Re: [Autoconf] Summary of re-chartering discussions.

[Jari Arkko <[e-mail]>]

Ryuji,

Let me try again.

I don't feel a need to delay the work on the distributed solution in any 
way, or bind its progress to the progress of the other deliverable that 
I wanted to see. I merely observed that as a part of the work for 
developing a distributed solution you will go through the usual steps, 
which involve documenting why that solution is needed, what the options 
there are in the design space, and so on. I'm fine adding the 
distributed solution to the charter, if those steps are in the charter.

Also, you used the phrase "real AUTOCONF work". Were you referring to 
new protocol development as real work, or saying something about what 
types of solutions are going to be needed and used? For what it is 
worth, it is not at all clear to me that the fully distributed solutions 
are the only interesting ones. Virtually all networks have Internet 
connectivity for various good reasons, and relying on some subset of 
nodes -- perhaps the Internet gateway router(s) -- for the ad hoc 
network address allocation may be just fine in these networks. 
Particularly if it comes with less complexity and less time spent 
waiting for new software. I know what I would do in my network.

Jari