being done today.

MDC: We have been talking about "ontology," but what is "semantics"? Isn't that also a term that is getting overloaded and turned into some kind of magical talisman?

FVdM: I think that most people understand that "semantics" is about getting to the meaning of data. "Ontology" includes that, but also looks at the relationships amount the things represented by the data and rules for these things. Yes, it's true that "semantics" is being thrown around a lot today, but you still need it—data means something, and you have to know what it means.

MDC: Felix, thanks for your time today. I enjoyed our conversation, including the occasional amicable sparring.

FVdM: A little sparring is not a bad thing, and you are most welcome.

SOURCE: The Rise of Ontology: A Conversation with Felix Van de Maele

  • Malcolm ChisholmMalcolm Chisholm

    Malcolm Chisholm, Ph.D., has more than 25 years of experience in enterprise information management and data management and has worked in a wide range of sectors. He specializes in setting up and developing enterprise information management units, master data management, and business rules. His experience includes the financial, manufacturing, government, and pharmaceutical industries. He is the author of the books: How to Build a Business Rules Engine; Managing Reference Data in Enterprise Databases; and Definition in Information Management. Malcolm writes numerous articles and is a frequent presenter at industry events. He runs the websites;; and Malcolm is the winner of the 2011 DAMA International Professional Achievement Award.

    He can be contacted at
    Twitter: MDChisholm
    LinkedIn: Malcolm Chisholm

    Editor's Note: More articles, resources, news and events are available in Malcolm's BeyeNETWORK Expert Channel. Be sure to visit today!

Recent articles by Malcolm Chisholm



Want to post a comment? Login or become a member today!

Be the first to comment!