Information Technology Reference
In-Depth Information
necessary to assess its adequacy, e.g. how well-suited it is for generating classes
and attributes in a class model. The definition and application of metrics using
the i MDF M method [ 20] is a possible path to follow.
Traceability among models. Traceability is a classical problem in MDD methods
[ 52] and as such it needs to be properly managed.
6 Conclusion
In this chapter, we have defended the position of shifting the focus of the i com-
munity from pure research to a more practical view that may help in transferring
the framework to practitioners. For the sake of brevity, we have focused on four
challenges that have been described in detail, but several others are out there. These
challenges solve some of the drawbacks that the empirical study by Estrada et al.
has pointed out [ 18] . We have focused on scientific challenges, but there are also
some other community-oriented issues worth to be considered, among them we may
mention:
Lessons learned. For putting i into practice, it is needed to have feedback about
its use by practitioners. We think that the facts observed in the experiences
mentioned at the introduction of this chapter should be consolidated by all the
participants in these and others collaborative experiences, packaged into lessons
learned, and the consequences fed back into the community as a main driver for
identifying lines of future research.
Tool support. In our opinion, taking into account the size of the core i research
community, there is an excess of tools. Given that developing and maintaining
such tools has a considerable cost for the research groups, a possible strate-
gic action could be to join efforts for producing a common subsystem at least
for the more basic capabilities (e.g., i editor) configurable enough to adapt to
each group's specificities, importing/exporting models in e.g. iStarML format
and with a well-defined API, with a plug-in based infrastructure for enriching its
functionality.
Population for experiments. A great deal of current proposals of modelling vari-
ations, analysis techniques, development methods, etc., undergo through a weak
validation (if any). The main reason for this probably is the difficulty on get-
ting population enough to run these experiments. A community-oriented view is
probably needed in order that the research groups allocate some of their effort
in participating in such validations. The i wiki may help on implementing this
idea.
We hope to see in the next years an increasing effort in these and other topics that
make the use of the i
framework in industrial cases not an exception but a usual
practice.
Acknowledgments This work has been partially supported by the Spanish project TIN2007-
64753
Search WWH ::




Custom Search