This paper will provide suggestions on workable alternatives to the full-frontal, brutal process necessary perhaps to elevate legacy software to either a higher level of safety integrity or to "modern" standards (expectations) of software safety. On the one hand, approaching legacy as if it were COTS, in the nominal term, would appear to be an alternative yet none of the COTS safety techniques can be applied. On the other hand, there is neither enough time nor money in legacy programs to accomplish the level of assurance that one expects from new software under today's standards. While this is still a difficult question, this paper offers two approaches and the criteria by which to choose these, or even another, set of processes.
展开▼