Date: 27 Dec 96 13:32:19 From: "Julian H. Lloyd" <email@example.com> Organization: Diamond Partners Inc. (Singapore Branch) References: 1 2
View raw article or MIME structure
>SNIP< >The really irresponsible > thing is when people who should know better sound off like that in > public. It may be worth adding that the number of variables involved in testing even a simple piece of software make it impossible to demonstrate mathematically that it is correct - unlike hardware circuits, where all possible states can be known and tested. The 777 is by no means unique in making extensive use of software in mission-critical applications - there are few, if any, commercial aircraft flying in the US today which don't - and then, think about the air-traffic control system (antiquated though it be). The question with software is not so much on of "will it fail?" but "what happens when it does?". As the Gimli glider incident shows, Boeing seem to have done a reasonable job of answering this question in a non-life-threatening manner. What remains to be seen is what happens in the case of true "fly-by-wire" aircraft such as recent military types and the A320, which are neutrally stable (I believe that's the term, apologies if not) and therefore difficult or impossible to control in the event of a software failure, regardless of what else has gone wrong...?