> Tucker Taft wrote:
> > When looking at a cost/benefit analysis, it is always easiest to
> > see the benefits of a favored solution, and the costs of a disfavored
> > solution.
> I haven't seen the report yet, only various summaries, but your reference
> to "cost/benefit analysis" has me wondering about the methodology used in
> the study. As I understand it, the Ada policy (and Ada itself) was intended
> to foster lower total life-cycle costs, even at the expense of higher initial
> development costs. However, your description of why Ada may not be appropriate
> for some domains seems to emphasize time-to-market and development cost
> factors only. How extensively does the report discuss life-cycle cost
> comparisons? Does it make assumptions about where the long-term maintenance
> of software will be performed (depot vs. contractor vs. COTS vendor), for
Life cycle costs were a major factor in shaping the nature of the
proposed Ada requirement. If the DoD is not directing maintenance,
then there is generally no specific Ada requirement, since the product is
presumed to be a commercial product, with the commercial supplier spreading
the life-cycle maintenance burden over a largely non-DoD customer base.
We did not distinguish depot vs. contractor, but rather DoD-directed
vs. standard "commercial" maintenance. The use of depot maintenance
seems to be somewhat up in the air in general, but if the DoD is
paying, the committee felt that Ada provided advantages independent
of who performed the maintenance.
Life cycle costs also helped determine the split between mandated
and non-mandated domains. In non-"DoD-dominated" domains, the bulk
of the "value" of the system is presumably in commercially-based
technologies, with commercial-technology-based maintenance, even
if DoD is directing the maintenance (in the sense of prioritizing
which bugs get fixed first). Where the goal is to capitalize on
commercial investment and infrastructure, the committee felt that
the benefit of specifying a particular 3GL was not worth the costs
of being cut off from maintenance infrastructure investments made
based on other technologies.
Finally, life cycle costs were a major driver for the committee's
push for a more encompassing Software Engineering review process
to replace the waiver process. It was felt that
the waiver process did not necessarily involve
all the relevant "stakeholders," and in particular did not necessarily
involve the organizations responsible for eventual support and maintenance
of the systems. The Software Engineering Plan Review Board is required
to include representatives of the long-term maintenance organization,
as well as post-deployment support.
> By the way, I was stunned to see that the November 4 issue of Aviation Week
> and Space Technology already has an article on the Ada study. The punchline of
> the article: "The report said that the Defense Dept. should either provide
> that support [$15M/yr] or drop the Ada requirement entirely." Given the
> current budget situation, I'd say that decision has already been made...
The committee made a strong case that the return on investment of
the $15M/year would be very good, but that without it, the Ada policy
would be non-functional. Our sponsors seem to be taking this part
of the recommendation very seriously, and apparently have already moved
to restore some of the funding that was being removed from the AJPO.
Note that the $15M was not primarily designed to "prop up" the Ada market,
but rather to represent an ongoing DoD investment in improving and enhancing
its own use of Ada as a critical defense technology.
> LMTAS - "Our Brand Means Quality"
> For more info, see http://www.lmtas.com or http://www.lmco.com
-Tucker Taft [log in to unmask]