From mboxrd@z Thu Jan 1 00:00:00 1970 X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on polar.synack.me X-Spam-Level: * X-Spam-Status: No, score=1.9 required=5.0 tests=BAYES_50,INVALID_DATE, LOTS_OF_MONEY autolearn=no autolearn_force=no version=3.4.4 Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!ucbvax!grebyn.com!karl From: karl@grebyn.com (Karl Nyberg) Newsgroups: comp.lang.ada Subject: More on AFATDS Message-ID: <8903081532.AA09156@grebyn.com> Date: 8 Mar 89 15:32:55 GMT Sender: daemon@ucbvax.BERKELEY.EDU Organization: Grebyn Corporation List-Id: [Ed - Forwarded] AFATDS is about 1.2 million source lines. For more information about AFATDS, check the 6 March issue of Federal Computer Week for the letters column. I replied to their recent article about an Army comment on 'early management problems' due to the contractor. Since I was a member of Magnavox' original AFATDS team, I have a great deal of knowledge about AFATDS. AFATDS and Magnavox, the developer, have been trying to get through the FQT hoop for three years. Obviously, the tests have been limited in some way to get past a milestone. This kind of thing is typical in large enterprizes of all kinds when they get in trouble. What is interesting is, that this version of the program is the basis for all future develoment. One wonders how much future budget will be expended for 'improvement' which is actually spent on making this monster work according to requirements. Congress has funded AFATDS Army program management shop for about $12M annually for three years as an R&D project, while Magnavox has been burning its own money (since April 1986). Its chief Congressional opponent was Rep. Chappel, who lost his bid for re-election last fall. Whether the new Congress is willing to buy into the Army's committment to AFATDS is an open question. Probably the greatest difficulty AFATDS faced is the immaturity of the state of software engineering technology and management practise. Also, the Army's support contractors which assist it in managing its acquisition, are the pits. If a contractor is trying something new, it will be held to the letter of its contract, no matter how dumb it is. Ada was a technical challenge, particularly using the early Telesoft compiler, with a VAX host and M68020 target. Another gotcha was not having significant debugging available on our custom 3-processor target hardware--we had to build our own, in assembler and Ada.