I'm also very much in favor of retaining the two-step process.
One potential improvement would be for PS RFCs to specify how to perform
interoperability tests, and where/how to report the results of
interoperability tests. A lot of the effort at moving to IS is
getting people to do the interoperability tests and report the
results. And if the tests were designed along with the protocols,
maybe that would create an incentive to make the protocols less complex.
Keith
On 3/17/19 5:07 PM, Carsten Bormann wrote:
On Mar 14, 2019, at 17:51, Spencer Dawkins at IETF <spencerdawkins.ietf@xxxxxxxxx> wrote:
… you might or might not love https://www.ietf.org/archive/id/draft-loughney-newtrk-one-size-fits-all-01..txt
Actually, I don’t agree with that document at all.
The statistics used as an (pretty much the main, actually) argument in that document are really not that relevant.
It argues that, at the time, we produced around a hundred new PS documents per year, and about ten new IS documents. So what?
I’m experiencing the upgrade of a standard from PS to IS right now.
In this process, we have the unique opportunity to really focus on interoperability.
The question is not what to add, but what we need to remove or (slightly) tweak because it didn’t quite interoperate.
The additional quality that we can instill to the final outcome of this process is significant.
So, yes, I see a substantial benefit in producing Internet Standard documents as defined in RFC 6410, and I will be a vocal proponent of keeping the two-step process, even it is only exercised by a small percentage of the standards documents and the fine points about it may not be understood by external experts.
In the end, the unique selling proposition of the IETF is the quality of the standards it produces, and the PS to IS advancement process is rather unique compared to the various review and advancement processes that other SDOs have.
My only critique of RFC 6410 would be that it wasn’t radical enough in revisiting the names we gave to the two stages — they are just not very intuitive (read: outright misleading) for somebody who is familiar with other SDOs. This is where most of the confusion comes from, not from the two-stage process itself.
Grüße, Carsten
