At 6:39 PM +0000 3/4/02, Sean Doran wrote:
>| Nevertheless, I lean definitely to one document or document set from
>| the IRTF to IETF, whether or not that's the way it's been done.
>
>It isn't totally clear to me that there *is* a way to reconcile
>the two reqs documents, however, if you're volunteering to lead
>the effort of editing together a single document in a relatively
>short period of time, consider yourself encouraged.
Maybe there's another way to do this, and I'm willing to try. I would
be willing to do a "flow edit", not changing any substance, on the
Group A document. The only thing I'd try to reconcile is subject
headings, and note when a topic seems to be in one and not the other.
Meanwhile, if someone involved in Group A has the time, do exactly
the same thing to the Group B document. What I'm describing here is
sort of an intelligent diff. I don't think we can start reconciling
until we have two structurally similar documents.
If those two "normalized" drafts were available, I'd be willing to
take on facilitating the reconciliation, fully recognizing that such
a document is quite likely to identify different requirements and/or
different architectural approaches to the same requirement. I'm
honestly not sure if splitting into requirements and architectecture
sub-documents would make the job easier, but I do find that
requirements and architecture are uncomfortably intermixed at the
moment.
>
>(Relatively short - well, I'd really like to see us issue the
>equivalent of a WG last call towards the end of IETF week, assuming
>the suggested improvements from the mailing list have quieted
>down by then... An intense editing period should not push this
>off by many weeks.)
Don't know whether that time frame is realistic, but more like 2-3 weeks.
>
>| BIG CAVEAT. If there turns out to be several distinct viewpoints
>| about RoutingNG--hypothetically, one based on control theory, one on
>| map exchange, and one on the technology of the Psychic Friends
>| Network--it's entirely appropriate to have several such documents,
>| much as there were several competing approaches for IPng. Even so, it
>| would be useful that these documents be written with something of a
>| common structure so they can be compared and contrasted.
>
>I fully agree. Indeed, I predict that this is exactly what will happen.
>
>That said, let's learn a bunch of "don't"s from the ROAD process...
>
>#1: don't put all your eggs in one basket
>#2: don't market your chickens while they're still eggs
And remember to have the project goosed periodically.
>
>I would really like to see it be possible for EACH of these
>routing architectures to be operated IN PARALLEL in different
>parts of the Internet,
Quite a good idea, especially if there could be a bakeoff using
something like 6bone.
>but then I get accused of smoking strange
>things sometimes, so take that with a grain of salt.
Sean, some friendly advice. Coffee, it is true, sometimes can be
improved with a grain of salt. I cannot think of anything smokable
that would be improved with salt -- smoking just doesn't get hot
enough to vaporize the NaCl. :-)
>
>| Frank, your personal opinion above isn't necessarily unreasonable. My
>| gut tells me that by the time something hits the IETF, there should
>| be one consensus document.
>
>Ultimately, I personally would like to hand the IETF something against which
>they just engineer -- generating MIBs and tweaking message formats --
>as if the RRG were an author putting forward a standards track document
>or set thereof. There is quite alot of work to do before then.
>
>I would love people's comments on this, particularly those of
>of you who e.g. chair IDR or manage the Routing Area.
>
> Sean.
This archive was generated by hypermail 2b29 : Mon Aug 04 2003 - 04:10:03 EDT