Re: Is RC4 a good choice?
From: Tom St Denis (tomstdenis_at_iahu.ca)
Date: 08/05/03
- Next message: Thor Russell: "reversible whole number license key"
- Previous message: Mrsjunecarey: "Re: Is RC4 a good choice?"
- In reply to: Mrsjunecarey: "Re: Is RC4 a good choice?"
- Next in thread: Mark Wooding: "Re: Is RC4 a good choice?"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]
Date: Mon, 04 Aug 2003 23:10:50 GMT
Mrsjunecarey wrote:
>>Subject: Re: Is RC4 a good choice?
>>From: mdw@nsict.org (Mark Wooding)
>>Date: 8/1/03 5:22 PM GMT Daylight
>
>
>>It's an /enormous/ code book.
>>
>
>
> Yeh but it's still a code-book.
> And it still sucks.
Science practices like this only make me wonder how civilization is
surviving this long.
>>Why's what a problem? There's nothing which says that a message has to
>>be a whole number of blocks in length, for example. Consider modes such
>>as CBC-with-stealing, CFB, OFB, OCB, etc.
>>
>
>
> It's inconvenient.
Ciphers can also be used as hashes. [or hashes as ciphers!]. Stream
ciphers are mostly only useful as ciphers
Also if you cannot encapsulate a cipher in CTR mode you shouldn't be
writing software.
>>Of course, stream ciphers are always trivial objects. Compare AES or
>>Blowfish with SEAL, for example.
>
>
> SEAL uses a cryptographic hash function which means it sucks very badly; almost
> as bad as a block-cipher, but not quite.
And your cipher uses horribly jibberish non-portable C++ code. Yea, way
to chalk one up for your argument.
BTW RC5 is a block cipher that is WAY simpler than Lxy.
>>[theory lesson deleted]
>>
>>Counter mode turns a PRF into a PRG. Thus, a block cipher in counter
>>mode /is/ an additive stream cipher.
>
>
> Wrong.
>
> A block cipher in counter mode is just a block cipher in counter mode.
>
> It is only becomes a binary additive stream cipher when you XOR its output with
> some plaintext.
Um, what did you think you did with the key stream? Same thing for L14,
I could use it as a PRNG [I wouldn't but I could] which means it
wouldn't be an additive stream cipher.
>>A professional should know the things I've described above, and in the
>>formal terms used in the literature rather than the watered-down
>>handwaving I've used above.
>>
>
>
> You're assuming that when I said "professional" I was meaning "professional
> cryptographer".
>
> And I never said that.
Well as an auto salesman you're still not that convincing!
Tom
- Next message: Thor Russell: "reversible whole number license key"
- Previous message: Mrsjunecarey: "Re: Is RC4 a good choice?"
- In reply to: Mrsjunecarey: "Re: Is RC4 a good choice?"
- Next in thread: Mark Wooding: "Re: Is RC4 a good choice?"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]
Relevant Pages
|