Re: HP33S Checksum Message #8 Posted by Ivan Nejgebauer on 9 June 2006, 7:31 a.m., in response to message #1 by Derek
Quote:
Can someone possibly tell me how the label check digits on a 33S are calculated? (shift left -> mem -> 2 ->shift right -> show). It's a long story but basically, I need to duplicate the process programmatically.
Ah, that's one of the reasons I wanted to get a 33S myself (another long story.) I don't have a ready answer to your question, but I do have an idea how to identify the checksum algorithm, if it is a common one. I believe it is, because
- By all accounts, development of the 33S firmware was a bit of a rush job, as witnessed by the number of initial bugs, complete disregard for memory efficiency and the (relatively) short time frame.
- That being the case, I seriously doubt that any significant time was spent on analyzing the frequency and type of program entry errors and finding some funky CRC polynomial that would minimize checksum collisions.
Therefore, I am reasonably certain that the chosen algorithm was something well known and widely available, such as CRC-32 or Adler-32. These are 32-bit checksums, so probably the upper 16 bits of the result are discarded or exclusive-ORed with the lower 16 bits.
So, how to find out? Depends on what is the shortest program you can write and get the checksum for. Let's assume it is a label alone, e.g.,
A0001 LBL A
You know it is three bytes long, and you know the checksum. You don't know the byte values, nor the algorithm. Suppose that the algorithm is CRC-32, and check all 2^(3*8) values to see which ones produce the known checksum. You will get a number of collisions, so try to find a value which makes sense for the label step.
Once you have this tentative value, extend the program by a single instruction, say RTN, and repeat the process. Try another algorithm if the assumed one gives nonsensical results, etc. Mapping the entire instruction set (and constants, and in-program equations) is likely to be gruelling work, but I believe it could be done.
Regards,
I. Nejgebauer
|