Information Technology Reference
In-Depth Information
To reduce the complexity of validating a large software system, testing is usually
performed in stages. In the first stage of testing, each individual module of the system is
tested independently. It is easier to isolate and fix the causes of errors when the number
of lines of code is relatively small. After each module has been debugged, modules are
combined into larger subsystems for testing. Eventually, all the subsystems are combined
in the complete system. When an error is detected and a bug is fixed in a particular
module, all the test cases related to the module should be repeated to see if the change
that fixed one bug accidentally introduced another bug.
Successful software systems evolve over time to meet the changing needs of their users.
The evolution of a software system resembles the creation of a software system in many
ways. Software engineers must understand the needs of the users, assess the strengths
and weaknesses of the current system, and design modifications to the software. The
same CASE tools used to create a new software system can aid in its evolution. Many of
the datasets developed for the original system can be reused when validating the updated
system.
There is evidence that the field of software engineering is becoming more mature (Fig-
ure 8.10). The Standish Group [56] regularly tracks thousands of IT projects. As recently
as 1994, about one-third of all software projects were canceled before completion. About
one-half of the projects were completed but had time and/or cost overruns, which were
often quite large. Only about one-sixth of the projects were completed on time and on
budget, and even in these cases the completed systems often had fewer features than
originally planned. Another survey by the Standish Group in 2009 showed that the prob-
ability of a software project being completed on time and on budget had doubled, to
about one in three. Only about one-quarter of the software projects surveyed were can-
celed. Slightly less than half of the projects were late and/or over budget, but the time and
1994
2009
On time/
On budget
Canceled
Canceled
On time/
On budget
Time/
Cost overruns
Time/
Cost overruns
FIGURE 8.10
Research by the Standish Group reveals that the success rate of IT projects in
2009 was twice that of 1994. Today, about one-third of software projects are completed on
time and on budget.
