Hi! Mvan,

On Wed, 13 Sep 06, you wrote to me:

 ML>> Policy 4, Section 2.2 - "How to Obtain a node number" does not
[ ...trimmed.. ]
 PQ>> What happens if you use a 1:343/999 -aka- instead?
 PQ>> (Or, insert your nearest NC's Net number... for the
 PQ>> purposes of the exercise.)

 ML> Unless the procedure's to use node 999 ...

That's the intent of P4, surely.

 ML> Nope. BinkleyTerm refuses to operate with an unlisted address.

Apparently from your testing, then, it appears that BT does _not_ comply
with P4 and should not be used.  Mmm.

 ML> From my understanding if it's a valid address then it might get
 ML> routed. But 0:0/0 or -/-1 or even 1:65535/65535 shouldn't, and the
 ML> destination system's mail processor should place those messages to a
 ML> defined BAD_AREA.

It (node/999 outbound) ought to route to the NC.  There's no problem there,
if it's ever delivered (marked as 'direct' flavour).

 ML> The applicant would then contact their NC by phone to let them know
 ML> mail's waiting. Or they could simply wait until the NC inspects their
 ML> BAD_AREA.
 ML> For the purpose of applying for a node address, an applicant should
 ML> only need to demonstrate that their system can send Netmail.

I never had to phone no stinking NC.  The first time I did was just a few
years back, to organise a sysop piss-up.

In your case, and if you can do BinkD, then I suggest you try a BSO-style
mailer that can do both dial-up and IP like: Argus or Radius (forget about
Taurus as it's dead-meat, now).  Being more modern software I'm sure they
(it, in reality) won't have a silly need to modify your local copy of the
nodelist, just get them (it) to execute to 'waiting for caller' mode.

Cheers,
Paul.

--- Radius 4.010/21.01.2005(Final)