ted@grebyn.com (Ted Holden) (12/07/89)
From Terri Richard: Ohio State University Computer >I think most people who read this conference are really tired of hearing >this. If you love C++, great. Just don't hassle productive conversations >on THIS conference. I for one am extremely tired of hearing it. Is somebody actually forcing you to read these articles?? Just hit the "n" key and don't worry... be happy! (Apparently, the truth doesn't set everybody free; some, it just makes squeamish...) From Jeffrey Stewart: Software Engineering Institute, Pittsburgh, PA >>The literature and every conversation I have ever had with real-world >>people who have been forced to actually attempt to USE Ada have totally >>convinced me of the failure of this project. By the same token, C++ >>could very nearly, if not entirely, fill the bill. >Well Ted, I guess you never talked to me or the group I worked with at >Hercules Defense Electronics. >We have just finished a project to implement all of the control software >plus some signal-processing software, in Ada, for a Millimeter-Wave radar >seeker for the Maverick missile. >Its about 7500 lines of code, based on an object-oriented design, does >signal-processing, provides dual-loop control for the antenna, provides >mission control for the entire missile, plus some other things. And it >does it at 600Hz, on a 10MHz 80286. There is NO ASSEMBLER. The goal >was a complete Ada artifact, and the goal was achieved. You're not really telling us very much here... most importantly, what all did you have to turn off? What if anything differentiated your use of Ada from most people's use of Pascal (other than cost)? >Projects that DO have a problem with Ada tend to fall into the following >categories : > The classic "10 pounds of stuff into a 5 pound bag" problem. You won't > solve this with C either. Many times, the requirements stack the deck > against you. Period. > Less than production-quality compilers. I'll admit this was a problem, > and is less so every day. Perhaps you are referring to a recent comment by one of your fellow Ada supporters on the net: "My only problem with Ada at this point is the cost ($ and hardware resources) of a compiler for my XT clone. Both IntegrAda and Janus require more memory than DOS 4.01 leaves available. This is BAD DESIGN. There is no excuse for a 551K executable in a PC (pass 2 of Integrada). Janus Ada requires > 580K available to run, and rumor has it that the Integrada compiler is a repackaged Janus compiler." The funny thing is, I don't really have that much of a problem with your basic thesis other than the fact that it's import is rather limited. Basically you seem to be saying that, of the ten or fifteen Ada compilers in the DOS world which we hear about, you were able to pick one which, given enough restrictions, could produce code which would execute rapidly enough for a fairly sophisticated project, without saying anything about compile times and the resultant impact on development times. I have no real problem with that. You seem, however, to have little appreciation of the situation in which the average poor slob who is forced to use Ada finds himself. The Ada compiler he gets to use is just the one provided on the contract by, you guessed it, the lowest bidder. Now, this would seem to indicate that there is something to be said for keeping your basic programming language spec simple enough that more than one out of ten or fifteen compiler vendors might succeed at building a reasonable compiler given 10 years (as Ada has had). For instance, I would lose little or no sleep at the prospect of being forced to use the lowest bidders C compiler between MicroSoft, Borland, Zortech, Watcom... Could you say the same as regards Ada compilers? In the same vein, when you attempt to enforce something like Ada upon the numbers of people involved, you should have some idea of how these people live; for instance the fact that a great many of them are just now getting Unisys/Arrete 5000/95 and/or similar computers to work with. Or, for instance, the fact that almost none of them are having any luck keeping Ada programmers or "software engineers" around as regular government workers. Consider that you will soon have to justify Ada on a cost basis, and that almost all Ada programmers are contractors... minor detail worth mentioning. > Whiners who will use every excuse to blame the language rather than the > design. The two ARE different, design and implementation, don'cha know? > Many of these projects fail due to poor design, regardless of > implementation language. When you hear the whole world whining, there's probably some real reason for it. >I refuse to assign percentages to these categories. Let that be an exercise >for the reader. >By the way Ted, part of the rationale for Ada was that the techniques and >languages in current use (mid 70's) would be inadequate for the million-plus >SLOC systems anticipated for the 90's. What are the successful C-language >projects of this size? UNIX, WordPerfect, DBASE-IV, X-11, Quattro, Sprint, News... What if any million-line projects have actually been successfully implemented in Ada (without having major language problems in 10% of the code)? The truth is, I wouldn't want to be using C on huge projects anymore; that's what C++ and the object-oriented paradigm are about. C++ has it now. Ada might have it in the future, but: 1. How many years from now? 2. At what cost in clunk factor in a language whose clunk factor is already too high for most users. In the Zortech C++ manual, p 62, the authors note: "An overwhelming design requirement was that the speed, efficiency, and simplicity of expression of the C language was to be retained in the new language." The corresponding statement, ten years from now, will probably read: "The unholy clunk factor of Ada just got a lot unholier..." The best description of the applicability of object-oriented programming to large projects which I have seen is the section of the Pinson/Weiner book (Intro to OOP & C++) which shows a C++ linked-list application with and without late-binding, along with the requirements for maintaining the two sections of code. Night and day. >To sum up, Ada DOES work, for its intended purposes. Maybe you should >consider becoming an equal opportunity basher, i.e., cite some C/C++ >failures as well. (And don't tell me there aren't any. Your use of >absolute terms is part of what detracts from your credibility.) It isn't really like there aren't any other losers in the realm of computer languages: there are Fortran, COBOL, PL/1, Forth, Prolog, and any number of others; I would actually choose Ada over one or two of these for certain kinds of applications, since Ada is at least a structured language. The thing is, these others are just quietly withering away on the vine and/or dying of old age of their own accord and in accordance with ordinary free-market principals; they do not seem to require a stake through the heart. The thing is, when you go to ENFORCE a language on the entire U.S. military, you owe it to them to ensure that that language is the best available, and not one which, according to everything in print, most compiler vendors have never learned to deal with after ten years. Actually, it would be better to simply allow our much-touted free-enterprise/free-market system to do it's thing; that's what's supposed to seperate us from the communists, isn't it? Ted Holden HTE
rsd@sei.cmu.edu (Richard S D'Ippolito) (12/08/89)
I appreciate Dave Emery's request for a moratorium on flames and a return to business. I will do both, but still respond to Mr. Holden's recent address: In article <14075@grebyn.com> ted@grebyn.com (Ted Holden) writes: >From Jeffrey Stewart: Software Engineering Institute, Pittsburgh, PA > >>We have just finished a project to implement all of the control software >>plus some signal-processing software, in Ada, for a Millimeter-Wave radar >>seeker for the Maverick missile. > >>Its about 7500 lines of code, based on an object-oriented design, does >>signal-processing, provides dual-loop control for the antenna, provides >>mission control for the entire missile, plus some other things. And it >>does it at 600Hz, on a 10MHz 80286. There is NO ASSEMBLER. The goal >>was a complete Ada artifact, and the goal was achieved. > >You're not really telling us very much here... most importantly, what >all did you have to turn off? What if anything differentiated your use >of Ada from most people's use of Pascal (other than cost)? What you mean is, he's not really telling you what you want to hear. If you are really interested in this project, you can talk to someone who attended the recent TRI-Ada conference in Pittsburgh, where Mr. Steve Pate, the lead engineer from Hercules, spoke at length about the whole design and implementation, including compiler selection. Better yet, you may wish to loosen your purse strings and attend the AdaJUG meeting in San Diego in February, where the same gentleman will speak. >You seem, however, to have little appreciation of the situation in which >the average poor slob who is forced to use Ada finds himself. The Ada >compiler he gets to use is just the one provided on the contract by, you >guessed it, the lowest bidder. This is pure nonsense! The choice of compilers was up to Hercules. As far as "poor slobs" go, Jeff and I were members of the team in "the situation", so we certainly couldn't lack appreciation of our own endeavors, could we? Furthermore, this is not the first such project -- we have spent over three years producing just this kind of improvement of the Ada programmer's lot AND the resulting products. >In the same vein, when you attempt to enforce something like Ada upon the >numbers of people involved, you should have some idea of how these >people live; Mr. Holden, see above. >Consider that you will soon have to justify Ada on >a cost basis, and that almost all Ada programmers are contractors... >minor detail worth mentioning. And worth refuting. I can give you the names of program officers who will insist that our methods with Ada reduce life-cycle costs and, more importantly, make possible products that aren't economically possible with any other language. Finally, in this business, contractors are companies; programmers are people. >UNIX, WordPerfect, DBASE-IV, X-11, Quattro, Sprint, News... What if >any million-line projects have actually been successfully implemented in >Ada (without having major language problems in 10% of the code)? I'm sorry, but your insistent mis-interpretation of this reference forces me to expose your unethical methods: I will reproduce several quote form the December 88 "Journal of Electronic Defense" to show that you are attempting to pull quotes out of context to support your position: Moreover, the recent ABICS-3 tests established tha Ada does not impose any significant restraints on system speed specifically when applied to flight-critical avionics. [In the F-15] Ada-based systems were transparent to the aircrew, and match[ed] the performance of standard units programmed in assembly or FORTRAN. ...[using Ada,] an error rate reduction [over other languages] of 25% was experienced after...three projects' worth of experience. More critical errors seem to be found sooner in Ada development than in programs written in other languages. ...as Ada projects get bigger, they get cheaper. So the bottom line is that the much-maligned edict for the use of Ada is now beginning to look like a superb move. The above quotes are from the guest editorial by Hugo Poza on pages 32 and 33. I am assuming, from your use of this issue, that it is unlikely that you would have skipped over these two pages in your search for negative material. If you did indeed miss these and other similar items, then I apologize here and now for condemning your actions, but respectfully suggest that you improve your research methods. Rest assured, however, that I will always respect your opinion of Ada by not inviting you to be on any of my Ada project teams. Enough said. Richard S. D'Ippolito, Project Leader Software Architectures and Design Software Engineering Institute -- When you can measure what you are speaking about, and express it in numbers, you know something about it. Lord Kelvin rsd@sei.cmu.edu -----------------------------------------------------------------------------
gerhardt@ajpo.sei.cmu.edu (Mark Gerhardt) (12/10/89)
I suggest that the people who desire to flame about their quasi technical language issues form a new newsgroup: comp.lang.bigot Or even better, go somewhere where your level of technical contribution will be appreciated for what it is not - wither compuserve or maybe even public office. Both sides of this discussion have gotten unbearably boring. I agree with emery. Mark Gerhardt Esl, Inc. Sunnyvale, CA
ted@grebyn.com (Ted Holden) (12/11/89)
From: Richard S D'Ippolito, Software Engineering Institute >>You seem, however, to have little appreciation of the situation in which >>the average poor slob who is forced to use Ada finds himself. The Ada >>compiler he gets to use is just the one provided on the contract by, you >>guessed it, the lowest bidder. >This is pure nonsense! The choice of compilers was up to Hercules. As far >as "poor slobs" go, Jeff and I were members of the team in "the situation", >so we certainly couldn't lack appreciation of our own endeavors, could we? >Furthermore, this is not the first such project -- we have spent over three >years producing just this kind of improvement of the Ada programmer's lot >AND the resulting products. I can believe that individual teams using PCs (for which Ada compiler costs appear to run no higher than 1.5 times that of the machine) might occasionally get to choose their own Ada compiler. Most others haven't been as fortunate. Remember this point, however, when we get to the point concerning left-out JED quotes below. >>Consider that you will soon have to justify Ada on >>a cost basis, and that almost all Ada programmers are contractors... >>minor detail worth mentioning. >And worth refuting. I can give you the names of program officers who will >insist that our methods with Ada reduce life-cycle costs and, more >importantly, make possible products that aren't economically possible with >any other language. Finally, in this business, contractors are companies; >programmers are people. I've seen people insist they were Napoleon Bonaparte. Unless you are claiming to be either: 1. A government worker or 2. Working for GS-9 - GS-12 wages, than the above is hardly what I would call a refutation. I think almost everybody reading this knows the difference in cost between government workers and contractors. >>UNIX, WordPerfect, DBASE-IV, X-11, Quattro, Sprint, News... What if >>any million-line projects have actually been successfully implemented in >>Ada (without having major language problems in 10% of the code)? >I'm sorry, but your insistent mis-interpretation of this reference forces me >to expose your unethical methods: You're out to set the entire U.S. military back 20 years in basic computer science and you think I'm being unethical??? The exact quote was and remains: "In October, Magnavox reported that AFATDS contains 1,245,802 lines of code. Ninety percent of the software requirements were met with no major software problems." What's to misinterpret? Since the entire article is about Ada, I assume that the ten percent of the requirement which met with "major software problems" involved Ada. The (relatively simple) point which I was making was that most people probably don't read these articles past the headlines but that, if you did, you would more often than not find evidence of the entire Ada program being in serious trouble; things such as a major league project having major language-related problems with 124,580 (roughly) lines of code out of 1,245,802. And this, of all things, in articles and magazines whose vested interest, one would have to assume, is in supporting Ada. I'm afraid all you've exposed is your own inability to comprehend a rather simple article and a rather simple rhetorical point. >[several quotes from JED] >The above quotes are from the guest editorial by Hugo Poza on pages 32 and >33. I am assuming, from your use of this issue, that it is unlikely that >you would have skipped over these two pages in your search for negative >material. If you did indeed miss these and other similar items, then I >apologize here and now for condemning your actions, but respectfully suggest >that you improve your research methods. You think maybe I left out too many valuable quotes from the AFATDS article????? No problem at all: 1. "The language was very resource intensive in the host environment. The compilers the project used were huge; very few compilations saturated the mainframe, and some Ada libraries were extensive." 2. "The program libraries required a lot of recompiling. 3. "The tasking implementations were so inefficient that portions of the data management system had to be re-designed to decrease the number of tasks." 4. "The programs were made larger by overhead and runtime checks." 5. "Ada's portability features made it difficult to store data efficiently." 6. "Ada variant record structures often consumed more memory than expected." Almost sounds like these guys might honestly be regarded as some of the "poor slobs" I mentioned in the previous article". Anybody still think that point was "Nonsense"? In fact, it almost sounds like the person writing all of this might have had major software problems with at least ten percent of his requirements and that this might have been caused by Ada. You could search journal articles about C or C++ until the cows came home and you would never find those kinds of comments. C would simply not be around if this were not the case. Such is a free system, as opposed to yours. >Rest assured, however, that I will always respect your opinion of Ada by not >inviting you to be on any of my Ada project teams. Enough said. Nonetheless, I plan on making quite a bit of money with Ada; by cleaning up the mess and teaching courses in C language and C++ after the fall. I mean, to me at least, it's always seemed funny how some people seem to have all the answers right up to the day when it's all over. Take Eric Honnecker, for example... From: Horst Kern, PCS GmbH, Pfaelzer-Wald-Str. 36, 8000 Muenchen >The recent discussion initiated by Ted Holdon's article surprised me >a bit. Nobody answered the argument that I thought would be THE one >in favor of Ada: > It's not the programmers that decide which language to use in the future >In the European Market we expect to have laws in 1992 that resemble the >American Laws of Product Responsibility. As far as I know, those can make >an implementor responsible for any consequences of using an implementation >language which is not considered the best choice. And there are strong >signals that Ada will be the default 'best choice' for lawyers. Is this >not true for the US? Das gefaehlt mir sehr; Advocaten VERDIENEN Ada (und alle andere Wehe und Ungluecke). Most other victims of Ada are, however, innocent and undeserving. Ted Holden HTE
rsd@sei.cmu.edu (Richard S D'Ippolito) (12/12/89)
In article <14090@grebyn.com> ted@grebyn.com (Ted Holden) writes: >Nonetheless, I plan on making quite a bit of money with Ada; by cleaning >up the mess and teaching courses in C language and C++ after the fall. >Ted Holden OK, Ted, you've convinced me. When you get get to that point, please consider putting the SEI at the top of your student list. I can't speak for my colleagues at Mitre, but many of us here apparently really need your help. Sorry to have been so dense. At least it will give us a few months to convince the computers that the Ada programs don't really work... Rich -- When you can measure what you are speaking about, and express it in numbers, you know something about it. Lord Kelvin rsd@sei.cmu.edu -----------------------------------------------------------------------------