Date   : Mon, 30 Jan 2012 09:18:35 +0100
From   : rick@... (Rick Murray)
Subject: CRC-8 algorithm

On 27/01/2012 05:07, J.G.Harston wrote:

> I specifically didn't want to just pick one (at random?),

If you're implementing your own CRC routines, and you are using it for 
your own purposes (i.e. not interoperability with something else), then 
does it matter *which* method so long as it works?

I'm afraid I don't understand maths well enough to know why one 
polynomial is better than another; however the end result is that you 
would like a hopefully unique number for <random block of data> so data 
corruption can be easily detected.

Thing that worries me, what is your block size? If you're using an 8 bit 
CRC, that's only 256 distinct values. How likely is it that a corruption 
could result in the *same* CRC as was expected? Have you tried the CRC 
with injecting some garbage data, and also runs-of-zeroes and 
runs-of-bit-swap (etc)?


Best wishes,

Rick.

-- 
Rick Murray, eeePC901 & ADSL WiFI'd into it, all ETLAs!
BBC B: DNFS, 2 x 5.25" floppies, EPROM prog, Acorn TTX
E01S FileStore, A3000/A5000/RiscPC/various PCs/blahblah...