Create key's over 4096 bit ????

Christoph Anton Mitterer cam at mathematica.scientia.net
Thu Dec 22 13:43:05 CET 2005


Johan Wevers wrote:

>Christoph Anton Mitterer wrote:
>  
>
>>- And even from a cryptographic point of view this wouldn't make sense 
>>(as far as I know), as currently hashfunctions are the weak point of the 
>>whole system.
>>    
>>
>
>That depends on what you consider important. Hash functions are only used
>for signing; for encryption, currently the 256 bit algo's are the strongest.
>  
>
Yes and no,...

(btw: The strongest has should have 512 (SHA512), or am I wrong?)

It is true that you don't directly use hash functions when encrypting data.
But you need it indirectly too.
If you encrypt to another key,.. your implementation is going to check 
the validity of that key (either you've signed/certified it yourself or 
via some trust-path). And these certificates are "bound" to the hash...

Ok,.. you could argue that one use its key for local encryption only,.. 
but perhaps one should use other tools for that task...


Chris.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: cam.vcf
Type: text/x-vcard
Size: 449 bytes
Desc: not available
Url : /pipermail/attachments/20051222/dd0e65a6/cam.vcf


More information about the Gnupg-users mailing list