Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!usc!cs.utexas.edu!sun-barr!ames!vsi1!ubvax!igor!rutabaga!jls From: jls@rutabaga.Rational.COM (Jim Showalter) Newsgroups: comp.object Subject: Re: ada-c++ productivity Keywords: Looking for a few lazy men Message-ID: Date: 21 Mar 91 21:10:36 GMT References: <11966@pasteur.Berkeley.EDU> <1991Mar15.224626.27077@aero.org> <1991Mar16.000624.2513@leland.Stanford.EDU> <1991Mar16.205228.4268@grebyn.com> <1991Mar21.024445.8746@grebyn.com> Sender: news@Rational.COM Lines: 134 >Agreed. The language issue (Ada) is simply another symptom of the same >disease (terminal stupidity) as the other symptom (2167/A). This is a contentless argument, since you provide nothing to back up the claim that Ada is the result of stupidity. Your uniformed biases don't constitute evidence. >Object oriented languages are a partial solution to the problem; the >best answer American computer science has yet devised. Actually, this is complete bullshit. Stroustrup himself says that C++ addresses 5% of the actual problem, the other 95% consisting of problems in design, architecture, and process. Computer science (actually, more properly, software engineering) has developed some methods for addressing this other 95% of the problem. >Weiner's book on C++ and OOP describe means of using these techniques to >eliminate complexity and vastly simplify maintenance of programs. Too >bad Ada doesn't have these features. Ada simply adds (greatly) to the >complexity which programmers must deal with, and offers no paybacks or >quid-pro-quo for the added complexity. Again, contentless argument. In what way does Ada add to the complexity? Ada introduces strong typing, separation of specification and implementation, separate compilation, algorithmic and metatype parameterization (via generics), opaque types, and exceptions. These features were the result of years of first-rate research into ways to REDUCE the complexity of programs. Oddly enough, these SAME features constitute 80% of the difference between C++ and C. >One such is the popular VMUSIC multipart musical routine for PCs >(formerly thought to be impossible), available on BBSs. You didn't answer my question (why am I not surprised?). What I asked was: "How complex is the stuff you work on?". Your response here tells me nothing. I want to know: a) How many SLOC is it? b) How many people worked on it? c) How many development platforms did the team work on? d) How many configurations does it support? e) How many targets does it run on? f) How many separate processes does it contain? g) What is the method of inter-process communication? h) What documentation standard was it developed under? i) How many subcontractors were involved? j) How many years did it take? On first blush, something called a "routine" seems more like a two-guys-in-a-garage-typing-away project than something even remotely as complex as the sorts of projects I'm impressed by. >If programmed >in Ada, it would sound like two or three dogs growling at eachother, Again, this is not supported by evidence. You merely reveal your prejudice against Ada with no justification. This is generally called "ignorance". >that is, if it could be programmed in Ada. I doubt it. Are you aware of the concept of Turing equivalence? Anything can be programmed in anything. If you find a counterexample, you get a Nobel. >Bullshit. Unix is written in C, Yeah, and that's one of the reasons IBM had to rewrite it from scratch for the RS6000. I used to work in a UNIX shop (we made 32-bit superminis [didn't everybody, once upon a time?]). Looking at the guts of the kernel was a VERY scary experience. One guy, describing UNIX, paraphrased the description of a helicopter: "A bunch of spare parts flying in loose formation". Every time I execute a UNIX command and my system doesn't crash, I'm amazed. >WordPerfect, Ami, and most modern >software. "Most" seems rather excessive, since you then go on to say: >The bulk of American scientific software is in Fortran. The >bulk of American business software is in Cobol, Which is it? After all, COBOL represents 65% of ALL software in the world. (Incidentally, COBOL and FORTRAN are acronyms, so one typically writes them in all uppercase.) >I damn Ada from the various horror stories I read and hear regarding it. Uh huh. So the short answer is that you don't know how to write in Ada. I figured as much. You say you've read various horror stories. Interesting: have you ever bothered to read success stories--there are many of those, you know. Or would reading that Ada had saved some sites millions and millions of dollars be too inconvenient for your worldview? Most people, when they don't know enough about something to form an opinion based on fact, have enough sense to keep their mouth shut. You, on the other hand, don't let your nearly complete ignorance of Ada keep you from shooting off your mouth about it. >I have managed to avoid it in my personal life, other than having to >write interfaces between it and low-level file-handling routines written >in C. Doing that, I personally watched an Ada compiler take 25 minutes >to compile a 30 line program into a 600K byte executable; I never saw a >C compiler do that. So you had a crappy compiler. Does that make the LANGUAGE bad? Try to keep in mind that Ada pushes the state of the art of compiler technology and that early compilers ate dust bunnies. Compiler technology has continued to improve, and today you can buy Ada compilers that produce code every bit as tight as the best C compilers. Furthermore, Rational sells compilers that perform incremental compilation, and for many classes of change--even to specs (even to the implementation of the private types in specs)--we can compile in the change in a matter of seconds. Indeed, our effective batch compilation rate often exceeds 100KSLOC/minute. I've never seen a C compiler do THAT. >>Productivity rates range from no gain to order of magnitude gains. We >>have lots of success stories backed up by factual accounting data if >>Mr. Holden would care to read them. >I found the tales on the Adawoe BBS to be more of an indication of the >real effects of Ada than the kind of bullshit you're describing. So anecdotal evidence is better, in your eyes, than the balance sheets of major corporations? So far you've proven to my satisfaction that you'd make a lousy Ada programmer or software engineer, but this last comment of yours demonstrates, in addition, that you'd make a lousy scientist or accountant. What DO you do for a living, anyway? -- ***** 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."