* If you want to get really complicated, encode the file *twice* with two
keywords of _different_ lengths. It doesn't matter what order, i.e. if you
encode first with 'abc' then with 'wxyz', you can decode with either
'abc' or 'wxyz' first, as long as you do both. The effect is as if you
encoded with a very long keyword, making it hard to detect a pattern.
* _DON'T_ use keywords like 'abc' or anything less than three letters.
* _DON'T_ insert or delete any *codeable* characters to the file after it's
been coded. CRYPT will ignore spaces, CR's and control characters, but any
letters or numbers will screw up the code.
Enhancements:
* Changed the code using some techniques I saw in QKRCHK.ASM. Main effect--
you don't have to type the '.DO' extension anymore. In fact, if you
type the wrong extension, it will be ignored.
* Changed the loader to allow for either relocating load OR loading to the ALT
LCD buffer. Had to shorten some prompts to allow for this. ALT is the
default location.
* I already said not to use 'cute' lines of characters, but never, _NEVER_ use
lines of @s or `s <Shift-2 or Grph-[>. They will show up as your keyword in
the coded file!
* Regarding EasyPlex/Forum mail or other BBS transmission of coded files--make
sure that your coded file meets the line length specifications of whatever
editor you use. If you upload a coded file with lines that are too long,
characters may get chopped off the end of the line, thus putting the coded
file out of sync with the keyword.
I think these refinements make CRYPT pretty much complete <for most purposes>. I would encourage others to write similar programs using the same algorithms <see CRYPT.SRC> for other computers <an IBM-compatible version might come in handy for some of you>. At any rate, I think I'm done with this for a while - at least till I come up with something better!