Re: More spec-ulations

[[e-mail]]

On Thu, 26 Mar 1998, Jon Callas wrote:

> At 11:41 PM 3/25/98 -0500, [e-mail] wrote:
>    1. If we are going to have a lame cipher for conventional encryption, why
>    not also have one for the PK algorithms, e.g. DH, RSA, and the
>    conventional key material plaintext as an MPI integer (for exportable
>    samples).  Would we need this for hashes and/or signatures too?  Any weak
>    (ROT-N like) alternatives?
> 
> Or we just remove the lame algorithms? There are places where it makes
> sense to have unencrypted secret keys, and these were the only places
> algorithm 0 has actually been used. I'm planning on noting such in the spec
> and striking the lame algorithms.

Yes and no.  The string to key has to use conventional encryption
algorithms, and no current implementation I know of really sets this,
rather it is left to default to IDEA/SHA1 (I use unpublished globals
preset to these).  This also creates interesting problems moving secrings
between implementations.

Maybe some additional text somewhere:

Conventional Algorithm Byte:  For secret keys 0 means unencrypted.  It is
invalid in any other usage.

--- reply to tzeruch - at - ceddec - dot - com ---