About this series of articles

While my first series of articles on writing natural language rule statements^[1] explored a wide variety of issues in a rather organic and hence random manner, this series takes a more holistic and systematic approach and draws on insights gained while writing my recently-published book on the same topic.^[2]  Rule statements recommended in these articles are intended to comply with the Object Management Group's Semantics of Business Vocabulary and Business Rules (SBVR) version 1.0.^[3]

The story so far

In previous articles in this series (see the "Language Archives" sidebar) we have looked at standardised rule statements for various types of operative rules:^[4]  data rules,^[5] activity rules,^[6] and party rules.^[7]  Each specific type of rule statement has a common formulation, which we have discussed in both a relatively informal way and by way of rule statement patterns.

In this article, we will look at the various forms that rule statements for definitional rules^[8] can take.  You may recall that, in the previous article in this series,^[9] I proposed that such statements should be distinguished from operative rule statements by the former's use of the phrase "by definition" after the subject of the rule statement.  First though let's review the various types of definitional rule.

Types of definitional rule

Definitional rules can be categorised as:

1. formal term definitions, which define business terms formally.


2. categorisation scheme enumerations, which define members of categorisation schemes.


3. category transition constraints, which define constraints on the transitions between categories in a categorisation scheme.


4. valid value definitions, which define the valid values of measures.


5. standard format definitions, which define standard formats for data items.


6. data calculation rules, which define algorithms or formulae for named quantities or conversion factors between units.


7. complex concept structure rules, which define constraints on the components of complex concepts.

Formal term definitions

A formal term definition defines a particular business term in a formal manner:

1. A formal intensional definition defines the subject business term using an intensional definition: one that cites both a hypernym (a term that signifies a superset of the set signified by the original term) and the characteristics that distinguish members of the set signified by the original term, e.g.,

2. A formal extensional definition defines the subject business term by using an extensional definition:  one that lists a complete set of hyponyms (terms that signify subsets of the set signified by the original term), e.g.,

3. A symbolic literal definition defines the subject business term using one or more literals, e.g.,

One question that is often put to me is "why do we need formal term definitions?" since the business glossary or fact model should contain the definitions of all terms.  However, there is no redundancy if you consider formal term definitions to be merely renderings of (<term>, <definition>) pairs in the form "A <term> is by definition <definition>."  In fact, an important role for such definitions (whether they reside in the business glossary or fact model or as rule statements) is to reduce redundancy.  For example:

1. There may be multiple rule statements that refer to the close of business each day, and there is the possibility that that closing time may change.  If each of those rule statements refers explicitly to 5pm, each of them would need to be changed, whereas if each of those rule statements refers to close of business, only rule statement RS241 needs to be changed.


2. There may be multiple rule statements that restrict parties to transactions to immediate family members of a person, and there is the possibility that the definition of immediate family member might change.  If each of those rule statements refers explicitly to the list of related persons, each of them would need to be changed, whereas if each of those rule statements refers to immediate family member, only rule statement RS238 needs to be changed.

Categorisation scheme enumerations

A categorisation scheme enumeration defines the members of a categorisation scheme that is both mutually exclusive (i.e., no individual object can be a member of more than one category in the categorisation scheme) and jointly exhaustive (i.e., each member of the class with which the categorisation scheme is associated belongs to a category in the categorisation scheme).

One way of expressing this type of rule is in terms of subsets of the categorised term, by way of a rule statement very similar to that of a formal extensional definition (already discussed above), e.g.,

The difference between these rule statements and formal extensional definition statements is because the categories listed in these rule statements are mutually exclusive (whereas those in formal extensional definition statements do not have to be).

Another way of expressing this type of rule is in terms of allowed values of a category attribute, e.g.,

One use of categorisation scheme enumeration statements of the second type (e.g., RS247 and RS248) is as a Source of Truth for value set rules, such as:

This situation (the co-existence of RS248 and RS249) prompts the question "do we need both?"  I find that if there are multiple rules governing specification of a category attribute in different circumstances, verifying value set rule statements against a single categorisation scheme enumeration statement such as RS248 is easier and less error-prone than comparing with other value set rule statements, particularly as some of the latter may demand only a subset of the possible categories.

An important function of categorisation scheme enumeration statements of the first type (e.g., RS243 – RS246) is to support quality assurance of rule statements.  For example, since domestic flight and international flight are the only subtypes of flight, "for a domestic flight or international flight" is redundant in:

By contrast, since return journey and multi-stop journey are not the only subtypes of journey, "for a return journey or multi-stop journey" is not redundant in:

Further, since one-way journey, return journey, and multi-stop journey are the only subtypes of journey, RS252 is logically equivalent to RS251:

Category transition constraints

A category transition constraint specifies allowed or disallowed transitions between categories or statuses.  For example, if we only allow the values never married, married, widowed, and divorced for the marital status attribute,^[11] only the following transitions are possible:

• never married to married but not vice versa
• married to widowed and vice versa
• married to divorced and vice versa.

This state of affairs could be expressed using the following rule statements:

The inference to be drawn from this set of statements is that any other transitions are possible.  Alternatively, of course, one could list the five possible transitions, with the inference that any other transitions are impossible.

A better alternative is to define the circumstances in which transition to or from a particular status is possible, as in the following rule statements:

Constraints such as this document our understanding of the real world in which the organisation operates rather than the data it uses to record that real world.  For example, if we record the marital status of employees, you might think that there is a state transition constraint (as described in Part 12^[12] in this series) corresponding to each of the rule statements RS258 – RS260.

However, data may be incomplete or not up to date.  This means that:

• an additional unknown data value is required, and
• a direct transition from never married to widowed or divorced is possible for employees who have not notified their employers that they have married but have then notified those employers of their subsequent change of status.

Therefore, while there is a state transition constraint (RS132) corresponding to RS258, there are no state transition constraints corresponding to RS259 or RS260.

Valid value definitions

A valid value definition defines the valid values of a particular measure, as a range or (occasionally) as a list of discrete values, e.g.,

One use of this type of definitional rule statement is as a Source of Truth for range rules, such as:

While 0 – 14 is the absolute range of possible values for pH, reasonable ranges for water samples from the environment are considerably narrower:  for example, 4.3 – 6.3 for precipitation (rain water), 6 – 8 for "streams" (rivers and other inland watercourses), 7.5 – 8.4 for seawater.  This presents us with a dilemma, since the occasional sample may exhibit pH values outside these ranges, so rule statements such as the following are not strictly correct:

The SBVR uses the concept of enforcement level (or level of enforcement) as a property of operative rules to distinguish between those that are strictly enforced (cannot be violated without consequence in any circumstance) and those for which violations are allowed (with different enforcement levels defining different actions to be taken in the event of a violation).  While RS262 should be strictly enforced, violations of RS263 – RS265 should be catered for, in such circumstances as extreme environmental conditions, faults in measuring equipment, or errors in analysis.

However, the co-existence of rule statements RS262 – RS265 presents us with a further dilemma.  Each of RS263 – RS265 implies RS262, so if the rule book is to be non-redundant (which it should be) RS262 should be removed.  Yet there still ought to be a rule preventing pH values outside the absolute range from being entered in any circumstance.

In situations such as this, business stakeholders often suggest we replace 'must' by 'should' in rule statements RS263 – RS265.  However, whereas 'must' can be taken to imply that some standard error handling occur when a non-compliant value is supplied, it is not clear what should happen when data doesn't comply with a 'should' rule statement.

My question to stakeholders in such situations is therefore "what should happen if a value outside the relevant reasonable range is supplied?"  Initially the answer is that there should be a message from the system receiving the data, along the lines of "are you sure?"  I then suggest that the user should not simply reply "yes" to the "are you sure?" message, but supply additional information to support the out-of-range value.  This allows verification of this data to be managed with rule statements such as the following:

These rule statements replace RS263 – RS265, but RS262 is retained.

To be continued...
The next article in this series will look at the remaining types of definitional rule, as well as patterns for all definitional rule statement types.

References

[1]  The first of which is:  Graham Witt, "A Practical Method of Developing Natural Language Rule Statements (Part 1)," Business Rules Journal, Vol. 10, No. 2 (Feb. 2009), URL:  http://www.BRCommunity.com/a2009/b461.html  

[2]  Graham Witt, Writing Effective Business Rules.  Morgan Kaufmann (2012).  

[3]  Semantics of Business Vocabulary and Business Rules (SBVR), v1.0.  Object Management Group (Jan. 2008).  Available at http://www.omg.org/spec/SBVR/1.0/
     The font and colour conventions used in these rule statements reflect those in the SBVR, namely underlined teal for terms, italic blue for verb phrases, orange for keywords, and double-underlined green for names and other literals.  Note that, for clarity, these conventions are not used for rule statements that exhibit one or more non-recommended characteristics.  

[4]  A rule that states what must or must not happen in particular circumstances, and which can therefore be contravened, by contrast with a definitional rule.  

[5]  A rule that constrains the data included in a transaction (a form or message) or a persistent data set (e.g., a database record).  

[6]  A rule that constrains the operation of one or more business processes or other activities.  

[7]  A rule that makes a distinction between different parties or the roles they play.  

[8]  A rule that defines a construct created or used by an organisation (or the industry within which it operates), or defines a property of such a construct, and which cannot therefore be contravened, by contrast with an operative rule.  

[9]  Graham Witt, "Writing Natural Language Rule Statements — a Systematic Approach — Part 18:  Operative and Definitional Rules," Business Rules Journal, Vol. 15, No. 1 (Jan. 2014), URL: http://www.BRCommunity.com/a2014/b740.html  

[10]  This is the case in Australia; other dates apply in other countries.  

[11]  I have omitted the de facto and separated values to simplify the discussion, as treatment of these values varies according to jurisdiction.  

[12]  Graham Witt, "Writing Natural Language Rule Statements —a Systematic Approach —Part 12:  Spatial Data Constraints, Data Update Rules," Business Rules Journal, Vol. 14, No. 6 (June 2013), URL: http://www.BRCommunity.com/a2013/b707.html  

[13]  Any rule statement that is marked with an initial dagger is, for one or more reasons, not recommended; in this case a preferred alternative will also be shown.  

# # #