Re: explain to an idiot why gpg output slates towards 0xd5


jiminyor...@xxxxxxxxx wrote:
> thanks for your explanation. but as far as suggesting
> my counting program is flawed beyond my apparant lack
> of programming skills, it just aint.

Well I ran "gpg -c test.src" [test.src is actually a copy of the
counting program] and the resulting test.src.gpg even with the constant
passwd of "1" is different each time and passes the eyeball test.

> i've been getting this dismal result with walkers ent
> program and with the counting function in winhex.

Um, how big is the file you are encrypting? If it's small you can get
a "sizable" [or local] skew to one byte [but shouldn't be the same
byte].

> here is what 70 megabyte of isaac output looks like in ent.exe
>
> E:\>ent 70mb
> Entropy = 7.999997 bits per byte.

> here is the same isaac file encrypted with gpg
>
> E:\>ent 70mb.gpg
> Entropy = 7.999995 bits per byte.

That's such a bias....

As for the list of all chars ... How many times have you run this test?


Either your build of GPG is broken, your test is somehow flawed or your
lying about the results. I mean I did 2 mins of work to eyeball the
most freq chars and it changes each time I encrypt. There is no
consistent bias towards byte 0xd5 or any other.

Tom

.

Relevant Pages