On 20 May 2024, at 10:13, Bob Hinden wrote:
On May 19, 2024, at 7:22 PM, Dave Crocker dhc@xxxxxxxxxxxx wrote:
On 5/10/2024 2:33 PM, Dave Crocker wrote:
Companies don't work that way.
That's an overly broad universal statement.
It was meant to be quite broad. Because I am pretty sure that I've never seen a counter-example in 50 years.
Companies do not make public, future commitments for implementing standards. And when there are attempts to get them to, they waffle and evade.
Some do, some don't. We've had experience in the email community where people participate quite openly, show up for the hackathon with example code, and discuss their implementation plans quite openly. We've also had experience where (mostly large) companies do exactly what you describe. The requirement is not that all participants in the WG make a commitment to implement; just two or more.
Except that the issue is not whether companies choose to show up, but whether they choose to make a commitment early in a process to implement the result of that process.
If you know of examples of this being done -- in the IETF context, especially -- please document them.
That you think showing up at a hackathon is equivalent to satisfying this wg's charter requirement for a priori commitment might suggest the nature of the very basic disconnect that prompted this chartering error.
Also, I believe, the IETF has wisely never tried to impose this burden.
I regularly hear the question posed in BOFs. Perhaps that's just lip-service, but it's certainly being taken into consideration. And I think it is a perfectly wise thing to impose, particularly in a space where we've seen multiple proposals over the years that individuals bring to WGs, get large amounts of input and direction, only to discover that nobody is interested in implementing it aside from the implementer.
Interest is fundamentally different from commitment. In the
earliest days of the IETF, it was actually pretty common to
explicit assess community interest. Actively, if subjectively. A
query for show of hands about 'interest' in implementing is what I
recall.
But again, that's profoundly different from inventing a formal requirement for explicit, public commitment.
Again, if the goal is to limit this working group to only take on specifications that are already in use, then just say that. It's simpler, clearer, more direct and, frankly, more pragmatic.
That is not the intention as far as I understand it, and in fact it is something that I would rather see us decrease or eliminate rather than encourage. If people want to work on specifications outside of the IETF, they should publish them outside of the IETF. We have seen the ill effects of bringing in work that is mostly or completely "done". And there is IMO no reasonable way to truly assess the IETF consensus for work that is mostly or completely done outside of the IETF.
My comment is, really, about the effect this requirement will
have. In practical terms it is essentially mandating coming to
the IETF only after the work is fielded.
Because that is the practical effect of what's in the charter.
To further Dave’s points, “implementation” is not the same as deploying it operationally at scale. That would be a significant commitment for someone to make.
Actually, that seems to be opposed to Dave's points rather than furthering them: Deploying at scale would be too significant a commitment to make. Implementation is not, at least for some folks.
1. Implementation effort is not free.
2. Public promise to implement is making a public, corporate commitment.
The first has been required only selectively, over the decades, and with specific justification. There is no precedent for making it a blanket, ongoing requirement.
The second is, to my knowledge, without any precedent in the IETF and is broadly against typical corporation's policies.
d/
-- Dave Crocker Brandenburg InternetWorking bbiw.net mast:@dcrocker@mastodon.social
