>
> Sorry, Paul. It is not at all your fault that you didn't know what
> this amorphous "site" thing is, and I apologize for blindsiding you
> in the meeting. (Though I *was* hoping to blindside you with
> the flaw I found in your proposal, but you deprived my of that
> pleasure by discovering it yourself. :-)
The latter would've been ok. One deserves to be blindsided with ones own
flaws.
I think it wasn't so much the blindsiding that bothered me, but the fact
that, unknown to me at the time, you continued to use the word "site" in its
common English sense rather than in the rigorous sense I was assuming. It
should have been clear to you that I was attempting to speak in a rigorous
sense (like when I suggested that two office networks in two cities
connected by a dedicated link with only a single ISP connection surely would
be considered a single "site" in the networking sense. Obviously at that
point in time we couldn't have been discussing the dictionary meaning of the
word "site".)
Anyway forget about it...apology appreciated and accepted.
Thinking back, I remember you started that thread with the statement that
many people equated a site-local with an IPv4 net-10 address, and that this
was incorrect. Other than the fact that an individual site-local address
has a higher probability of being globally unique because of the EID field,
what else is different? I can't re-create your argument in my head.
By the way, I can certainly see that nailing down the definition of "site"
would be difficult. I think the problem is that we use site in two
different contexts: One for "site-local", and the other for SLA (Site Level
Aggregator). It is not necessarily the case that the boundaries of a
site-local "site" match those of a globally aggregatable SLA "site". In
fact, I suspect that if you decoupled the assignment of SLAs for a
site-local prefix from that of SLAs for a globally aggregatable prefix, you
could in fact achieve smooth renumbering when merging two "sites" (if you
had near-unique site-local addresses).
But I'll save that for version two of the internet draft.
PF
--------------------------------------------------------------------
IETF IPng Working Group Mailing List
IPng Home Page: http://playground.sun.com/ipng
FTP archive: ftp://playground.sun.com/pub/ipng
Direct all administrative requests to [EMAIL PROTECTED]
--------------------------------------------------------------------
