Towards a Linguistic Analysis and Representation of
Business Rules
Pieter Joubert
Department of Informatics, University of Pretoria, Lynnwood, Pretoria, South Africa
Abstract. The paper explores linguistics as a basis for analyzing and
representing business rules. The actual business rules are seen as the text to be
analyzed. A very simple, straightforward methodology is explained and illu-
strated with part of a bigger case study. In essence every business rule is ana-
lyzed to determine nouns, verbs, adjectives, adverbs, prepositions and other
lexical categories. The business rules are then structured into sentences which
basically have the structure of conjunctions, subject, predicate, and direct ob-
ject/adjunct.
1 Introduction
Business rules have gained prominence over the last few years. They are seen as
important assets of organisations that should be managed carefully [3]. Business rules
can also be seen as an important (maybe the most important) link between business
and IS. [4]
Business rules:
Are constraints or tests designed to maintain integrity of data (Ross, 1997 in [2]).
Are statements that aim to influence or guide behaviour and information in an
organisation [5]
Define how the business is actually run [2].
Define or constrain some aspect of a business [6].
Determine business structure [6]
Influence the behaviour of an organisation [6].
Are statements that influences business behaviour towards desired objective
(Plotkin in [2])
Are assertions that constrain patterns of the enterprise behaviour. (Morabito, et al,
2001 in [4])
Some characteristics of business rules are [4]:
They exist in various forms, from simple to very complex and dynamic.
They can originate internally, mostly derived from strategic processes, or exter-
nally, from government, industry or a specific profession.
They can be based on explicit or tacit knowledge.
Joubert P. (2009).
Towards a Linguistic Analysis and Representation of Business Rules.
In Proceedings of the Joint Workshop on Advanced Technologies and Techniques for Enterprise Information Systems, pages 105-113
DOI: 10.5220/0002217601050113
Copyright
c
SciTePress
They can be found in documents, procedures, policies, regulations, user manuals,
information systems, etc.
They are normally expressed in the form of principles, procedures, facts, figures,
rules, formulas, etc.
Explicit business rules are a manifestation of a richer underlying knowledge.
The purpose of this paper is to show that that a “guided” linguistic analysis can
help to analyse and represent business rules. The methodology does not require the
modellers to have a deep understanding of linguistics. A straight-forward, common-
sense approach to linguistic analysis will be followed, so that this methodology is
usable by non-linguists.
To show this the following will be done: Firstly, a very basic linguistic analysis
methodology will be proposed and illustrated by analysing and representing a number
of business rules from various sources. Secondly, the linguistic representations of the
business rules will be discussed to indicate how this analysis contributes to better
understanding the business rules.
2 A Basic Linguistic Analysis Methodology
2.1 Introduction
The methodology suggested in this section combines a morphological and syntactical
analysis. A semantic and pragmatic analysis can also be included but has been ex-
cluded from this study because of space restrictions.
Every business rule can be seen as linguistic text to be analysed. In essence the
methodology takes every business rule and breaks it up in the following fixed parts:
A conjunction linking related clauses.
A subject indicating the person or thing which the clause is about.
A predicate describing what the subject did, what action was done to the subject
or what state of existence the subject is in.
A direct object/adjunct indicating persons or things affected by the action of the
verb.
For instance the business rule “each policy must have an expiry date” can be
represented in structured sentence format as:
Conjunction Subject Predicate Object/Adjunct
(Each) Policy (Must) have Expiry date
Or diagrammatically
Fig. 1. Diagrammatical representation of a business rule.
106
2.2 The Steps in more Detail
The steps involved are listed below and business rule “IF acceptance of renewal
notice from insured is not received within 14 days THEN send reminder” will be used
to illustrate the process:
1. Identify verbs
: Take each business rule and identify the verbs or verb phrases.
Look for both action and existence verbs. (Action verbs portray actions, e.g. He
walked
slowly forward; while existence verbs indicate states of existence, e.g. He
is
a man.) There will normally be one row in the resulting table for each verb
identified in the business rule. Make verbs infinitive, present tense where possible,
for instance walked and walks become walk. Place auxiliary parts of a verb phrase
in brackets. Place one verb only in the Predicate column.
Two verb phrases can be identified in the example business rule: is not received
and send. Place these two verb phrases in the Predicate column of two subsequent
rows. Make verbs infinitive, present tense, i.e. is not received becomes (Not) receive.
Conjunction Subject Predicate Direct Object /Adjunct
(Not) Receive
Send
2. Identify nouns: Place one noun only in the second Subject, Direct Ob-
ject/Adjunct column as applicable. (A noun denotes persons, places or things
that we can either perceive by our senses or conceive in our minds.) Make all
nouns singular and show the plural parts in brackets. There does not always have
to be an explicit subject (although there is always an implicit one). These are
nouns that are directly or indirectly the agent or doer of an action. Most human
nouns are. There must be a direct object or adjunct or both. If there is more than
one adjunct they must follow on subsequent rows. Simplify clumsy noun phrases,
for instance, courses that can be offered should rather be offered course; and
courses that can be scheduled should rather be scheduled course. Note that the
two examples in the previous sentence implies that out of all possible courses,
only some will be offered and only some will be scheduled.
In the example business rule neither verbs have subjects specified. For the first
verb receive place the direct object acceptance in the second Direct Object/Adjunct
column. For the second verb phrase send the direct object is a genitive case noun pair
that needs to be resolved as shown in the next step.
3. Identify all direct and indirect genitive case noun pairs. Genitive case nouns can
be identified by the following means: (1) the proximity of nouns, for instance,
room equipment implies the room’s equipment; (2) special words like their and
for, for instance, rooms and their equipment and equipment for rooms also implies
room’s equipment; and (3) nouns with an apostrophe s, for instance, the room’s
equipment. Translate all these genitive case noun pairs into the basic genitive for-
mat, noun of noun, for instance, equipment of room.
In the example business rule acceptance of renewal notice becomes acceptance
(of) renewal notice. At this stage the table looks as follows:
107
Conjunction Subject Predicate Adjunct
(Not) Receive Acceptance (of)
Renewal Notice
Send Reminder
4. Identify adjectives: All adjectives must be placed within brackets together with
their referent nouns. (Adjectives specify the attributes of a noun or pronoun e.g.
The tall
girl danced.)
In the example business rule the 14 in 14 days is an adjective describing the noun
days.
5. Identify prepositions
: If a preposition is followed by a noun or a noun phrase it is
part of the Adjunct (and placed in the first column of the Direct Object/Adjunct
section), however, if it is followed by a verb or verb phrase it becomes a new row
and the preposition (or its conjunction equivalent) is placed in the Conjunction
column. (Prepositions indicate a semantic relationship between entities, for in-
stance: (1) location of one entity in relation to another, e.g. the book is on / under /
above / below / near the bookshelf; (2) direction, e.g. he travelled from his house
to work; and (3) accompaniment, e.g. with/without salt.)
In the example business rule, related to the first verb are two prepositions: from
and within. Specify on the next line in the first direct object/adjunct column the
prepositions and in the second column the nouns on subsequent lines. At this stage
the table looks as follows:
Conjunction Subject Predicate Adjunct
(Not) Receive Acceptance (of)
Renewal notice
From Insured
Within (14) days
Send Reminder
6. Identify conjunctions: If you have more than one predicate per business rule, the
resulting clauses are many times linked to each other. The conjunction column is
used to link the various rows together.
In the example business rule the conjunctions IF and THEN are both followed by
sentences and are therefore placed in the Conjunction column.
Conjunction Subject Predicate Adjunct
If (Not) Receive Acceptance (of)
Renewal notice
From Insured
Within (14) days
Then Send Reminder
7. Identify adverbs. All adverbs must be adjusted so that they can be placed into the
adjunct. For instance, for the phrase double booked place book in the Predicate
108
column and double in the Adjunct column. (Adverbs modify verbs (he sang
loudly), adjectives (a very tall building), other adverbs (unbelievably quickly) and
sentences (sadly, he died).)
There are no adverbs in the example business rule.
8. Identify pronouns: For the lexical analysis ignore pronouns and replace them with
the nouns or noun phrases that they take the place of (their referents). For in-
stance, in the phrase instructors, and their availability, the pronoun their refers to
instructors and should be replaced by that. (Pronouns are words that are usually
used in place of nouns or noun phrases, e.g. she
looked him in the eye. The noun
or noun phrase that is replaced by a pronoun is called the referent (or antecedent)
of the pronoun.)
9. Handle lists: A list separated by commas implies that everything applying to the
first item in the list also applies to the rest of the items.
10. All other lexical categories: Ignore all other lexical categories.
The final representation of the business rule is:
Conjunction Subject Predicate Object/Adjunct
If (Not) Receive Acceptance (Of)
Renewal notice
From Insured
Within (14) day(s)
Then Send Reminder
This structured sentence can also be represented diagrammatically as follows (All
noun phrases are placed in rectangles, verb phrases in ellipses, conjunctions in di-
amonds, prepositions are on the lines linking it):
Fig. 2. Diagrammatical representation of example.
Specifying the business rule in this way clearly shows the structure of the business
rule, missing information and the key components of the business rule:
This business rule in essence is an action condition which, if not true, is followed
by another action. It can be stated as follows: IF (NOT) ACTION THEN
ACTION.
109
Both actions lack a subject, which can lead to questions by the analyst like “who
receives the acceptance” and “who must send the reminder”, ensuring more com-
plete business rules.
The fact that the business rule is involved in 4 objects (Acceptance of renewal
notice, insured, days and reminder) is also very clear. It is possible to group busi-
ness rules by object, for instance, by reminder.
A basic logical check on the verb can help to ensure a more complete business
rule. For instance, the verb receive, leads to the following questions (based on the
Zachman questions plus other questions), what
is received, who receives it, from
whom do we receive, when must it received, where must it be received, why is it
received. In this example, what = acceptance, when = within 14 days, from whom
= insured. These questions can be asked for any action.
2.3 A more Extended Example
The following example (adapted from [6]) will be used as a more extended example:
EU-Rent has 1000 branches in towns in several countries. At each branch cars,
classified by car group, are available for rental. Each branch has a manager and
booking clerks who handle rentals.
Most rentals are by advance reservation; the rental period and the car group are
specified at the time of reservation. EU-Rent will also accept immediate (‘walk-in’)
rentals, if cars are available.
At the end of each day cars are assigned to reservations for the following day. If
more cars have been requested than are available in a group at a branch, the branch
manager may ask other branches if they have cars they can transfer to him/her.
Diagrammatically this example can be represented as follows:
110
Fig. 3. Diagrammatical representation of EU-Rent example.
2.4 A Discussion of the Resulting Analysis and Representation
The above analysis and representation provides the following results fairly straight-
forward:
Existence verbs can indicate static relationships between objects. For instance, the
has in business rules 1, 2 4 and 5, can be represented in an ERD as follows:
Fig. 4. Static relationships based on existence verbs.
111
Existence verbs can also indicate static relationships between objects and their
attributes. For instance, the are in business rules 9b and 11b, can be represented in
an ERD as follows:
Fig. 5. Object attribute relationships based on existence verbs.
By just listing all of the objects (as represented in blocks) a quick overview of the
entities involved in the business rules can be identified.
All subjects where the predicate is an action verb provide a list of all potential
actors (in for instance, use case modelling). In the example business rules 6, 9a,
11c, 11e gives the list Booking Clerk, EU-Rent, Branch Manager and Branch can
be deduced. Business rules 8, 10 and 11a are action predicates without subjects
which then will logically lead to asking questions to determine the corresponding
actors.
3 Conclusions
This paper shows that even a very basic linguistic analysis can help towards firstly
understanding and analysing a business rule better; and secondly, representing it
better. Based on this initial analysis further more advanced analysis can be done much
easier.
This paper gives only an indication of what is possible with linguistic analysis and
does not constitute completed research. Furthermore, at this stage the methodology
only considers morphological and lexical analysis. Semantic and pragmatic analysis
can be addressed in further research and will expand the richness of analysis even
further.
References
1. Bajaj, A. and Rockwell, S. 2004. COGEVAL: A propositional framework based on cogni-
tive theories to evaluate conceptual models. Ninth CAiSE/IFIP8.1/EUNO International
Workshop on Evaluation of Modeling Methods in Systems Analysis and Design
(EMMSAD04).
2. Steinke, G., Nickolette, C. Business rules as the basis of an organization's information
systems Industrial Management & Data Systems. (2003) Vol. 103, No 1, 52 – 63.
112
3. Ram, S., Khatri, V. A comprehensive framework for modeling set-based business rules
during conceptual database design. Information Systems, Vol. 30, Issue 2, (2005), 89-118.
4. Bajec, M., Krisper, M. A methodology and tool support for managing business rules in
organizations. Information Systems, Vol. 30, Issue 6, (2005), 423-443.
5. Von Halle, B. The business rule roadmap. Database Programming and Design archives,
(2000) available at www.dbpd.com/vault/9710arch.htm (last accessed 24 July 2005)
6. Hay, D., Healy, KA. Defining business rules-what are they really. White Paper, The Busi-
ness Rules Group, (2000).
113