| ??? 06/16/05 00:54 Read: times |
#95054 - Design Verification Responding to: ???'s previous message |
" It is a known fact that the fact that something "works well" is not a verification of a design. "
Agreed - but with some reservations. Like when the prototype is the end product ( non-mission critical ). The cost and time involved in setting up scenarios for simualting errors/faults/ failure mode analysis may not fit into the scheme of things. In such cases is it not customary to accept the design if the product does what the user wants and exits gracefully for all "known" faults ? Since most of my designs fall into the one-off category, I am sincerely interested in knowing more on proving the design in such cases. Thanks Raghu |
| Topic | Author | Date |
| AT89C51RD2 without supervisor IC | 01/01/70 00:00 | |
| RC is never good | 01/01/70 00:00 | |
| watchdog ad absurdum | 01/01/70 00:00 | |
| and uses a bunch of pins | 01/01/70 00:00 | |
| wd | 01/01/70 00:00 | |
| paranoia revisited - wd 2 | 01/01/70 00:00 | |
| when the full story | 01/01/70 00:00 | |
| WD3 | 01/01/70 00:00 | |
| if code run astray | 01/01/70 00:00 | |
| Has been discussed before | 01/01/70 00:00 | |
| External ones are better | 01/01/70 00:00 | |
| External ones are better in one respect | 01/01/70 00:00 | |
| Yes, but... | 01/01/70 00:00 | |
| Nope, never use an external when | 01/01/70 00:00 | |
| Exactly! | 01/01/70 00:00 | |
| Combined Watchdogs | 01/01/70 00:00 | |
| External better in another respect | 01/01/70 00:00 | |
| I'm using the internal and works well | 01/01/70 00:00 | |
| works well - how do you know | 01/01/70 00:00 | |
| who should test it? | 01/01/70 00:00 | |
| nobody | 01/01/70 00:00 | |
| the manufacturer should | 01/01/70 00:00 | |
| Design Verification | 01/01/70 00:00 | |
one off | 01/01/70 00:00 | |
| The meaning of "well" | 01/01/70 00:00 |