Archival storage

On Tue, 2005-10-11 at 22:19 -0700, Fred Cisin wrote: Sort of. AFAIK DNA does not self replicate, that is if there is some DNA laying around it won't just start making copies of itself. There is a lot of supporting machinery in a cell that needs to be in place. Some things in the cell aren't even coded for in the main DNA like mitochondria. So maybe the more accurate analogy would be a cell, with the DNA as payload but with the payload containing information and description for making all necessary machinery to decode the data and make copies. Cell/DNA analogy only goes so far, since although it is data, bit errors are tolerated and the organism deals with it or dies. In an engineered system no data loss is permitted (though error correction is OK). The issue I was seeing was that there seems to be a requirement for some bootstrap code that is never corrupted. This bad since the whole point of this is that we trying to tolerate errors by adding error correcting codes. Is it possible to make executable code which is resistant to bit errors? Say just assume X86... is it possible to write some minimal bootstrap code that can tolerate errors in itself, but run long enough w/out crashing the machine to decode a secondary boot loader that is padded with error correcting code? That initial bootstrap code can be inefficient as hell... -- John.