Information Technology Reference
In-Depth Information
complex data, which requires a good data base and an efficient
querying system. In case of automatic processes, the situation
is much better since we can replay the sequence and tune
the trace configuration. Traceability is often used for change
impact and covering analysis. It is more complex with SPLE
since one modification has an effect on re-usable artifacts,
than on several products. This holds good for other analysis
like covering. The use of MDE to model trace links seems
obvious, it provides automation, and can capture links during
the tool chain. However, it could also permit the tuning of the
granularity of artifacts and the granularity of links needed for
a successful analysis of trace links.
As a result, a traceability framework has to be configured
for each use by the software architect. The architect has to
choose the kinds of artifacts he wants to trace, the dependency
links he needs to observe, and also insert the correct trace
actions in the right places. Such a framework provides the
basic support for storing the chosen artifacts and dependencies
using interface managers. In addition, it is possible to provide
some interactive help for inserting trace actions in the source
code.Theinterestedreadercouldlookattheproposaldescribed
in [ANQ 09].
2.9.5.
Product line evolution
Software evolution is defined as software artifacts evolving
over time due to updates and changes during software
maintenance. The reasons to maintain software assets are
usually classified into four categories: corrective, perfective,
preventive, and adaptive. Change in one asset may impact
several other dependent artifacts and also several products
and probably several versions of the products. Thus, evolution,
anticipated or not, is an important challenge in the
management of a product line. Analyzing the reasons for
changes and the direction of these changes can greatly help in
anticipatingtheevolution.[McG03]analyzeshowevolutionary
Search WWH ::
Custom Search