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_00,FREEMAIL_FROM autolearn=ham autolearn_force=no version=3.4.4 X-Google-Language: ENGLISH,ASCII-7-bit X-Google-Thread: 103376,bc1361a952ec75ca X-Google-Attributes: gid103376,public X-Google-ArrivalTime: 2001-07-30 15:35:48 PST Path: archiver1.google.com!newsfeed.google.com!postnews1.google.com!not-for-mail From: hfrumblefoot@yahoo.com (Hambut) Newsgroups: comp.lang.ada Subject: Attributes of a development which require Ada (Was Re: How to make Ada a dominant language) Date: 30 Jul 2001 15:35:47 -0700 Organization: http://groups.google.com/ Message-ID: References: <3B6555ED.9B0B0420@sneakemail.com> <9k3l9r$10i2$1@pa.aaanet.ru> <3B656345.64AB603A@sneakemail.com> <9k3oa1$2qg8$1@pa.aaanet.ru> <3B657715.7EC592D9@sneakemail.com> NNTP-Posting-Host: 62.188.137.98 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit X-Trace: posting.google.com 996532548 18515 127.0.0.1 (30 Jul 2001 22:35:48 GMT) X-Complaints-To: groups-abuse@google.com NNTP-Posting-Date: 30 Jul 2001 22:35:48 GMT Xref: archiver1.google.com comp.lang.ada:10793 Date: 2001-07-30T22:35:48+00:00 List-Id: Russ Paielli <18k11tm001@sneakemail.com> wrote in message news:<3B657715.7EC592D9@sneakemail.com>... > I am trying to sell Ada for a safety-critical program, and I am getting > little or no support from my organization. This is sad to hear. For the safety-critical problem domain I would have expected this would be a relatively easy sell. Anyway this got me to thinking 'What attributes of a development most forcefully require the use of Ada?' (if this sentence makes any sense - typically it's late at night) The bottom line of my thinking is: 1. Will using Ada make the developers life sufficiently easy that the cost of training and tooling up is worth it? 2. Are the verification requirements of the application such that Ada is the only language with the appropriate tool support to be able to meet the requirements? The longer version of these points are: 1. Simplification of Developers Life ------------------------------------ I can point to two (kind of) non-technical things that have pushed the use of Ada on recalcitrant managers, in my experience: o Requirements of standards - The developer is forced to use Ada because of the standards that are contractually applied. o Perceived requirements of regulators - Ada is chosen because the developer believes they will get a bigger 'credit' at certification from the regulator. So a developer will adopt Ada if the Developer believes it will make their (contractual) life easier. Is this likely to be true for your development? Is there an independent safety organisation that would very much prefer to see the application developed in Ada? 2. Verification Requirements ---------------------------- >From a technical point of view I think that the most convincing drivers for the use of Ada are the verification requirements. In other words how much testing and software analysis do you have to do? The more you have to do the more attractive Ada becomes. Particularly if it's necessary to *demonstrate* very high levels of assurance in certain properties of the final application (for example absence of run time errors). Demonstration of high assurance would probably entail, in addition to high levels of testing, both a lot of very deep thought about the software and high levels of software static analysis, for example deep flow static analysis. Currently I believe that Ada is the only language to cost effectively support high levels of software static analysis. I'm probably not saying this too well. What I'm trying to get at is that Ada has probably the most cost effective toolset for doing things like: o Showing that all variables are initialised before use o Proving absence of run time errors o Program proving against a formal specification So if you're application absolutely must never experience a run time error (e.g. Nuclear Power Plant Protection Systems?) then there may be a very good argument for the use of Ada based on the rigour of verification that's required. (Additionally ASIS looks like it will provide an excellent tool to do other analyses.) >I get forwarded email > messages from full professors of CS at MIT claiming that Ada is being > replaced by Java even in their studies of software reliability. I would be very interested if you have publicly accessible examples of such studies. Cheers, Hambut