THE DATA ADMINISTRATION NEWSLETTER – TDAN.com
ROBERT S. SEINER – PUBLISHER

Business Rules – November 2009

You Need Verbs, Not Just Nouns: About Fact Models

Published: November 1, 2009

Data models and class diagrams have never much focused on verb concepts. That omission is a harmful one. A well-organized business vocabulary not only has noun concepts as represented by terms, but also verb-ish connections between those noun concepts as represented by wordings. These verb-ish connections give structure to basic business knowledge – that is, they represent fundamental connectives in the operational business you need to know and talk about, especially in expressing requirements and business rules. A graphical fact model aids in visualizing these semantics, providing a multi-purpose blueprint to the inherent structure underlying your company’s data and data designs.

Excerpted from Chapter 1, Business Rule Concepts: Getting to the Point of Knowledge (Third Edition), by Ronald G. Ross, July 2009. ISBN 0-941049-07-8, http://www.brsolutions.com/publications.php. Reprinted with permission.

In the human body, structure is provided by the skeleton. The skeleton has two basic components: the bones and the ligaments that connect the bones. Even though the bones are larger and in a sense more basic, both components are essential.

In everyday business operations, structure is given by business vocabulary, or more precisely by the concepts represented in the vocabulary. There is a whole lot more to a vocabulary than you might imagine. Without any exaggeration, a good business vocabulary is no less important to effective business operation than a strong and complete skeleton is to the human body.

Like the skeleton for the human body, a business vocabulary likewise has two components: noun concepts as represented by terms, and verb-ish connections between those noun concepts as represented by wordings. These noun concepts and verb-ish connections are equivalent to bones and ligaments, respectively. They give structure to basic business knowledge – that is, they represent fundamental things in the operational business you need to know about. The terms and wordings given to them, in turn, let you talk about those things in a standard way.

We like to visualize this structure by means of a graphical fact model, representing the skeleton, or blueprint, for the basic knowledge needed in business operations. I will illustrate later.

About Connections Between Noun Concepts and Wordings that Represent Them

Noun concepts can be connected to each other much as ligaments connect bones in the human skeleton. Connections between noun concepts are generally expressed using verbs and verb phrases relating appropriate terms. These noun-and-verb constructions are called wordings – phrases of predictable types that permit sentences, especially expressing business rules, to be made for business operations.

Examples of wordings are given in the table. Note that each wording involves a verb or verb phrase (italicized in the table) to connect relevant terms. Selection of the best verbs and verb phrases to succinctly represent connections between noun concepts is fundamental to building a robust business vocabulary.


Here are some important observations:

1. Wordings extend the common business vocabulary in important ways. Most obvious is that wordings add standard verbs and verb phrases. Less obvious but equally important is that by connecting terms they bring structure to the business vocabulary (think ligaments). For this reason, we like to say structured business vocabulary.
   
   
2. The sample wordings in the table actually represent types of connections, called fact types, rather than individual connections, called facts. For example, for customer places order an actual fact might be Global Supply, Inc. has placed the order A601288. Structured business vocabularies are generally more concerned with identifying fact types rather than actual facts.
   
   
3. A fact type is always worded with a verb or verb phrase (wording); therefore fact types are often called verb concepts.
   
   
4. In English and many other languages, every wording follows a strict subject-verb-object structure – for example, customer places order. The wording thus provides a building block for constructing sentences of arbitrary complexity that unambiguously express business rules or other kinds of knowledge.
   
   
5. A fact type does not imply or establish any business rule on its own, nor does any associated wording. For example, the wording customer places order creates no business rule. It would be inappropriate to express a wording as: A customer has always placed at least one order. This latter statement is more than a fact type – it expresses a business rule pertaining to the fact type.
   
   
6. Verbs (e.g., places) used in wordings do not represent or label any action, task, or procedure per se (e.g., place order). Any such operation represents a different aspect of business operations – the power or “muscle” aspect. Think of a structured business vocabulary as providing the most appropriate way to organize knowledge about the results (or potential results) of such operations. By most appropriate, I mean anomaly-free and semantically-clear. In other words, a business vocabulary organizes what we can know as the results of operational processes or transforms taking place in the business.
   
   
7. A majority of connections of core interest for structured business vocabularies involve exactly two terms – e.g., customer places order. Connections involving more than two terms, however, are sometimes appropriate (e.g., person visits city on date). It’s also possible for a wording to concern only a single term (e.g., person smokes).
   
   
8. In formal logic, each wording represents a predicate. More precisely, a wording represents the meaning of a predicate. Although not directly important for practitioners, this point is a crucial one for engineers and others concerned with tooling and formalisms.
   

Certain important kinds of connections between noun concepts come in handy, pre-defined ‘shapes’. Two of these are illustrated briefly below. These special-purpose elements of structure extend the reach and precision of the structured business vocabulary significantly.


A structured business vocabulary establishes the full scope of potential discourse (in business operations and any supporting systems) in a very fundamental way. If a worker or some automated process produces or expresses knowledge about some other concept or connection not in the vocabulary, we literally have no way to communicate or share such knowledge in a standard and consistent fashion.

Using Graphical Fact Models

You might have noticed that even though structural models are often rendered graphically, no diagrammatic examples have yet been presented. This is not because diagrams are not useful. Just the opposite is true; they are very useful.

Rather, I wanted to emphasize that a structured business vocabulary is first and foremost about what we can know, and about how we can communicate about what we can know. The bottom line is business communication. Knowledgeable workers on the business side must originate and understand the vocabulary.

Getting all the terms and wordings in a vocabulary to fit together as if in some large jigsaw puzzle can be hard. This is where a graphical fact model plays an important role.

When creating a blueprint for remodeling your house, you can quickly see when the pieces are not fitting together. The eye often spots the problems quite easily. A graphical fact model serves a similar purpose in developing a business vocabulary. Just remember, sponsors and business people should sign off on vocabulary – the terms, definitions, and wordings – not on any diagram per se.

Figure 1–1 presents a simple fact model in graphical form. The various connections in this fact model are listed below it. This list includes several that are unlabeled in the diagram – these connections are based on the special-purpose elements of structure mentioned earlier.



Connections in Figure 1–1

Explicitly-labeled:
Unlabeled, based on connector type: