VMs: "Running encoding" -- Used in period? Properties?

[elvogt@xxxxxxxxxxx]

Hello all,

Inspired by Steve's "Fold the vellum" paradigm, I wondered about some kind 
of "running encoding" of any text. (Don't know if that's the right term -- I'll 
just call it that.)

The idea is essentially to write down a table of your cleartext letters:

A B C D E
F G H ...

Or scramble the letters, as you please.

Now what you write down is not the sequence of letters, but their relative 
positions.

Eg to encode "Bee", you would write

"Jump to 1st line, 3rd column
Move 2 cells to the right
Repeat last letter"

how to reach every letter from the previous cell.

Now you encode this _instructions_ how to jump from table cell to cell. 
(Essentially, you've got some 26 different moves, so for each move you could 
invent one particular symbol. Hint hint.)

The funny thing about this would be that different letter sequences could be 
mapped to the same encoded sequences (eg every doubled letter "tt", "ee", etc. 
would be transformed into a jump-instruction and always _the same_ "Repeat"-
instruction).

Now the questions --

I browsed some period encoding schemes, but didn't find anything like that. Is 
it true that none of these "running encoding" schemes was ever used? (I 
understand they would be vulnerable to the loss of individual letters, but you 
could recover by introducing codes to "synchronise", ie to give you absolute 
rather than relative positions again.)

Off the top of your head, what would the statistical properties of a text 
encoded that way be? Could this possibly explain the high repetivity of the VM, 
provided the right code table was used?

Just a thought during a slow day at work,

   Elmar


-------------------------------------------------
debitel.net Webmail
______________________________________________________________________
To unsubscribe, send mail to majordomo@xxxxxxxxxxx with a body saying:
unsubscribe vms-list