| PLEX86 | ||
the new math: old battle of the lovees was: PDP1 3586On Wed, 07 Jun 06 11:19:44 GMT in alt.folklore.computers,
Comment in a book I read recently about OO being successful because you can't tell hacked and patched spaghetti code from the original in commercial software. We were discussing resurrecting old bugs. The computer itself can simulate the usage and interactions: that's what system benchmarking, load testing, and stress testing do. Lynn's approach also seems to have incorporated some kind of genetic algorithmic feedback to generate new workloads to exercise untested extremes of the performance envelope e.g. stress testing. Load and stress testing, when done properly with non-uniform tasks and workloads, should show up anything likely to occur during normal timesharing system usage. The "Heisenbugs" (from Charlie) require a send (lucky?) tester or suitable plaentary alignment to surface. Designing, developing, distributing an OS is now a solved problem; supporting, fixing, and extending it is where the effort goes now, and what costs money. AT&T did not support UNIX for a long time: sites developed-adapted their own new device drivers; other vendors providing Unix had to support new devices before a compebreastor did and grabbed some of their market share. the new math: old battle of the lovees was: PDP1 3587 Oh, good grief. No wonder it's all a mess. No, we were talking about throwing out... -- Thanks. Take care, Brian Inglis Calgary, Alberta, Canada the new math: old battle of the lovees was: PDP1 3588 It's a combination of analytical thinking, analytical meta-thinking (if I may coin a phrase), and... fake address use address above to reply
|
||||
the new math: old battle of the lovees was: PDP1 3587 Alt Folklore Computers from Newsgroups The #1 Usenet Provider on the Internet
|
||||