thoughts on w.g. concensus process

Bob Hinden hinden at IPRG.nokia.com
Mon Mar 24 15:03:05 CET 2003


I have been the chair of several working groups where we were able to reach 
consensus on divisive topics.  I have several observations on the process 
based on this experience.

In order for a working group to reach a consensus on a single solution 
there first has to be a consensus that doing so is desirable.  This might 
not be obvious, but there can be lots of reasons why people won't agree 
that this is important.  Some reasons I have seen include (in no particular 
order) 1) Fear that their solution won't be picked, 2) Not wanting any 
solution in this area (e.g,. raising objections instead of working toward 
solutions), 3) No compelling market need so there is little reason to reach 
closure (e.g., more research focus than product focus), 4) Disagreement 
that a single solution is appropriate (e.g., a single solution may not 
really solve the problem, or people disagree on the cost/benefit ratios of 
different solutions).

Unless the working group chairs can lead the working group to the point 
where there is agreement that a single solution is valuable, it is unlikely 
to happen.  It is easiest to do this when the different solutions are only 
different in details (e.g., how to encapsulate IP on a particular link 
type) and/or there is short term market need.  If the solutions address 
different problems and/or there is no pressing market need it may not be 
possible.  Sometimes there can be a middle ground where the group agrees 
with most of what is in a document, and there is only disagreement on a few 
items.  In this case, a push by the working group chair(s) for the need for 
completion can work to get people to put aside their differences.

I don't think reaching a consensus is directly related to the size of a 
working group, but it can be easier if the group is smaller.  Even a small 
group can deadlock if the conditions mentioned above aren't 
resolved.  Large groups are also known to reach consensus too.

In my view, the most important job of working group chair(s) to lead the 
w.g. toward a consensus.  It's is a process that requires leadership more 
than anything else.

Bob

p.s. I think that choosing from among different solutions is the core work 
of the IETF.  The IETF is not a research group where solutions are 
invented.  It is an engineering group whose focus is to pick from available 
solutions and create an open standard.  The IETF's consensus process is 
well suited to picking from available solutions.  It is not a good approach 
for research.  Sometimes I think we get this confused.









More information about the Problem-statement mailing list