RE: draft-eastlake-rfc2777bis-selection-01.txt

[Christian Huitema <huitema@windows.microsoft.com>]

> Donald Eastlake has requested that
> draft-eastlake-rfc2777bis-selection-01.txt be published as
> an Informational RFC obsoleting RFC 2777.  Harald has asked the NomCom
> WG to review this doc before taking the request to the IESG and IETF
> last call.
> 
> Please take some time to review and to send comments to this list.

After reviewing, the draft appears nearly identical to RFC 2777, the
only very significant change being a recommendation to not use stock
prices and related market indices as sources of randomness. This is
based on past experience, that giving a list of stocks can backfire when
companies merge or get delisted, or that planning to use some reported
numbers can become problematic when the listing unit suddenly changes.

The recommendation appears technically sound, but the process is very
much a geeky option compared to the previous "standard", picking names
from a hat. For that reason, I support a publication as an independent
document rather than incorporation in the base nomcom document: future
nomcom chairs should have the liberty to use the computer based system
described in RFC 2777 (bis), or use instead an analog system with
suitable guarantees.

-- Christian Huitema