10 Aug
2006
10 Aug
'06
12:16 p.m.
To make it on topic, I've found exactly the same
thing with classic computers
- with something that sold tens of thousands or more the bugs tend to get
ironed out; with something that hardly sold at all there can be all sorts of
errors. (that seems to be tied to the machine class though - e.g. mainframe
designers *know* that they won't sell many, so they set aside time to get the
service docs right and that gets reflected in the cost)
I do not think that is the reason. The engineering process for
mainframe computers is far more disciplined, in keeping with the RAS
philosophy. Sell one or sell a million - mainframe customers have very
strict demands for their systems, and the engineers must adhere.
--
Will