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=0.6 required=5.0 tests=BAYES_05,INVALID_DATE autolearn=no autolearn_force=no version=3.4.4 Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!uunet!igor!rutabaga!jls From: jls@rutabaga.Rational.COM (Jim Showalter) Newsgroups: comp.lang.ada Subject: Re: Ada9x Transition Plan (443 lines) ftp-ed from AJPO Message-ID: Date: 26 Mar 91 02:39:37 GMT References: <2916@sparko.gwu.edu> Sender: news@Rational.COM List-Id: > 3.2 FREE EDUCATIONAL ADA 9X COMPILATION SYSTEM > > Since this system is for educational purposes only, and > there is no intent to compete with industry but rather to stimu- > late the market, it may be that certain features of the language > are not supported; Bad idea, bad precedent. One of the great selling points of Ada over any other language is that it IS a standard. Introducing dialects, ESPECIALLY in an educational system, is regressive. Either teach Ada or don't teach Ada, but by all means don't teach "sort-of" Ada. > The original testing philosophy was to expose compiler errors, Seems like a good idea. Why depart from this philosophy now? If it aint broken, don't fix it. > whether or not an error was likely to be encountered by users or > would have been considered a significant defect by programmers. > The new, usage-based testing philosophy is to focus on potential > non-conformities that would impair the actual use of the lan- > guage. How does one guess ahead of time which particular compiler bugs will be "significant" in the eyes of a programmer? Personally, I view ALL bugs as an evil thing that must be stamped out. I would not like having to choose which bugs I wanted to live with and which bugs I wanted fixed. I want them ALL fixed. That's the whole POINT of validation. > Furthermore, continued growth of the ACVC test suite will be > prohibited. WHY? Why set an artificial ceiling on the number of tests? The number of tests should be exactly equal to the number required to ensure validation, no more, no less. This seems elementary--who is pushing to limit the number of tests? > Tests may be modified/added/deleted but a fixed > ceiling of approximately 3800 tests (187,000 LOC) will be estab- > lished. While this might be a great idea for laws, it stinks for testing compilers. If you are limited to 3800 tests and there are 3801 things that can be wrong with a compiler, you are guaranteeing that one bug will be visited upon all Ada users. > The increased focus on usage and the prohibition of continued > growth is expected to provide vendors with more time to concen- > trate on meeting users special requirements for optimization and > high quality tools. I would rephrase this to read "The prohibition of continued growth [of the tests] is expected to provide vendors with a shortcut to achieving validation at the expense of the Ada user community". -- ***** DISCLAIMER: The opinions expressed herein are my own. Duh. Like you'd ever be able to find a company (or, for that matter, very many people) with opinions like mine. -- "When I want your opinion, I'll read it in your entrails."