Re: Late review management (Re: [Icar] independence of reviews; variability)

[David Meyer <[e-mail]>]

On Wed, Mar 10, 2004 at 07:26:23PM -0500, Margaret Wasserman wrote:
>> At 2:58 PM -0800 3/10/04, David Meyer wrote:
>> > >>  Sometimes, though, we seem to question those assumption,
>> > >> and we end up adding structures, policies or processes that are
>> >>> intended to protect us from the results of our own consensus-driven
>> >>> decisions.  I think that is (and has been) a mistake.
>> >
>> >	I couldn't really parse the last sentence in this
>> >	paragraph, i.e., what do you think is the mistake,
>> >	questioning the assumptions or adding structures,
>> >	policies, or processes?
>> 
>> I meant that I believe it is (and has been) a mistake to add new 
>> structures, policies and procedures at the end of the process that 
>> are intended to protect us from the less-than-perfect results of the 
>> WG consensus process, because I don't believe that is the most 
>> effective way to produce good work.
>> 
>> It's not that I think we should let small, insular WGs run amok and 
>> then blindly publish everything that they produce.  However, I think 
>> that we would do better to offer community guidance and input during 
>> the development of new work (through review, for instance) than to 
>> try to block that work at the end.
>> 
>> There is really nothing new in what I'm saying...  Most modern 
>> engineering organizations have reached the point where they know that 
>> more bugs can be found and corrected, with less schedule impact, 
>> through early review (such as design and code review) than by quality 
>> checks at the end of the process (such as traditional 
>> post-development QA processes).
>> 
>> IMO, we do need some type of evaluation at the end (as a final check 
>> to detect and correct failures earlier in the process), but the 
>> current situation where a large percentage of the work of the IETF is 
>> blocked and modified after the WG believes it is finished, is 
>> pathological.  IMO, we can't fix this problem at the end (by just 
>> publishing the work, for example) we need to fix it at the beginning 
>> (in the WGs). An important way to try to fix this problem is to get 
>> more community input/review, including review and guidance from IETF 
>> leaderhship, earlier in the process.

	I think we're in pretty close agreement.

	Dave

_______________________________________________
Icar mailing list
[e-mail]
https://www1.ietf.org/mailman/listinfo/icar