Documents need to conclude

[Dennis Glatting <dennis.glatting at software-munitions.com>]

Last night at the IESG's open mic at the Plenary I shared my concern
on document life cycle. I am writing to clarify my comments and offer
a suggestion I did not make at that time.

Yesterday, in the sprit of the law enforcement question in front of
us, a small group of people including myself held court in isolation
and passed judgment on some WG drafts. We found some drafts guilty of
a lack of entropy and some others a lack of momentum. Ladies and
gentlemen you have heard comments like this one many times before:
work items need a time limit. The IESG's response was the same you
have heard many time before also: we're an organization of volunteers
and it is unfair to make such demands.

I agree with the IESG and that it is a difficult problem. However, we
have two related issues before us. The first is we have to figure out
how to scale the IETF. The second is documents that do not conclude in
reasonable time tend to defocus WGs. More specifically, the longer a
WG persists the more items that are put in front of the WG with the
result of spreading limited energies.

For every case I can put before you of documents that need closure I
can put before you documents that have excelled from prolonged
exposure. The IPsec documents are great examples of documents that
have excelled. Therefore, I offer to you this rule to consider:

	Once something is committed to paper in a WG a timer 
	starts. The document has 24 months (6 IETF sessions) 
	to either be sent to the IESG for advancement or 
	with WG consensus the Chair petitions the AD for a 
	two session extension, which can be extended in the 
	same manner again. Otherwise the document is
	withdrawn.

I believe this rule to add something the IETF sorely needs but is
unfair to impose: a little bit of project management. It's advantage
is very low overhead.

Comments?