Jussi Koskinen, Heikki
Lintinen, Henna Sivula & Tero Tilus
Evaluation
of Software Modernization Estimation Methods Using NIMSAD Meta Framework
Publications of the Information Technology Research
Institute 15, 2004. Jyväskylä, Finland: University of
Jyväskylä.
This report charts and compares some of the most promising
methods and approaches available for 1) estimating the profitability of
software modernizations and for 2) supporting the actual modernizations. Thus,
we’re concerned both with providing arguments for informed decisions regarding
software modernizations, and charting effective technical possibilities to
support the actual modernizations. In this report the focus is on the first
objective. Profitability is affected by benefits, risks, and costs. There
exists multiple approaches for evaluating these issues, including strategy
selection of legacy system evolution and modernizations. We shall provide a
comparison of 12 well-known approaches. The sample includes 6 general
approaches suitable for strategic-level decision support, 2 risk evaluation based
approaches, and 4 approaches for cost estimation. The evaluation is organized
based on NIMSAD meta framework for evaluating methodologies. Reliable
estimation of (long-term strategic-level) benefits especially appears to be
difficult. Risk-based evaluation approaches seem practical bases for process
improvements. Cost estimation has its peculiarities on software maintenance
field, which should be taken into account. Empirical validation of most of
these models, unfortunately, is either relatively weak or non-existent. We feel
that an appropriate process development strategy includes improvement of the
existing expert decision making (regarding software modernizations) iteratively
in cooperation with software industry including the explication of risks and
possibilities involved with modernizations. We will also briefly survey 10 main
branches of approaches for actual modernizations.
Updated: Sept. 28, 2004
by Jussi Koskinen.