Data Modeling with the Entity-Relationship Model - Database Processing: Fundamentals, Design, and Implementation - page 186

In Figure 5-6(b), the two small circles indicate that the Computer_Assignment relation-
ship is an optional-to-optional (O-O) relationship . This means that an EMPLOYEE need
not have a COMPUTER, and a COMPUTER need not be assigned to an EMPLOYEE. The
Computer_Assignment relationship is thus a 1:N, O-O relationship.
Finally, in Figure 5-6(c), the combination of a circle and a hash mark indicates an optional-
to-mandatory (O-M) relationship . Here an EMPLOYEE must be assigned to at least one
SKILL, but a SKILL may not necessarily be related to any EMPLOYEE. The complete specifica-
tion for the Qualification relationship is thus an N:M, O-M relationship. The positions of the
circle and the hash mark are important. Because the circle is in front of EMPLOYEE, it means
that the employee is optional in the relationship.
By ThE Way Sometimes when interpreting diagrams like Figure 5-6(c) students become
confused about which entity is optional and which is required. An easy way to
clarify this situation is to imagine that you are standing in the diamond on the relationship
line. Imagine looking toward one of the entities. If you see an oval in that direction, then that
entity is optional; if you see a hash mark, then that entity is required. Thus, in Figure 5-6(c),
if you stand on the diamond and look toward SKILL, you see a hash mark. This means that
SKILL is required in the relationship. If you then turn around and look toward EMPLOYEE,
you see a circle. This means that EMPLOYEE is optional in the relationship.
A fourth option, a mandatory-to-optional (M-O) relationship , is not shown in Figure 5-6.
But if we exchange the circle and the hash mark in Figure 5-6(c), then Qualification becomes an
M-O relationship. In that case, an EMPLOYEE need not have a SKILL, but a SKILL must have at
least one EMPLOYEE.
As with maximum cardinalities, in rare cases the minimum cardinality is a specific number.
To represent the relationship between PERSON and MARRIAGE, for example, the minimum
cardinality would be 2:Optional.
Entity-Relationship Diagrams and Their Versions
The diagrams in Figures 5-5 and 5-6 are sometimes referred to as entity-relationship (E-R)
diagrams . The original E-R model specified that such diagrams use diamonds for relationships,
rectangles for entities, and connected ellipses for attributes, as shown in Figure 5-2. You may
still see examples of such E-R diagrams, and it is important for you to be able to interpret them.
For two reasons, however, this original notation is seldom used today. First, there are
a number of different versions of the E-R model, and these versions use different symbols.
Second, data modeling software products use different techniques. For example, Computer
Associates' ERwin product uses one set of symbols, and Microsoft Visio uses a second set.
Variations of the E-R Model
At least three different versions of the E-R model are in use today. One of them, the
Information Engineering (IE) model , was developed by James Martin in 1990. This model
uses crow's feet to show the many side of a relationship, and it is called the IE Crow's Foot
model . It is easy to understand, and we will use it throughout this text. In 1993, the National
Institute of Standards and Technology announced another version of the E-R model as a na-
tional standard. This version is called Integrated Definition 1, Extended (IDEF1X) . 4 This
standard incorporates the basic ideas of the E-R model but uses different graphical symbols.
Although this model is a national standard, it is difficult to understand and use. As a na-
tional standard, however, it is used in government, and thus it may become important to you.
Therefore, the fundamentals of the IDEF1X model are described in Appendix C.
Meanwhile, to add further complication, a new object-oriented development methodol-
ogy called the Unified Modeling Language (UML) adopted the E-R model but introduced
4 Integrated Definition for Information Modeling (IDEF1X) , Federal Information Processing Standards Publication
184, 1993.