al@ames.UUCP (07/26/84)
>I very much agree with the let's get a piece of hardware and bang on it >philosophy... >What is needed in space systems is what in other fields we call Engineering >common sense... >My attitude from years of work in automation systems is that the best >results are gained by getting prototype hardware and software together as >earlier as possible. Your team gets experienced with the nature of the >specific problem at hand, and then you use the prototype as a basis for a >REAL system specification. This philosophy works fine on the ground, and will work fairly well in the space station era. Unfortunately for current space operations, there is a major flaw. The problem is that it is not possible to test things in their target (orbital) environment. Specifically, thermal issues are much different in orbit, there is tremendous vibration during launch, lack of atmosphere changes things a great deal, there are no repairmen available on short, if any, notice, forces are much different (the shuttle arm cannot lift itself in 1G, how do you test it?), and so forth. If we had a workbench in space to debug prototypes we could do things the way suggested, but we don't. As a result one must rely on analysis, simulation, and over design. To make a bad situation worse, the data base for analysis and simulation is frequently sketchy and extremely expensive to extend. To imply that space systems engineers do not use 'Engineering common sense' is insulting and inaccurate. I ask the author - have you ever carefully inspected a single space system design or talked in depth with a single space system designer? In short, what data are you basing your accusations on?