Re: draft-housley-two-maturity-levels

[Peter Saint-Andre <[e-mail]>]

On 10/27/10 10:35 AM, Dave CROCKER wrote:
> 
> 
> On 10/27/2010 8:53 AM, [e-mail] wrote:
>> three level is one level too many. Simplifying things and
>> eliminating process clutter is helpful in and of itself.
> 
> 
> By my reading of the proposal, this means that any spec with a couple of
> interoperable implementations can become a (full) Internet Standard.
> 
> This means that the assignment of that final status has nothing to do
> with real-world deployment and use, or even inclusion in products.
> 
> In other words, it has nothing to do with demonstrated utility.
> 
> Is that really what the IETF community wants?

I think there is a disconnect between the proposal as currently
structured and the core principle that the IETF community has always
professed: "rough consensus and running code".

As I understand it, running code means that the technology has been
deployed across the full breadth of the Internet, in services large and
small, by individuals and companies and service providers and
universities and government agencies and all the other kinds of entities
that are connected to the network. The technology doesn't need to be
universally deployed, but it does need to be widely deployed.

We can't achieve that kind of deployment in 6 months. We might achieve
that kind of deployment in 6 years.

We also can't measure that kind of deployment through rather informal
reports from the producers of two interoperable implementations. And if
we measure the wrong thing, we'll get the wrong results.

I don't think anyone in the IETF community wants to see the term
"Internet Standard" applied to a toy protocol merely because the
original designer just happened to convince one of his friends that it
would be cool to implement that protocol in a second hobby project.

"Internet Standard" is a weighty designation, and I think that to earn
it a technology community needs to show some weighty proof.

So I'm in favor of two maturity levels because it recognizes reality and
simplifies things in a reasonable way, but I think we need to reflect
more carefully and completely on how we determine if something is indeed
an Internet Standard.

And no, I don't yet have text to propose, although I'll get to work on
that...

Peter

-- 
Peter Saint-Andre
https://stpeter.im/