munck@MITRE-BEDFORD.ARPA (Bob Munck) (04/07/86)
From previous traffic: >> If they just make the contractors use the few people who are >> productive, instead of the thousands of mostly brain dead bodies >> as is typically done, then a very large savings would be realized. > ... the state of the practice at present is not ready > to support (the above) ideal. I don't believe that it is the state of the practice in SOFTWARE ENGINEERING that keeps us using armies of ants for software projects, with resultant productivity rates of 500-1000 LOC/MY and costs of $100-200 per LOC. Rather, it is the way our procurement process perceives and handles "risk." A typical software procurement might be on the order of 100K estimated LOC and two years duration. Given what I have seen and experienced, I'd be happy to give such a job to a group of, say, FIVE smart, experienced people who proposed to use individual ATs hooked into a LAN with good compilers, debuggers, word processors, etc. Cost, assuming the five are well-paid for motivation and not saddled with the overhead of a big company, would be maybe $1.5M Fixed Price. My confidence would come from becoming acquainted with the people, seeing their past accomplishments, etc. FORGET IT! That approach would be seen as EXTREMELY high risk by all but the most enlightened in the procurement effort. To them, safety lies in hiring a big defense contractor for maybe $15M CPFF and holding him to an 8.25% profit TO SAVE MONEY. The contractors, knowing that their profits will be limited this way, have all bid ant army approaches with 50-75 people because that maximizes their total profit. Past experience allows us to say that the project WILL be late and over budget. I see it as a generalization of the maxim "No one ever gets fired for buying IBM." That is, "No one ever had their career hurt by hiring a big aerospace company." In fact, the $1.5M approach might be more likely to hurt your career, because you've failed to spend all the money you were supposed to. My great hope for the Ada effort is that it may make the smart, experienced people so extremely productive and so extremely successful in building large systems that DoD procurement will be forced to recognize the validity of this approach and stop hiring the armies of ants. -- Bob Munck (speaking for himself and a small circle of friends)