Re: [i2rs] Results of WG Adoption Call:draft-keyupate-i2rs-bgp-usecases

[t.petch <[e-mail]>]

----- Original Message -----
From: "Russ White" <[e-mail]>
To: "'Keyur Patel (keyupate)'" <[e-mail]>; "'Alia Atlas'"
<[e-mail]>; <[e-mail]>
Sent: Friday, August 16, 2013 1:29 PM
>
> > I'm a bit confused at this point about when those decisions were
changed,
> > and how we intend to proceed.
>
> I have one possible suggestion, after looking at the structure of the
drafts
> again. What I think we might want to do is to pull the bgp use cases
from
> draft-white to bolster the bgp use cases draft, moving the bgp use
cases
> draft to an editor/contributor format (since it already has a lot of
co's).
> The remaining use cases can be put into a rib uses cases draft, where
we can
> collect all the various "routing protocol agnostic" use cases. Again,
we
> could move this to editor/contributor --I'll be happy to hold the
editor
> baton for this new draft.
>
> There may be some question about what fits where (are all the bgp uses
> really limited to bgp? Do we want a separate overlay use cases, or
should
> these fall into the protocol related draft they work with?), but I
think
> this is a workable model to get all the use cases organized into a
> reasonable set of drafts that can be added to, etc., until the WG gets
to
> the point of having a 'first set' of use cases on the table. After
this
> initial batch is actually published, then it's going to be better to
> continue with individual drafts for individual use cases, rather than
> bis'ing this set, I think.
>
> Thoughts?

Russ

The recurring thought I have is that we are pressing on with design -
architecture, info-model etc - and have not, AFAIK, agreed requirements,
which the use cases should do, or at least bring out into the open.  So
I would oppose the adoption of any design document until the
requirements have greater clarity.

Your comment about boiling oceans I see rather as boiling separate
lakes, each lake being firmly founded on a different view of use cases,
and I think that that has emerged in the discussion of the info-model.

So, for me, use cases comes first, one document preferred, more than one
ok; the rest should wait.

Tom Petch

>
> Russ