Environmental Engineering Reference
In-Depth Information
•
Technical description of the software re-
lease that is provided for testing or verii-
cation. This should include information on
the overall number of procedures and func-
tions (modules), number of modiied, sup-
plemented and deleted modules in compar-
ison with the preceding release, and also
the modular structure of the software for
the possibility of converting the software
to a state graph. The technical description
is compiled directly by programming engi-
neers, manually as a rule.
tion for the determination of input parameters to
calculate quality metrics (reports on fulfillment
of verification and testing processes, technical
description of the software release, source code
and so forth); information compiler (project man-
ager, system analysts, programmer-managers and
test engineers and so forth); degree of objectivity
(reliability and completeness) of the presented
information.
It should be emphasized that the process of
determining numerical values of input parameters
for calculation of reliability metrics manually is
unfeasible because of complexity and consider-
able amount of initial data. Therefore, the task
of developing (or selection) support tools for
gathering the initial information and automatic
determination of input parameters for the calcula-
tion to metrics is needed. For this reason software
source code analyzers are used as objective sources
of information for determining input parameters
for calculation of software quality metrics.
Based on the examined standards and analysis
of publications (IEEE, 1988, a), (ISO, 2000) and
(Pressman, 1997) a database of reliability indica-
tors (metrics) and tools has been developed, which
allows one to make a choice of indicators with
consideration of the previously listed classifica-
tion features, and also the stage of the lifecycle
at which the reliability evaluation is made, and to
produce their quantitative value. The calculation
of measures is carried out by using deterministic
methods of evaluation described in a standard
(IEEE, 1988, b), while for calculation of predictive
measures various probability reliability models
can be used.
•
The source code and executable code,
which allow one to determine objective
characteristics of the software; number of
lines in the source code, size of the pro-
gram, number of operators and operands
used, and also the total number of their oc-
currence on the program and so forth. The
software source code can be used along
with corresponding utilities to produce a
state graph of software as a whole and of
individual modules. The software presen-
tation in the form of a graph is necessary
for calculating individual metrics and for
developing test benches and executing the
testing itself.
•
Technical documentation for the software,
which includes requirements for the soft-
ware (speciication), technical description
and so forth. In calculating certain met-
rics the production of a comprehensive
result requires that the experts propose
evaluations or assignments of weighting
factors (signiicance) for diferent soft-
ware characteristics, input parameters or
intermediate results. These factors can be
determined based on analysis of corre-
sponding sections of the software technical
documentation.
Evaluation of Software Development
and Verification Processes
An evaluation of software development processes
is accomplished by examination and analysis of
documentation in accordance with the require-
ments for methods, devices and documentation
of development described above (see Figure 4).
Thus, the executed analysis allows one to in-
sert additional classification features for quality
metrics and their parameters: source of informa-
Search WWH ::
Custom Search