On Wed, Mar 14, 2012 at 12:05:01PM -0700, Barry Gershenfeld wrote: > You've examined the costs, but what about the benefits? >=20 > There's only a narrow area where a simple checksum misses an error that a > CRC would catch. According to the paper, it's not actually so narrow. It's the reason that CRC "dominates" Fletcher's in all instances and all Hammond Distances. > What error rates do you expect? Most digital connections don't make > errors, and when they do, the term "noise" often means "I haven't taken t= he > trouble to find out what the problem is". Now, when there's an analog > link (telephone, radio, modem), then a CRC is a very good idea. I can se= e > that a scheme which times the width of pulses might be counted as "analog= " > and therefore subject to jitter, aka, noise. Like I said, it's a 6 foot serial cable. I fully expect zero errors.=20 >=20 > What is the risk of a bad download? Medical devices, certainly you deman= d > liability. Lab experiments and consumer toys, not so important. Maybe on= ce > in its lifetime there will be a bad load, and trying the download again > will fix it. Which sounds an awful lot like your failure scenario with C= RC > anyway :) It's a development setup for educational testing. The only harm will be user frustration. This is one of the reason's I threw the subject out to the list. Sometimes it's too easy to overthink the problem, or to get a fresh perspective. I've already gotten two good options (don't worry too much, use bit by bit CRC because there's enough time and the cost is minimal). Thanks. BAJ >=20 > Barry > --=20 > http://www.piclist.com PIC/SX FAQ & list archive > View/change your membership options at > http://mailman.mit.edu/mailman/listinfo/piclist --=20 Byron A. Jeff Department Chair: IT/CS/CNET College of Information and Mathematical Sciences Clayton State University http://cims.clayton.edu/bjeff --=20 http://www.piclist.com PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist .