Mining Finite-State Automata with Annotations - Mining Software Specifications: Methodologies and Applications - page 47

TABLE 2.3: A Pattern of Data Reuse

# Events Attributes Actual

Values

GO RI

1 hasNext

2 next return.availSeats 5 1 0

3 getAvailableSeats return 5 1 1

4 add flight.availSeats 5 1 1

5 hasNext

6 next return.availSeats 7 2 0

7 getAvailableSeats return 7 2 1

8 add flight.availSeats 7 2 1

9 hasNext

10 next return.availSeats 8 3 0

11 getAvailableSeats return 8 3 1

12 add flight.availSeats 8 3 1

13 hasNext

14 next return.availSeats 9 4 0

15 getAvailableSeats return 9 4 1

16 add flight.availSeats 9 4 1

17 hasNext

representation. For example, Table 2.3 shows the events generated in an exe-
cution that differs from the one reported in Table 2.2 for the number of times
the same sequence of operations is executed. The extra iteration is classified
as a new sequence by the global ordering rewriting strategy due to the new ab-
stract symbol that is generated to rewrite it. On the contrary, the relative on
instantiation rewriting strategy rewrites values according to the same schema.
Thus, the relative to instantiation rewriting strategy successfully captured the
common behavioral patterns represented in the traces.

2.5.1.4 Access Rewriting Strategy

The relative to instantiation rewriting strategy can deal with the gener-
ation and subsequent use of new concrete values, but does not handle well
the case of multiple production of concrete values. This case is exemplified in
Table 2.4, which presents three sequences of events that reuse the same at-
tribute values multiple times across iterations. We can observe that symbols
in column RI in Table 2.4 do not fully capture the repeated pattern of values
that are generated and used multiple times. The relative to access rewriting
strategy is designed to capture this case.
The relative to access rewriting strategy replaces the first occurrence of a
concrete value with 0, and the subsequent occurrences with a number that