david@ztivax.UUCP (03/19/86)
Someone sent me mail saying I should not just give flippant remarks about Ada, but I should be more specific. After all, this is a forum for discussion. So, here goes: I do not think Ada is significant nor a Good Thing for computer science in general for the following reasons: 1) A "common programming language" is not significant. It is not even DESIRABLE. Do you think Ada is good for Avionics? Even with a stack oriented calling convention??? Even with dynamically scoped exception handling??? (remember, these things have to be RELIABLE) Better than SQL for database access? Better than report generation languages for defining report layout? Than screen def languages for interactive screens? A spreadsheet for tabular computation? NO WAY. Besides, how long does it take YOU to become proficient in a language? One week? Two weeks max? (I am assuming the reader is both intelligent and has studied languages, at school or in the 'real world'). How then is this such a big deal? 2) APSE. Why limit ourselves to some specified environment when the free market (which is MUCH MUCH MUCH larger than the DoD market) is doing it, and better? Read Smalltalk, UNIX, Macintosh ... 3) Do strongly typed languages REALLY help? I will admit they do something, but I am not convinced that SYSTEMS developed using such techniques are any less bug ridden than loosely typed systems. There are MANY MANY examples of this (look at compilers written in Pascal versus C - Which is more reliable????) 4) Software engineering is in its infancy, and will continue to be until something is REALLY discovered which reduces effort and bugs. As Parnas goes to great lengths demonstrating, the current SOA of software engineering as embodied by Ada DOES NOT MAKE ANY SIGNIFICANT IMPACT in either the effort required for implementaation nor the number of bugs in the final system. Ada and the engineering support offered by APSE only helps if you are re-inventing the wheel. 5) The government could do many things to reduce the cost of DoD programs to a very small fraction of the current budget. I feel a very bad way is to force a standard (Ada) which is stone-age (because all currently used software technology is stone-age). If they just make the contractors use the few people who are productive, instead of the thousands of mostly brain dead bodies as is typically done, then a very large savings would be realized. Force people to use a poor technology for standards sake, ... Aguments? seismo!unido!ztivax!david