Dash wrote:

> On Oct 4, 9:13 pm, rober...@hushmail.com (Walter Roberson) wrote:
>> In article <1191506453.632349.193...@n39g2000hsh.googlegroups .com>,
>>
>> jma <junkmailav...@yahoo.com> wrote:
>>> I am pretty new to the subject so if I write sth stupid just point it
>>> out to me loud... I want to encrypt chunks of data that include ascii
>>> character. So it's not plain text but rather a binary stream. My
>>> question is whether an algorithm like AES is efficient and what
>>> problems I may face.
>> Discussions about the relative efficiencies of various cryptography
>> algorithms are probably best referred to sci.crypt .
>>
>> If efficiency is your main concern, then use NES instead of AES.
>> NES (NULL Encryption Standard) is as efficient as you can possibly
>> get, and can be easily written in a one or two lines of code:
>> just copy the input buffer to the output buffer unchanged.
>>
>> If that doesn't sound quite as secure as you would like, then chances
>> are that you are willing to give up at least a -little- bit of
>> efficiency for the sake of security. If so, then you should start
>> asking yourself how *much* security you need, and what kind of
>> events you are attempting to secure against, and how much inefficiency
>> you are willing to put up with to achieve that level of security.
>>
>> AES is considered to be fairly efficient for the level and
>> variety of security it gives you -- but it might not be the
>> appropriate level of security for you, and the drawbacks of it
>> might not be acceptable in your situation.
>
> A general question ... are you from Microsoft?
> I am asking you this question because you have spoken line a true MS
> techie but provided no help at all to the poor soul who asked the
> question.


I'd say he gave him the right advice quite nicely:
- He shall make up his mind about his demands, concretely formulate them and
then ask again (if the answer doesn't already jump into his face immediately).
- If you don't know what cipher to take: Simply use AES.