Hi SM, thanks for providing some feedback.

> -----Original Message-----
> From: [email protected] [mailto:[email protected]] On Behalf Of SM
> Sent: Monday, November 07, 2011 1:22 PM
> To: [email protected]
> Subject: Re: [marf] Working Group Last Call on draft-ietf-marf-redaction
> 
>  From Section 2:
> 
>    "4. Compute a digest of that string with any hashing/digest algorithm
>        such as SHA1"
> 
> A reference to RFC 6234 could be added for "SHA".  I'll leave it to the
> security folks to determine which algorithm to pick.  Using any
> hashing/digest algorithm will be viewed as unsecure.  It would be
> better to mention the minimally acceptable one.

Rather than RFC6234, I'm changing this to "such as one defined in 
FIPS-180-3-2008", since that's the SHA reference DKIM used.  I've copied the 
actual reference information over.

> I suggest separating Section 3 into different sections, one for
> security and the other for privacy.  The current section only covers
> privacy.  Does redaction (see algorithm) create any security issues?

Yes, your suggested split is appropriate, I think.  The IESG might find the 
current section name a little unusual.

The most common issue brought up when using message digests in a protocol is 
the impact of collisions.  I've added a paragraph about that post-split.

>  From Section 3:
> 
>    "If further protections are required, implementors may wish to
>     consider establishing legal contracts or other non-technology-based
>     agreements between the relevant entities."
> 
> This is a technical specification.  Legal contracts do not increase the
> level of privacy.  It would be better to remove that sentence.  I
> suggest focusing on the information disclosure angle if the aim is to
> have a privacy considerations section.

I've changed it to "establishing out-of-band arrangements between the relevant 
entities."

> BTW, the title of the draft is "Redaction of Potentially Sensitive Data
> from Mail Abuse Reports".  My reading is that the algorithm is to only
> redact the local-part of an email address (message header and body).
> Most, if not all, the WG participants know how to circumvent that. :-)
> The title could be "redaction of email addresses from mail abuse
> reports".

The algorithm can redact anything, though the local-part is the obvious and 
most common target.  In that sense, I think the present title is more 
appropriate.  Are there other opinions about that?

-MSK

_______________________________________________
marf mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/marf

Reply via email to