On Fri, May 17, 2019 at 10:23 AM Mark Nottingham <mnot@xxxxxxxx> wrote:
> On 17 May 2019, at 10:33 am, Ted Lemon <mellon@xxxxxxxxx> wrote:
> On May 17, 2019, at 5:38 AM, Mark Nottingham <mnot@xxxxxxxx> wrote:
>> .... but it works horribly when there's any controversy or disagreement about scope. For example, the ATOMPUB WG was probably the worst experience I've ever had in a standards body, and in my opinion it's largely because the group never met face-to-face (except for a single pre-chartering meeting). This was sold as enabling small developers to participated, but what it ended up being was a mess, because people had little shared context and no established trust.
>
> What could have been done differently?
>
> For example, were there f2f interims?
There were not. I think they could have had a set of people meet and build trust. While there's a risk of creating an "in group" there, if done well you have a group of folks who have mutual respect and demonstrate that to the rest of the group, creating a culture that's respectful and productive.
I've heard folks in other venues opine that the real purpose of F2F meetings is just this; socialisation, so that when you're arguing with someone on the other side of the planet behind a keyboard, there's some shared context / good will.
The IETF process and meetings are more the result of evolution rather than design. Before the net existed, the only way to collaborate effectively was through in person meetings. It would have been difficult to design SMTP entirely on mailing lists (yes, there were earlier mail protocols but...).
That said, I am pretty sure that the calming effect that in-person meetings have is one of the major reasons we continue them. It is certainly one of the reasons W3C started in-person meetings for WGs.
I can certainly imagine circumstances in which it would make sense for a semi-detached WG to meet only at interims. But mostly where they are in effect a joint working group with another body. IETF chose not to do policy work in PKIX. That is now done in CABForum as a result.
I think we are going to face choices about how we want to work at some point and it is far from clear to me that holding plenary meetings three times a year is the best approach. One of the considerable problems with the current model is that the meetings also fund the secretariat which is better than selling standards but limits scope for innovation. Another problem with the current approach is that it is ideal for people who do a small amount in a large number of groups. It is less than ideal to progress one particular work item swiftly.
