Re: Apple ][gs encryption

[neufeld@helios.physics.utoronto.ca (Christopher Neufeld)]

In article <1850728349.5586912@adbbs.adbbs.antioch.edu>,
John L. Graham <John_L._Graham@adbbs.antioch.edu> wrote:
>If you're trying to protect your files from non-Apple II users you might just
>use ShrinkIt.
>
   No, ShrinkIt, or rather 'nulib', exists for any machine with a C
compiler. Security by obscurity is never a good strategy.
   For general encryption of files containing 8-bit data one might first
BinSCII the file and then use the encrypt option of MicroEMACS. If you're
really worried about brute-force attack on the password you might strip
out the FiLeStArT line, since that represents a known string which can be
checked quickly. Without that there, the brute-force attack would have to
decrypt enough of the file to allow a BinSCII check of the checksums, and
so would slow down the attack. I seem to recall, though, that the
MicroEMACS encryption algorithm uses some embedded checksum information,
and that it can tell you that you've given it the wrong password, so this
might not buy you anything.
   In any case, my approach is not one you want to use if you have to
read those files frequently.
   A few other points to consider: don't decrypt the file to a temporary
disk file, or if you do, make sure to overwrite the temp file with
garbage before deleting it, otherwise a sector snooper or exhumation
program will get the plaintext of your protected program with little
difficulty. Similarly, don't let the stuff sit in memory on an unguarded
machine, it's often fairly simple to get at the contents of a ][GS's RAM.
Oh, and don't write the password down on the side of the computer
monitor.

   If you're ambitious, get the sources to GNU 'crypt' and
'uu{en|de}code', and combine them to a program which reads a file from
disk and puts the resultant 8-bit file in the RAM disk (keeping in mind
what I said about leaving stuff in RAM).
   If the files you're trying to protect are flat ASCII you can forgo the
uuencode step, but I'm not certain that 'crypt' can handle 8-bits, that's
why I suggest that step.

-- 
 Christopher Neufeld....Just a graduate student   neufeld@physics.utoronto.ca
 "Don't edit reality for  | "Genius may have its limitations, but stupidity
 the sake of simplicity"  | is not thus handicapped."  -- Elbert Hubbard