[Cryptography] rot13 for the 21st century

[Dennis E. Hamilton]

From: Stephen Farrell
Sent: Tuesday, September 28, 2021 06:05
To: cryptography at metzdowd.com; Viktor Dukhovni <cryptography at dukhovni.org>
Subject: Re: [Cryptography] rot13 for the 21st century

> On 26/09/2021 05:08, Viktor Dukhovni wrote:
>> On Thu, Sep 23, 2021 at 11:14:55PM -0400, John Levine wrote:
>>
>>> Fans of the traditional rot13 cipher should check out rot8000, same
>>> idea except that the character set ie expanded to 16 bit Unicode, and
>>> you rutate by 0x8000 rather than 13.  It has the nice characteristic
>>> that encoded ASCII looks like Chinese.
>>>
>>> Try it out here:
>>>
>>> http://rot8000.com/
>>
>> Cute.  Thanks!
> Yep. Wonder if there's a string that's meaningful in both versions?
> S.
>>
>> ????? ??? ???????? ??? ????
>>
[orcmid]
Methinks we are up against code page detection and whatever the browser/mail 
writer/reader is set for.
While the computation works in unsigned 16 bit values, it is best to employ 
UTF8 in the encoding of text sent and received.  This will prevent spurious 
code-page selections, endian problems, and ISO 8859-x guessing.  It will also 
make the rot8000 reversible once converted to wide characters.

It is an interesting challenge for edge-to-edge confirmation of text 
communication as experienced by sender vs. receiver.  This is one of those 
incidents where "works for me" is a spurious resolution.

 - Dennis