[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: VMs: VMS Lookalike encipherment

[Jeff Haley: One method that would achieve this effect is to first strip all spaces from the plaintext and break it into groups of alphabetically ordered strings.]

KM: That is an interesting concept.

[Jeff Haley: The dead letters are nulls. I can't say this is definately the way a medieval cryptographer would have done it but it is simple. The nulls would have masked the meaning a little.]

KM: This, along with the alternate substitutions, should reduce the number of chains/phrases/repeated-word-series that are transferred from plaintext to ciphertext, which is needed (not forgetting Marke's chains). At the same time, it would leave bigrams and the amalgamated trigrams-to-bigrams intact, also needed. I do not know whether some CT trigrams would be sufficiently increased and others reduced.

[Jeff Haley: The word chol is very common in the manuscript so it was assumed at this time that the group cho was not in fact a verbose group.]

KM: But *Ao* in the table is considered a verbose group (VG)?

If *cho* and *ol* are VG's, *chol* would always be understood by an intended reader as *cho+ol* because there is no *chool* -- discounting at least three such strings in the Takahashi transcription in favor of the Majority Vote. The same could apply to other letters. That leaves *iii* ambiguous unless the glyphs can be interpreted differently (as they have been). In that scheme *eee* would have to be a VG distinct from *e* and *ee*. All conjectural on my part. Might fit in somewhere.

The trouble with n-grams for me is that letters overlap. *Ao* leaves *r* and *l* standalones. I would guess, though, that the fewer resultant VG's, the better. (?) I was only able to reduce the number of substitutions to around 60 or 65 by using any consistent method.

[Jeff Haley: I do not believe however that decipherment would be impossible as the choices are binary.]

KM: Maybe. If there are not too many. Ambiguity was not unknown in certain types of script. You might be able to remove some of it. This experiment might also lead to reduced ambiguity re. Latin, as well:
[Elias Schwerdtfeger: Eventually I'll try other tables to encrypt german or english text,]
Another script to decipher would be helpful and show how much ambiguity there might be in a long text.


[Jeff Haley: Only time will tell.]

KM: Many hurdles to jump yet. In any event, tying all this together is good work.


To unsubscribe, send mail to majordomo@xxxxxxxxxxx with a body saying:
unsubscribe vms-list