Re: FYI: Site Local

[Thomas Narten <narten@us.ibm.com>]

Geoff Huston <gih@telstra.net> writes:

> Are we having a vote?

You know, the IETF likes to say we don't vote, and then gets itself
into a pickle anyone tries to somewhat objectively measure a hum. I'm
not sure this is a feature.

Personally, I've come to the conclusion that counting  hands is
superior to humming because:

- counts are unambiguous; the votes are data, concluding from the data
  what the consenus level is a separate issue and can be debated after
  the fact. Blurring the two tends to get us into even more
  trouble. Especially when folks feel like the concensus call was not
  fair/objective/open and that a particular result is being railroaded
  through by the chairs and/or ADs.

- hums are not objective. They can sound very different depending on
  where in the room one is sitting. It's not uncommon for source of
  hums to be rather localized in the room. This can lead to people
  disagreeing to the strength of hums and what the consensus actually
  is.

- I've often heard the volume of hums (for or against) different even
  though it's not clear that the number of hums in the two cases is
  different.  How does one factor that in?

In the IPv6 case, I asked the chairs to count hands because I thought
it was important that the minutes have hard data rather than the
"there was consensus" or "strong consensus" or whatever vague term we
tend to use. The numbers were something like 100 to 20. Hard to argue
with those numbers.

On mailing lists, how is one to measure a hum in a way that doesn't,
in effect, including tallying votes?

Thomas