Path: utzoo!attcan!uunet!samsung!think!linus!eachus From: eachus@aries.mitre.org (Robert I. Eachus) Newsgroups: comp.sys.amiga Subject: Re: Any ADA compilers available? Message-ID: Date: 7 Feb 90 22:05:18 GMT References: <1367@crash.cts.com> <816@cs.nps.navy.mil> <9002050356.AA22489@fsucs.cs.fsu.edu> Sender: news@linus.UUCP Organization: The Mitre Corporation, Bedford, MA Lines: 55 In-reply-to: peterson@fsucs.cs.fsu.edu's message of 5 Feb 90 03:55:42 GMT In article <9002050356.AA22489@fsucs.cs.fsu.edu> peterson@fsucs.cs.fsu.edu (Eric J Peterson) writes: I read a few months ago in an issue of ACM's SIGPLAN that the AJPO was going to allow the trademark on the name "Ada" to expire without renewing it, although there was no mention of why this was happening. Is this true? If so, why? Yes, it is true. Why? For complex reasons, a trademark is not the right way to protect the name of a programming language. The original trademark was applied for anyway, so that the "threat" of a government backed trademark violation suite could prevent flagrant rip-offs. Now that many real Ada compilers exist, the AJPO feels that any company which markets a subset compiler will be harming their own reputation, but not that of Ada, so there was no need to renew. There is a trademarked Ada validation mark however, which can only be used on validated compilers. Here at FSU, we have the Janus Ada compiler for MS-DOS machines which takes up about 3 megs of disk space and a full 640K of RAM to run -- it doesn't even leave enough memory to run our networking software. And you need a 386 of at least 20 MHz to get any reasonable speed out of the compiler. What kind of performance can we expect of an Amiga Ada compiler in comparison? It should scream by comparison. The exceptions and tasking in the Amiga Kernel were intentionally designed as an implementation of Ada tasking and exceptions. (This is not the right way to say it, maybe: The Amiga gods saw Ada tasking and Ada exceptions and they were good, in fact just what they needed...) In any case, a fast Ada run time for the Amiga is childs play, since Ada tasks are AmigaDOS tasks, etc. Dan Eilers of Irvine Compilers said that it took four days to retarget their compiler to the Amiga. (One of the people who works there told me that the compiler was useless without a hard disk, since it required about 4 Meg for the compiler and libraries...I had to explain that in that case I might put it in a RAM disk for speed. :-) However, market demand for a compiler has to show up before it will appear on the market. Once you start selling an Ada compiler, figure a maintenance staff of one or two people per host forever. In addition, the first validation on a particular host ends up costing about a man year above and beyond any development work. Robert I. Eachus with STANDARD_DISCLAIMER; use STANDARD_DISCLAIMER; function MESSAGE (TEXT: in CLEVER_IDEAS) return BETTER_IDEAS is... -- Robert I. Eachus with STANDARD_DISCLAIMER; use STANDARD_DISCLAIMER; function MESSAGE (TEXT: in CLEVER_IDEAS) return BETTER_IDEAS is...