Re: Reading Middle Eastern Characters
Re: Reading Middle Eastern Characters
- Subject: Re: Reading Middle Eastern Characters
- From: Christopher Nebel <email@hidden>
- Date: Tue, 7 Dec 2004 10:50:01 -0800
On Dec 7, 2004, at 1:46 AM, Elliotte Harold wrote:
Christopher Nebel wrote:
Correctly determining which of the dozens of conceivable text
encodings a given hunk of data uses is essentially an AI-complete
problem -- that is, you need human-level intelligence, and even most
humans would have a hard time with some of the fringe cases.
Oh, I don't think it's nearly that hard. Simply spell check the source
text against a variety of dictionaries in a variety of encodings and
languages until you find one that matches. Cryptographers have been
doing this for decades without solving the AI problem. They have all
sorts of algorithms for determining whether a brute force description
of a message with a particular key actually represents the clear text
or random gibberish.
True, I overstated the problem -- the practical point is that for most
projects, it's more trouble than it's worth. However, I'd point out
that (as any good cryptanalyst can tell you) getting decent results
presumes having a sufficiently large sample of ciphertext to chew on.
Very small samples, such as filenames or volume names, can be
problematic.
--Chris Nebel
AppleScript Engineering
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Applescript-users mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden