8.1 Overview of Results
This exercise has demonstrated the achievement of a detailed mapping
from CMMI Continuous Representation to the Process Reference Model of
ISO 15504-2. All significant elements of the model have been successfully
addressed, and the exceptions were in areas that were expected. The
mapping has led to significant insights into the structure and consistency
of the CMMI, and has some impact on the overall goal of this project.
Some problems and anomalies have been identified; these have been reported
to the CMMI Product Development Team in the form of Change Requests.
The mapping was extended to also address the Staged Representation of
the CMMI. Some additional problems in relation to the inclusion of expressions
of attributes of process capability were noted. From the Staged Representation
mapping it has been possible to identify standard Process Profiles that
correspond to the levels of maturity expressed in the Model.
These mapping exercises provide a foundation for the development of
a translation mechanism to enable the expression of CMMI assessment
results into standard ISO 15504 Process Profiles. The Translation Mechanism
will need to be validated through field studies. Initial studies will
use results of assessments conducted previously as a basis for generating
ISO 15504 Process Profiles. As opportunities arise, these will be extended
to translation during the period of the assessment.
The principal impact of use of the translation mechanism will be to
increase the cost and duration of assessments in which translation is
required. This can be mitigated through use of a sub-team with specific
ISO 15504 knowledge and understanding to conduct the translation. Since
there is already consensus on the observations, there should be no additional
data to consider and the involvement of the entire team should not be
necessary.
8.2 Future Work
The primary initial concern is to progress the work towards the production
of a formal statement of conformance of CMMI to ISO 15504-2. In order
to achieve this, the translation mechanism proposed must be validated
through actual use. In order to achieve this, an appropriate tool set
that will make use of the translation mechanism feasible is needed.
A high-level specification for this tool has been defined, and a prototype
version is being developed. Without tool support, it is not considered
that performance of translation will be feasible. Once the translation
mechanism is demonstrated to be accurate and feasible, the formal statement
of conformance can be issued.
If the benefits from this study are to be effectively deployed, further
work will be required. The most obvious issue is the forthcoming revision
of ISO 15504. When this is published (late in 2002) the existing mappings
to ISO/IEC TR 15504-2 will become irrelevant; under the new version,
the Process Reference Model for Software Engineering will be drawn from
the Amendment to ISO 12207. The mapping will therefore have to be repeated
against ISO 12207. The existing results will be of substantial value.
Given the timeframe for the proposed release of Version 1.1 of CMMI,
the best option will be to map the new model against the revised standard.
The opportunity also exists to extend the scope of mapping to address
systems-specific issues, through use of ISO 15288 (when published) as
reference model for this domain. We have already indicated that some
of the scope issues in the current mapping are addressed in ISO 15288.
|