Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Posting-Version: version B 2.10.1 6/24/83; site denelcor.UUCP Path: utzoo!watmath!clyde!burl!ulysses!mhuxl!houxm!houxz!vax135!cornell!uw-beaver!tektronix!hplabs!hao!denelcor!lmc From: lmc@denelcor.UUCP (Lyle McElhaney) Newsgroups: net.space Subject: Re: Chicago Tribune Article Message-ID: <501@denelcor.UUCP> Date: Tue, 10-Jul-84 14:12:39 EDT Article-I.D.: denelcor.501 Posted: Tue Jul 10 14:12:39 1984 Date-Received: Fri, 13-Jul-84 03:20:00 EDT References: <1434@sri-arpa.UUCP> Organization: Denelcor, Aurora, CO Lines: 25 No, the computer will not go into the scrap heap (yet, anyway). Much more important (at least in the NASA way of doing things) is to discover exactly WHY the computer failed. It will most probably be sent back to the manufacturer to be tested to death in order to find out whether the fault was a fluke (and why) or whether it is a design error. Unresolved failures in a piece of man-rated hardware are very serious. If the problem is found to be a single error in that computer alone, and can be fixed without stressing other components, and the computer can be requalified, then it may join the available spare parts list. Testing may well stress it beyond its useful life, and it will be scrapped, but not before the secrets of its problems have been squeezed out of it. This is one of the major reasons why the exploration of space has been so expensive. In a commercial operation, that computer would be tested, and if it failed, it would be scrapped and replaced. Another $1.2 million burned. NASA will more than likely expend far more than the $1.2 million tracing the error. The largest part of that cost per computer, in the first place, is the paperwork that the manufacturer has to generate on each piece of hardware in order that a fault can be traced to the exact step where it was introduced. Yes, this probably kept a few astronauts from being killed in space; but it has made space exploration a very expensive proposition, indulged in only by governments. -- Lyle McElhaney (hao,brl-bmd,nbires,csu-cs,scgvaxd)!denelcor!lmc