Excuse the snide and cavalier attitude. I have a disgusting cold, had to
work 13 hours so far today despite, and I want my Mommy!
No problem, I understand completely.
There would seem to be a coordination problem between BITINC and the
Internet NIC. Seems like something that should be fixed. The NIC alone is
bad enough these days.
You're right on both counts. Unfortunately, I'm just a techno-peon; I have no
influence in these circles. The few recommendations I've made in this area have
gone unanswered.
Not all gateways will work for all domains
Are these purportedly Internet gateways? If so, the gateways should be
fixed or not called Internet gateways, yes? Gatemaybes?
As usual, the stupid mess is more complex than it should be. The file is
constructed as a list of domains and subdomains. Each domain or subdomain has
an accompanying gateway specified for it. Some gateways will only accept mail
for the domain they are set up to serve. Others are more tolerant. There is
no indication of which gateways are tolerant and which are not.
I'm so concerned about the implications of this phychological problem that
I'm very much moving in the direction of increased validation as addresses
are entered, rather than relying on delayed validation.
Given things like the evident lack of coordination between BITINC and the
Internet NIC, and likely other such silliness requiring manual human
intervention, it would seem to an outsider that you will stand a greater
risk of merely compounding the problem.
It is more a question of who is responsible. If local tables are the issue then
local management is responsible for making sure that they are correct.
Bouncing responsibility upwards simply leads to problems with no solutions.
Often as not mail I send to BITNIC is _never_ answered. Ditto for the NIC.
In the particular case under discussion, a host refused to deliver a
seemingly valid piece of mail addressed to a DNSable host. hence, we can
either consider BITNET part of the known universe, or a separate universe
with which one might by chance be able to communicate. I would prefer the
former, but could be convinced otherwise.
There are plenty of Internet hosts that still don't have access to the DNS. I
constantly struggle with sites running PMDF to proper set up DNS access.
Believe it or not, a lot of them have actually been advised not to use DNS by
various so-called experts!
This simply illustrates that the problem is not local to BITNET. BITNET has
a different set of problems. All of the BITNET problems could be cured by
properly deriving domain information from NIC tables in a timely fashion.
This would leave us with the NIC problems only. One set of problems is better
than two.
Various PMDF users have made this desire known to us in no uncertain terms
either. If this means that more configuration information is needed to
make it all work, so be it.
Considering that the current configuration information does not make it
work, should I be optimistic adding more? :-)
Well, given what I see as the loss of effective central management, sites
are placed in the position of having to assume more responsibility themselves.
It is nothing to be optimistic about, that's for sure. But nothing is going
to stop this move towards less central management if the central management
groups don't get with the program a little better.
But more seriously, is there anything I can do to help. I had great fun and
learned a lot while registering the .PE domain. I would happily learn more
trying to convince BITNET that Peru exists.
I had the exact same experience when I help get .EG (Egypt) registered.
Fun but frustrating.
Anyway, some suggestions:
(1) The first thing to do is get BITNIC to update the tables. As far as I know
the tables are still maintained by Hank Nussbacher,
HANK(_at_)VM(_dot_)TAU(_dot_)AC(_dot_)IL(_dot_) You
probably should start off by sending a message to him asking why .PE isn't
in this month's DOMAIN NAMES file. If he no longer maintains them he
probably knows who does.
I don't know the policy on these matters -- there may be some requirement
that somebody act as some kind of liason to BITNET to handle conflicts.
The tables I have are dated Feburary 1992 (no day is given). They usually
come out at the beginning of the month. You might try to argue for a
middle-of-the-month update or something.
(2) An action item for me (at least). I've decided to build the entire list of
3166 top-level domains into PMDF. This will not work with full generality
because of the liklihood of domain-specific gateways, but it is the best I
can do from the PMDF end of things.
Ned