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=3.7 required=5.0 tests=BAYES_20,MISSING_DATE, MISSING_MID,MISSING_SUBJECT autolearn=no autolearn_force=no version=3.4.4 From: utzoo!decvax!harpo!floyd!vax135!cornell!ddw Newsgroups: net.lang.ada Title: The Devil, Ada, and Tom Slone Article-I.D.: cornell.4394 Posted: Sun May 8 21:13:17 1983 Received: Tue May 17 23:02:15 1983 List-Id: Message-ID: <20200615230049.GviyL4j40D8PWq27v3W1yOeOSVwNAJeDt5WfRjkU8ek@z> Tom Slone (arizona!tom) has done a wonderful job of further muddying the already turbid waters swirling around Ada. (Ada is a TM of the AJPO and all that rot.) His article is a fine mixture of half-truths, hearsay, speculation, and a dash of paranoia on the side. The original article was quite political but I suppose the references to Ada made it legit for this newsgroup, so I'm replying here, and trying to avoid the politics (mostly). Taking Slone's points one by one: Ada is a monsterously [sic] large language which is difficult to implement. The NYU version of Ada, implemented in SETL, is unbelievably slow. It is indeed. However, it's odds-on that it's slow because it's written in SETL, which is monstrously slow. Jack Schwartz gave a talk here at Cornell a year or so ago on SETL; one of its funnier aspects was that NYU had a SETL optimizer written in SETL, but it was so slow that they hadn't yet gotten around to running it over itself because it would take too much machine time. Ada may well be difficult to implement, but it may just be that we haven't spent enough time thinking about it. I do worry about the size of the language; Dijkstra [name-drop time here] told me once that there are two denotational-semantic definitions of Ada, both over 500 pages long, and nobody has any idea if they're equivalent! If Ada is successful, it seems likely that its use will be limited to the DoD and its contractors. Furthermore, military computer programming would exclusively use Ada. This would solve the major dilemma for anti- military programmers. One would know immediately whether one is furthering bomb building, espionage and international hanky-panky. Hard to know where to start on this one. This "it seems likely" sort of non-logic is annoying to argue against. I would argue that once many programmers have been schooled in Ada, if they grow to like it they will want to continue using it even if they cease working on military applications. Also, a shop that does a lot of military Ada work would probably want to simplify its life by doing various in-house things in Ada as well. So a "We Use Ada" sign in the window may not prove much. The influence of the military on academia has NOT been to encourage general research and academic freedom, but to create secrecy and to create focus upon research which has military application. Most scientific research has military application, other than some of the more far-out stuff in mathematics (or CS or physics or other areas that look a lot like mathematics). Some of it is just more immediate. Besides, military research generally has civilian applications. How do you propose to guarantee a distinction? The widespread use of Ada in the military-industrial complex would cut apart those academicians and scientists who are for freedom from those who are for secrecy. It would cut apart those who are for peace from those who are for war and mutual fear. Other than pointing out that people who are for war, secrecy, and mutual fear are rather hard to find [well, they're hard to find outside of the Reagan Administration, anyway], I think the above statement is sufficiently idiotic that it does not need a response. I'd love to see someone even *try* and prove it. My fear is that Ada will either be not popular enough, and hence disappear as another forgotten language. Or that it will become too popular, i.e., ubiquitous. This latter possibility seems unlikely though, as an implementation of Ada on a micro-computer seems almost laughable. Since Motorola should have the 68020 micro out in a couple of years, and since it will have a full 32-bit address space, memory management, and all that good stuff, I doubt that it's "laughable". What I really don't understand is why Slone is worried that Ada might disappear or become ubiquitous. It might disappear, which would mean a lot of blown tax dollars, but what's the danger? I can't imagine it becoming ubiquitous; COBOL didn't, and DoD was behind *it*. Certainly Ada is an improvement over COBOL in many ways and life would no doubt be improved if Ada replaced COBOL. Sorry about the length of this article, but I didn't see any way to make it shorter without leaving things out. David Wright {vax135|decvax|purdue}!cornell!ddw ddw@cornell