Julian Mehnle wrote:
I can only remember 3 review requests of mine.
It wasn't my intention to prove that it's your "inflation", I
wanted to show that there is no "inflation" at all, 9 or 10,
who cares. I counted your three points including 63 as three.
As for the SPF(non-existent-domain) == PermError thing, I
already sort of admitted that bringing it up was a mistake
Yes. And I admitted that you're in theory right. But unlike
"sign" / "qualifier" this NX-beast would require time travel.
no one really seemed to understand, so I forced the issue.
ACK, I tried this with my "source routes". I'm a bit paranoid
with Bruce's review tool, declaring a MUST in 2821 as archaic
scares me. Plus the other potential traps in this part like
why mention "bangpaths" at all if they are irrelevant for SPF.
Bruce has not yet really started. And he'll find _all_ formal
problems if he's determined to find them. Hell, and I never
checked all referenced RfCs for "errata", Bruce also knows all
"errata" and tends to insist on references in certain cases.
He wants a _definition_ of "domain owner". He wants a proper
RfC source for POSIX timestamps, and above all he's not joking.
If there's a weird way why CSV or SPF might not work in a LAN
with private IPs he will find it and ask why the drafts don't
say so. This flak already hit CSV. Sometimes Ned or Keith
are able to stop him. It can also take two IESG members plus
Dave Crocker plus a public announcement of the IETF Chair to
stop Bruce (the 2234bis issue).
It is a game. The IESG will look how much controversy there
was in the "last call". It will be a lot of stuff, so they'll
check if anything is more than trolls, rants, and lies, and if
that's the case the "last call" fails, and maybe we get a WG.
<crystalball heavyabuse="off" />
What was the 4th one?
Only me adding 3 + NXDOMAIN, because NXDOMAIN had a formal "for
Council review" here, but NXDOMAiN belongs to the 3, my bad.
*I* do know what the difference between "authentication" and
"authorization" is.
It started somewhere (forgotten) with a request to use the
correct term with a definition in the terminology chapter, or
to use another term. There is an acknowledged source (2828)
defining both "auth" terms.
So it's unnecessary for Wayne to create any new definition or
to avoid the "auth"-words, he can just add 2828 (like 1938) to
the references, pick the right "auth", pointer, ready. It's
a minor point, William and I tried to help out, nothing more.
I'm not sure what you are complaining about here.
Not sure if this was a "complaint", more like "if they ignore
this input, and prefer to make up their own terminology, then
pray that it's compatible with whatever RfC 2828 says, or it's
'SPF 0 : Bruce 1' in the first minute of the 'last call' game".
The s/prefix/.../ issue may be of minor importance, but it
certainly wasn't brought up for fun. :-)
Solving it with a vote could be fun. Of course I'd be very
happy if Wayne tosses _my_ "sign" or "qualifier" coin... <eg>
Bye, Frank