SDSC a écrit:First of all let me correct a misunderstanding here, DNA is not about compression but about
encoding and decoding. The main goal of DNA is to store data effectively on a local storage device.
So key codes are (yet) not exchangeable between systems without the corresponding reference table.
What do you mean by 'not exchangeable between systems'?
Is the reference table on your PC different from what it would be on my PC when I would run your software?
Or to put it more exactly, let's say you encode (if that's a more appropriate term than 'compress' in this context) a 1MB file to a 256 byte .dna file on your PC, using your program + the 720 MB reference table, and delete the 1MB original. Then you're able to decode the 1MB file back from the .dna file, using your program + table. Right?
Now instead, after encoding + deleting the original, if you would copy your program + reference table and the .dna file to another PC and decode it there, that should still result in the same original 1MB file, right?
And now what if, instead, you copy the program + reference table to another PC
before encoding the original file. Then encode the 1MB file to .dna on PC number 1, copy the .dna file to PC number 2 (which already has the program + reference table installed) and decode the .dna file there. If you're not using black magic of hardware dependent trickery, that should
still result in the original 1MB file, right?
This is a very interesting and good question, and the answer is NO. Within the reference table nothing of the original files can be found back, and without the key code in combination with the number of iterations the original file can not be restored. This makes DNA already a very solid encryption tool.
Ehm, but does the 720 MB reference table change when you encode a file? I.e. does the content of the reference file depend on what files you encode?
Cause in that case, it appears you're simply moving data around, which doesn't make data storage any more efficient? Or suppose I have 720 files of 10 MB each (or 72 files or 100 MB each, i.e. way more than 720 MB in total, you get the idea). Can you encode them all to 256 byte .dna files, and decode them all correctly, still using only
one 720 MB reference file?