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=-0.0 required=5.0 tests=BAYES_40 autolearn=ham autolearn_force=no version=3.4.4 X-Google-Language: ENGLISH,ASCII-7-bit X-Google-Thread: 103376,bbba36730ac96f9a X-Google-Attributes: gid103376,public From: Ron Thompson Subject: Re: Gov't, non-DoD use of Ada Date: 1996/09/10 Message-ID: <513so4INNcir@faatcrl.faa.gov> X-Deja-AN: 179715742 references: <4vnlgn$mko@uuneo.neosoft.com> <50p6ldINNall@faatcrl.faa.gov> content-type: text/plain; charset=us-ascii organization: aos-420 mime-version: 1.0 newsgroups: comp.lang.ada x-mailer: Mozilla 1.2 (Windows; U; 32bit) Date: 1996-09-10T00:00:00+00:00 List-Id: gwinn@res.ray.com (Joe Gwinn) wrote: >In article <50p6ldINNall@faatcrl.faa.gov>, Ron Thompson > wrote: I did write, and I am glad to see a tech from a company able to talk to a tech from the gummint. We need this sort of exchange, and here goes: >I don't know how much credence I would put in the party line from the PR >dept. That's why I don't bother to ask. And I know technical people who >were there. I couldn't agree more with your "party line" hesitation. Note however that I said factual info. It is there and can be gotten, once one gets past any party lines from the PR folks. Those people have TWO jobs, answer for all of us in a uniform voice, and provide facts. >That isn't what the last two FAA RFPs I worked on said. They said that >one could choose between C and C++; Ada was excluded by omission. We have >it by usually reliable channels that this omission was not an accident. I have no doubt that the RFP said something along the lines of "in a High Order Language such as C or C++...". Ada may not have appeared in the "such as" list. Usually reliable channels could certainly suggest that it was no accident. >I agree, but so what? I'm not the FAA. And, I would be terrified to use >such trendy stuff anyway -- all that stuff is vaporware, and certainly not >something I would bet civilian airliners on. However, C is widely and >successfully used in ATC. My answer may have appeared to advocate the use of such a scheme. I apologize for its' lack of clarity. My only point was that the language did not in any way cause the debacle, therefore, the AAS debacle was not and should never be a "lesson learned" about using Ada. >Not at all. They are buying full-custom, and can have it in any language >they want. And, the main language of the ATC world is C, not Ada, so it >would be hard to criticise their choice of C. There is probably ten times >as much C as Ada used in ATC applications. It matters not at all that >Thompson/CSF has been using and touting Ada. I can assure you that fine companies all over the USofA listen closely to the awarding of contracts, and line up the lawyers for the appeals. Any doubts about that? Watch the news this week, Friday, when there is a really big contract award announcement. Contact us in a year to see how many appeals and suits are pending based on that award. The point was not the choice of C or C++. We wouldn't criticise anyone for submitting a system design using C or C++. I will bet a dollar that if my company got the job on an Ada bid, LOTS of companies would appeal/sue based on the fact ONLY that they DIDN'T get it. And I bet another dollar that part of the justification that they would use would be the AAS debacle and how it happened because of Ada, and all of the techs know that is not true. C and C++ are definitely used 10 times as much as Ada. If the contract is awarded based on cost/schedule, and not on what it is that makes a good system, Ada is doomed. Remember the huge curve that Ada allegedly has. Remember that C/C++ programmers are rapidly becoming the MBAs of the 90s. Cost and schedule rules. >See comment on item 3, above. It seems to me that the FAA is talking only >about the final implementation language. I'm not sure what a "system >concept written in Ada" would be, unless you mean Ada PDL. I don't know >that anybody has tried that wrinkle out on the FAA lately. We (Raytheon) >actually implemented a Canadian ATC system in C with Ada PDL in the mid >1980s, as Ada83 compilers weren't then ready for prime time. It turned >out to be a bad idea -- the language models were too far apart. We got it >to work anyway, but wouldn't do it that way again. Again my apology for the muddiness of the answer. If my company submits a bid on a system, and it is to use Ada for the software, and the FAA awards it to another company because my company chose Ada, my lawyers would be working overtime for the forseable future. Those are the facts of the CONTRACTS end of this business. A large and well known company has had a contract for approximately two years. It is behind schedule in a large way, and they are back at the trough with the right hand out looking for money to bring the thing up to speed. The FAA says no, we are cancelling the contract with you. It is later awarded to another company that was one of the bidders on the initial RFP. The company that ran themselves into the ground on it goes to court and sues the government because we won't play the way we used to. THOSE are the facts about the contracting end of this business. THAT is happening as we speak. > > >Joe Gwinn Again, thanks for the healthy debate. My only intention was to point out that the "AAS debacle" was NOT a language problem, and that limiting ourselves to any single language would be just as bad as any other large company limiting themselves to a single language or variations of it. The references to law suits and appeals is a reference to the things that go on after the awarding of a contract, and the public doesn't usually pay much attention to them. Your swing... rct The opinions above are mine and mine alone.