ietf
[Top] [All Lists]

[ietf-dkim] Re: WG Review: Domain Keys Identified Mail (dkim)

2005-12-22 00:56:30

We have had three proposal for some text on changes to prior work, the
current proposed charter, the text from the XMPP charter, and the text Keith
provided below.

The question that I think IESG should be asking themselves is how is this
similar and/or different from other groups the have chartered or will in the
future. Nearly every group has some people with a fairly strong idea of at
least one way to solve the problem. Without this, it is usually not clear
the work is even possible. Now some groups, XMPP for example, perhaps TLS
long ago, have substantial deployment with difficult backwards compatibility
questions - theses situation might require the charter to provide more than
normal limitations to the scope of the solutions that are possible. I'm
failing to see that dkim has existing deployments or difficult backwards
compatibility problem that would cause the need for some special text in the
charter more than you average WG. (They do have difficult backward
compatibility problems with existing email systems and I like the text that
limits the scope on that.) My current understanding is that the deployments
are small enough that changes are still easy and that non backwards
compatible changes are already expected. I fail to see the analogy to XMPP
but perhaps there is a good reason for something more like XMPP. I'd sure
want whoever approved this charter to be able to give me a clear reason why
they though this WG would produce better work by having this constraint.
It's going to set a precedent for things to come.

Practically speaking, I'm not sure it will make much difference to the work
that the WG produces. If the individuals that will form the WG feel that the
WG has consensus that they will not make changes beyond a a certain
boundary, they don't need the charter to set that boundary. The strong
arguments that this needs to be in the charter makes me wonder if the
individuals that will form the WG will agree to very limited changes or not.
If they will, it's barely worth arguing for here.

Related to how much the charter pre-supposes the solution, the sentence that
"Public keys needed to validate the signatures will be stored
in the responsible identity's DNS hierarchy." seems like a pretty heavy
constraint on the possible solutions and one that some proposals disagreed
with. 

I'm not arguing against the current dkim drafts, I am arguing that the
future of doing internet drafts should not be a process where we come to
agreement in a dark and smokey room with a small group of people then set up
WG where effectively only syntax level changes can be made. If we like these
drafts as is, let's skip the WG and just take them forward as AD sponsored
individual drafts and call it done, if we think we need a WG, allow it to
have change control to select a reasonable solution to the problem.



On 12/21/05 4:07 PM, "Keith Moore" <moore(_at_)cs(_dot_)utk(_dot_)edu> wrote:


I suggest the IESG replace the paragraph with the following:

While it is understood that the WG will use the current DKIM
specifications as a starting point, the WG is neither expected nor
constrained to specify a standard which is compatible with those
specifications.  The WG should feel free to make whatever changes are
necessary to produce a specification that is robust with respect to the
requirements and flexible enough to support a diverse set of usage
scenarios.
_______________________________________________
ietf-dkim mailing list
http://dkim.org

<Prev in Thread] Current Thread [Next in Thread>