I’m doing some investigation into the role of maintenance in software development and the use of metrics within the development cycle.
What I am asking is for a point in the right direction or possibly a direct example in which there has been software failures due to the oversight of maintenance costs.
More specifically I am looking for a case in which an implemented system has failed as the costs/effort for maintenance has been far too great as there has been poor or no projections from metrics.
I understand this is a very specific question, but I’ve been racking my brains thinking of all the software failures I know of and none of them fit the criteria.
3
The U.S. Army’s WARSIM project might be a good example. From this Wikipedia article:
A decade after its original scheduled delivery date, WARSIM has yet to support a single Army training exercise, but is still being funded, largely to vindicate those who conceived of the system and defended it over the lifetime of its development. The WARSIM schedule slipped many times. Moreover, WARSIM has a clumsy architecture that requires enough servers to fill a small room, while earlier “legacy” wargames run efficiently on a single standard desktop workstation.
That same Wikipedia article also talks about the epic failure that was the Denver Airport’s baggage handling system. I’m sure if you research one or both of these projects, you’ll find some good case study material on the subject of cost/maintenance oversights.
3