On Sat, 14 Mar 1992 08:48:58 CST Natalie Maynor <[log in to unmask]>
said:
>This is a good reason NOT to tell the net about such equivalencies. When
>I found out that sometimes mail would be changed to my bitnet address if
>the equivalency were registered, I was very happy that ours isn't. I
>don't want my bitnet address used -- EVER. It can add days to the
>routing.
I'm sorry to have to say this, but your attitude is at best improductive.
Your problem is that your BITNET connection is not performing
satisfactorily. The ideal solution is to make it perform to satisfaction.
Another solution is to declare it devoid of end-user value and remove it
completely. Both of these approaches make sense. Advocating the secrecy
of BITNET<->Internet address mapping just because your installation is
not willing to take either action and it happens, through a side effect
of a common piece of software, to cause the same behaviour as removal
from BITNET (but without requiring you to justify anything) is not a
productive attitude. It is what I call hiding one's head in the sand. It
may solve your own, personal problem, but it does not help users without
BITNET access who want to know the hostname of your machine, or users
without Internet access who only got an Internet hostname. It does not
help people who run the network in their troubleshooting.
>Since most lists have internet addresses also, I strongly recommend that
>subscribers use their internet addresses and forget the list's bitnet
>address.
And when an end-user asks a LISTSERV for a list about topic XYZ, what he
gets is a BITNET node name. If everyone did like you and made sure not to
register their Internet<->BITNET address mapping, the user would have no
way to find out that UGA can be reached as UGA.CC.EDU, unless someone had
told him.
Furthermore, if everyone, in addition to removing the registration for
their Internet<->BITNET mapping, also changed all subscriptions to an
Internet address form, the DISTRIBUTE protocol would be totally disabled.
There would be only Internet subscribers, all routed through the same
gateway at INTERBIT. All of a sudden, this one gateway (which is actually
spread over a handful of nodes) would have to handle hundreds of
thousands of pieces of mail a day. What happens to the gateway? What does
the executive director of the computing centre in question decide the
very next day?
Mailing lists can only exist (on the BITNET scale) through the use of
DISTRIBUTE. To be effective, DISTRIBUTE requires either BITNET addresses,
Internet addresses with a known mapping, or Internet addresses mapped to
a gateway serving a particular area, rather than acting as a gateway from
the whole BITNET world to the whole Internet world. DISTRIBUTE does not
provide optimal performance for each and every posting to each and every
list. It only provides optimal performance for the network as a whole.
Eric
|