Hi, On 4 Jul 2019, at 11:31, Eric Rescorla <ekr@xxxxxxxx> wrote: > Ignoring labelling for a moment, in a number of WGs (HTTP, TLS, and > QUIC) we have found it necessary to have full implementations and > large-scale deployments quite early in the design process, long before > anyone thinks that the document is done. I wondered whether there were two scenarios that were worth considering, separately: 1. Instances where design and implementation are both happening in communities that are engaged in the IETF. The examples above seem like handy examples. The people who are likely to implement are involved with or aware of the design discussions; the precise disposition of individual draft specifications is likely fairly easy to socialise within that group of people. [In this case it's not entirely obvious to me that there is a problem to solve, but perhaps I am over-stating the degree of overlap in people involved in implementation and design (the WGs you mentioned are not ones that I have ever really followed).] 2. Instances where design and implementation are separate. The easy example that springs to my mind is the implementation of EPP before the original publication of RFC 3730 by various domain registries who more engaged in other communities than in regext. [In the regext/EPP example I think it's clear that there were annoying interop problems resulting from different registries implementing different revisions of an internet-draft rather than the same, stable specification. It's not clear to me that in that case having notices of the form "this has been in the oven long enough that you probably won't get sick" would have helped, though; such notices could have been attached to precisely the same set of drafts that were implemented, and the result would still have been that a single EPP client would work with one registry but not another.] However, having typed all of that I seem to have convinced myself that neither case would benefit tremendously from a new disposition marker for the document, and I wonder again what the original suggestion is trying to achieve. If what is actually required is better coordination between people preparing early implementations, regardless of what communities those people might be engaged in, carrying an applicability statement in a section that is preserved and extended through revisions of an internet draft might be better than a short label. Such a section could be left in-place (or perhaps moved to an Appendix) during final publication in the RFC series as documentation of early implementation experience. It is surely possible today for any wg to decide to require such a section to exist and to be maintained as part of the workflow they adopt in carrying out design work. Joe
Attachment:
signature.asc
Description: Message signed with OpenPGP
