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=-1.3 required=5.0 tests=BAYES_00,INVALID_MSGID autolearn=no autolearn_force=no version=3.4.4 X-Google-Language: ENGLISH,ASCII-7-bit X-Google-Thread: 103376,42427d0d1bf647b1 X-Google-Attributes: gid103376,public From: ka@socrates.hr.att.com (Kenneth Almquist) Subject: Re: Tiring Arguments Around (not about) Two Questions Date: 1996/04/17 Message-ID: <4l1lgl$en3@nntpa.cb.att.com>#1/1 X-Deja-AN: 147896822 references: <00001a73+00002c20@msn.com> <828038680.5631@assen.demon.co.uk> <828127251.85@assen.demon.co.uk> <315FD5C9.342F@lfwc.lockheed.com> <3160EFBF.BF9@lfwc.lockheed.com> <31729C6E.4903@lfwc.lockheed.com> organization: AT&T Bell Laboratories, Columbus, Ohio newsgroups: comp.lang.ada Date: 1996-04-17T00:00:00+00:00 List-Id: >> Choosing among available compilers is not an easy task. > > Actually, it's very easy. For many host/target pairs, there's only a > few vendors. In some cases, only one. In that case, it can get VERY hard. The part where you evaluate the various compilers isn't so bad. Things get more difficult after the compilers fail your tests and you have to tell your boss to switch to different hardware. :-) > "Is there some way to modify the scope of the ACVC process to > improve compiler quality across all vendors? Or, is there something > outside the scope of the ACVC that could be done to improve compiler > quality across all vendors?" The customers for Ada compilers are a diverse lot. For example, in the PC world, people expect low quality and low price. A CS 101 teacher wants a compiler that does a good job with toy programs, again at a low price. This diversity limits what can be accomplished across all vendors. The basic rules for encouraging industry to produce software which meets your needs are: 1) Make sure your purchasing decisions reflect the effectiveness in meeting your needs. If the quality of the vendor's product is poor but you buy it anyway, the vendor has no incentive to improve the product. 2) Centralize purchasing decisions to increase clout over vendors. If vendors are not willing to go to a great deal of effort to get your project to buy from them, perhaps Lockheed needs a centralized compiler evaluation and purchasing organization. Or better yet, perhaps Lockheed and other companies involved in the production of safety critical software should form an industry consortium to evaluate compilers. Assuming that the members of the consortium agreed not to purchase compilers which didn't meet the standards agreed to by the consortium, compiler vendors would have a strong incentive to meet those standards. Point 1 is critical--if you can't manage this all bets are off. Point 2 is less important. The key point with centralizing purchasing decisions is that all the people involved must have common needs. Saying, "we only purchase validated compilers" is a centralization strategy, but validation serves as a central point for the entire Ada user population and thus cannot be expected to reflect your needs very closely. In fact, if you buy any compiler which is validated, regardless of how well it meets your need, then you are violating point 1. None of this answers your question about how to "improve quality across all vendors." But remember, low quality vendors don't hurt you as long as you don't buy their products. Kenneth Almquist