Erik,
(Thanks for explicitly inviting my clarification about standard
procedures...)
A spec which goes to entry-level standards status (i.e., Proposed
Standard) must be a stable spec, with no *known* omissions or errors.
It is not required to have any implementation or operations experience,
unless the spec has some flavor of "danger" associated with it (such
as changing the Internet's routing protocol) or is viewed as having
inherent technical obscurity which requires experimentation before there
can be adequate understanding of the technology's "physics".
In the case of Keld's document, I note that there have been strong
expressions of concern about its complexity (which may well be necessary,
but it impedes thorough understanding) and a lack of full understanding
of its relationship to other standards efforts. Since the bulk of
the work has been done outside of this wg and the spec's author is
comfortable with having the spec published outside of the standards
track, it would seem quite reasonable for the wg to support the
document's publication as Informational.
Dave