Melinda Shore wrote: > On 6/13/15 12:22 PM, Brian E Carpenter wrote: > > On 14/06/2015 01:19, John C Klensin wrote: > >> ... However, if a WG is > >> started with a "solution" and a group of people behind it, there are > >> some bad effects: > > Yes, and this is certainly a very real situation. I've personally > > experienced it in the past, and am currently experiencing it (without > > belligerence, fortunately). > > I'm actually pretty ambivalent about this one. I'd much rather see things > coming in that are relatively well-baked than see proposals that are just > problem descriptions. > It seems to me to be a more productive use of energy to negotiate > engineering differences than it is go try to figure out whether or not a given > problem statement reflects an actual problem that somebody is really > experiencing, or if there's the ability to come up with a useful solution. > Yes, it can be heated and horrible (and I actually left the IETF for several > years in part because of my experience along these lines in one particular > working group), but I think we're better off figuring out how to deal with > these situations than we are going with the problem statement/ use > case/gap analysis model, which is really beginning to annoy me as > unproductive, slow, and unmoored to much that's useful. If a research team brings in a prototype and 'throws it over the wall' to le the WG refine it, I agree. All too often though a product team brings in a beta, and expects a rubber stamp of their soon to ship product. Those are the times that get heated, because changing the 'requirements' means slowing down the release. I understand that many people hate the requirements/problem-statement/use-case documents, but without those the spec has no goals. Granted people driving a given solution will have a set in their head, but if you really expect consensus, and particularly if you expect the wider IETF to understand the context, it has to be documented and agreed on by the WG. Contention rooted in a difference of opinion about the requirements / use-cases can be dealt with by letting each faction solve their favorite subset. Saying 'there can be only one true way' will force heated discussions. Tony > > Melinda