Re: [Manycouches] soft meeting issues

[Michael Thomas <[e-mail]>]

On 5/14/20 9:11 AM, Michael Richardson wrote:
> Michael Thomas <[e-mail]> wrote:
>      > I wrote up a blog post about this a couple of weeks ago more in the
>      > vein of teleworking, but may be interesting here too.
>
>      > https://rip-van-webble.blogspot.com/2020/05/the-water-cooler-problem.html
>
> I was reading your water cooler post over the past few days between online
> meetings.   The "clueless" list was intringing.  I wonder if anyone has a
> version of this where everyone gets a pseudonym... which would allow a senior
> Cisco Fellow to ask questions too :-)

It was strikingly egalitarian. Fellows did need to have pseudonyms to 
ask... clueless questions :) It was sort of the point. I think Van 
Jacobson created the list, iirc.

>
> One of the useful things about the post-meeting "hang-around", (especially
> when it's at the end of the day: no rush off to lunch or another session), is
> that it allows anyone who isn't steeped in the art of that WG (a "tourist")
> to walk up to someone who was hogging the mic line to ask a ("stupid")
> clarifying question about some base part of the technology.
> This includes senior people (up to and including ADs) who have been suddendly
> made aware of some controversy in the WG.
>    "Hey, can you just clue me into this debate. Why didn't the WG do X?"
>    (there usually is a good, but not obvious answer, often, it would not
>    be backward compatible, or it would violate security system FOO)
Yes, it hard overstate how much of this is a social problem as much of a 
technical problem to overcome. Understanding those interactions is key 
to any attempt at providing a virtual analog. In your example, it's ok 
to be clueless one on one/few but it is not ok to do so in the context 
of a formal meeting. Getting all of those interactions down along with 
their nuances is what is really needed before going into solution space 
-- not that I'm suggesting this is the right venue for either. Which 
actually brings up a salient point: until these kinds of problems can  
be addressed somehow/somewhere/by something, virtual IETF meetings are 
definitionally going to be quite inferior to f2f meetings. I'd venture a 
guess that f2f meetings are worth about 50% of the meeting value, and 
50% of the soft meeting variety value. We could quibble about the actual 
split, but I hope nobody would disagree that the soft part is pretty 
big. So what on earth are we going to about it?
>
> One of the good hang-around places is Jabber in my opinion.

I haven't been to an IETF meeting in a good long time, so I really have 
no idea of how the Jabber dynamic has evolved. But in normal conference 
calls everybody hangs up and that it that. The thing thing that Jabber 
doesn't recreate (at last to my knowledge) is the dynamic of splitting 
off into smaller groups since it's just one jumble of a disconnected 
conversations and worse socially available for everybody to see when 
that might not be desirable. Whatever the solution might be would need 
to take that dynamic into account too.


>
> One of the major reasons, I think, why integrated Jabber into Meetecho is
> BAD, is because the identities asserted that way are *NOT* long-term, and
> because the chat appears to go away when the meeting ends.
> Back when I first started remote attending, I put the "hallway" and "ROLL"
> jabbers into my pidgin config, and and they remain, and I'm always logged
> into them.

Absolutely. You want more rather than fewer "meeting" spaces to have 
impromptu chats. But there remains a serious problem of judging somebody 
receptiveness to a private chat which is much easier in real life: are 
they standing around gabbing already, or are they looking like they're 
off to their next meeting? How would you even move a group chat into a 
more private chat? How do you know whether it is open to outsiders or is 
truly a private chat?


>
> It seems that getting reliable XMPP servers has become somewhat of a
> problem. It was worse a few years ago when jabber.org just became unuseable,
> but it got better, I think...  I suspect that LetsEncrypt certs for servers
> has helped.  I think that the IETF probably should operate it's own jabber
> server for end-identities.

I assume that text chat is not especially challenging for hardware these 
days, especially for a couple thousand people. A couple of 2 core vm's 
from Linode cost about $40/month. If there are reliability concerns, 
it's probably worth it internalizing it since it's not just a 
nice-to-have, but would be mission critical for virtual meetings. This 
is at least a very solvable problem :)

Mike