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

Re: VMs: VBScript for finding repeating strings

> (4) You might consider replacing both probable glyphs (like "cfh",
> "ch", "sh", "ee", etc) and probable verbose pairs (like "qo", "dy")
> with non-EVA letters (like "A", "B", "C", ...), and reducing your
> cut-off length, to see what other patterns you notice. This might
> reduce your reliance on "qoteedy" matches, and hence help show up
> other kinds of matches within the text.

I have done something else, but I hope it will give you the same

I'm running the repeating string analysis again, but this time not at the
character level, but on the word level. The algorithm searches for repeating
occurences of word pairs. So:

wordW wordX wordY wordZ wordX wordY wordW

Will give one match for "wordX wordY" at distance 3, coordinates (1,4)

As you can see I get the "black triangles" again in the same places. This
way I'm ignoring all glyph substitutions, since I'm working at the "whole
word" level. Does this satisfy you?

The plots of the results are here:


(I hope my server protection doesn't block acess to these files.)

Here is the output from the matching algorithm. You can see which word pairs
the algorithm flagged:


The list of input words and their positions is here, you can find which
matches the algorithm has flagged. And it should be quite easy to locate the
corresponding sections of the VMS. I leave that as an exercise for the
readers (please?) :


And the script itself is here:



Why word pairs? That looks like another random cut-off level.
Well, if I look for all matches of single words, I get so many matches that
Excel isn't able to produce a plot anymore. Looking for pairs reduces the
amount of matches drastically.


> I think this technique was at the heart of finding key lengths etc,
> first applied by the Friedmans?

This sounds interesting. Which key lengts, Vigenére?
Where could I find more information on this?


BTW - Thanks everyone for the feedback and the suggestions. Tomorrow I'll
stop experimenting for a while, I have a lot of other urgent things to do
... but feel free to continue where I left off.


> If you have not done so, right click on a point in the graph. Select
> "format data series" from the menu. Find the "size" window and set to
> lowest value. Note the "style (dot shape)" and "foreground (color)"
> options.

Yes, That's the button! It's a pity Excel can't handle more than 32000


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