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

Re: Evolution - was VMs: Inks and retouching



On Monday 02 August 2004 13:05, John Grove wrote:
> Thanks for the character counts. You reminded me of the one 'major'
> problem I have with a natural language solution: lack of double letters or
> double syllables. They are far too rare in the VMS.

Yes, they are not very frequent (except the suspected e, i). Here they are:

ee 5395
ii 4769
hh 84
oo 76
ll 34
ss 31
dd 25
aa 5
rr 2
yy 3
mm 1
kk 1
cc 1

Note, that I do not think that the vms is written in a "natural language" in 
the sense of it being a simple substitution (because of the low entropy), yet 
there are a number of statistical properties that are shared with natural 
languages and so these have to be explained somehow.

I am still puzzled by Cappelli's dictionary. I like the possibility of <y> 
being two different meanings, depending where they are in the word (as noted 
in D'Imperio <y>- =con, -<y>= us). This, I guess, introduces a false sense of 
redundancy (and helps to keep the entropy low?).

Another: I cannot see many problems with the word substitution based on a 
dictionary (one problem is word repetitions, I know). Note that this does not 
require any word construction rules as long as the words correspondences are 
unique and at the same time it allows the low entropy to be there.

For an extreme case:
AND = eeeee
THE = eeeeb
OF =  beee
FOR = eeb
etc.

Of course, I have no evidence that this could be the case of the vms but shows 
that it is possible for the ms not being low in information as originally 
suggested by Bennett. The problem may be (as you are also aware) only the 
representation (but this is not new, of course).

Cheers,

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