> And my reply was that if you want to talk bandwidth or resources in general,
> then let's talk bandwidth/resources, and the conclusion is that a usenet
> group with only 10k subscribers is a waste of resources and bandwidth due to
> the tremendous amount of hosts which will get information they are not
> interested in, batched or not batched.
 
I don't have the usenet arbitron stats available to me at the moment, but my
impression was that an average newsgroup had roughly 40,000 readers.  So a
10,000 reader initial base is farly respectable, especially given that many
others are likely to subscribe to a newsgroup that might never sign up for a
mailing list.
 
As for the supposed waste of bandwidth involved in hosts getting articles they
are not interested in, that is the point of my apples vs. oranges comment.
Usenet is based on the idea that some people have resources (bandwidth, cpu,
disk space) that they don't mind "wasting".  When I say that usenet is
efficient for large discussion groups, I say that not because it uses less
resources overall, but because it makes better use of limited resources at the
expense of wasting some which are abundant.
 
Hosts which are happy to waste bandwidth will get fairly complete newsfeeds
and won't care about articles that nobody reads.  Hosts on dialup UUCP links
that care about their bandwidth will only carry the newsgroup if there are
users at their site who are interested, and if they do, news will be more
efficient than mail for them.
 
me> The only people who can adequately judge that are the administrators of
me> the INTERBIT MAILER in question.
 
> Which you will recall is exactly what I said - it is for UB to decide.
 
Well, I'm not an expert on BITNET, but I thought that there were only two
public INTERBIT gateways on BITNET (you said there were more on EARN).  In the
case of UB, their MAILER may use UB's Internet connection, but in the case of
a large list on a non-Internet-connected site, the burden would fall on one of
the two public INTERBIT MAILERs.
 
> You got it wrong.  I have no problem with the existence of 200 unix list
> managers, as long as they don't call themselves 'unix listserv version x.y'
> and don't claim in their documentation to be a 'port of the bitnet
> listserver'.
 
As I said before, I don't care about the naming issue.
 
me> ...the INTERBIT mail gateway only has to expand one address, and the
me> peered Unix LISTSERVs can spread the mail load amongst themselves more
me> easily than appears to be the case with BITNET LISTSERV and the INTERBIT
me> gateways.
 
> And you're the one complaining about apple vs orange comparisons? Give me a
> break. Of course a peered unix list is more parallel than a non-peered
> BITNET list, but what makes you think the same thing can't be done just as
> easily by peering the lists between BITNET LISTSERV's?  In addition, this
> would keep the "single logical list" vision and automatically send new
> subscribers off to the best site.
 
I'm aware that peering a list amongst BITNET LISTSERVs is a very good way of
spreading the BITNET subscribers around (in fact, it probably does a better
job of spreading BITNET subscribers around than Un*x L*STSERV does with
Internet subscribers, since the BITNET topology is more static and visible
than the Internet topology).  But you yourself pointed out that BITNET
LISTSERV's idea of the best peer for Internet subscribers was likely to be an
EARN site (not very optimal for lists with mostly U.S. Internet subscribers),
and have proposed various roundrobin schemes to deal with the fact that on
most peered BITNET LISTSERV lists, all the Internet subscribers end up on one
peer.
 
> Then how come people on the Internet aren't asking for the 3500 existing
> BITNET lists to be split in this fashion for their convenience? ... If they
> liked the unix list servers so much, they would start a revolution and move
> all the lists to unix - wouldn't they?
 
Probably because most Internet subscribers don't know very much about
mail-based server technology.  Un*x L*STSERV is relatively recent, and the
interactive features have only become available in the latest release.  I
would guess that most BITNET LISTSERV subscribers don't even know that there
are other mail-based subscription mechanisms out there.  And I'm not proposing
that all the existing LISTSERV lists be moved to Unix, only that some lists
may want to provide gatewayed Un*x L*STSERV lists as an option for their
subscribers.
 
me> I believe that both implementations of LISTSERV support mail
                   ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> This is exactly the kind of statement  that I object to. The software you
> are using is  NOT an implementation of LISTSERV. It  isn't compatible, it
> cannot connect to LISTSERV as a peer, it just does about the same sort of
> things. Internet people normally say "implementation of X" only when they
> have a  piece of software which  complies to some RFC  or other standard;
> however they seem to make an exception for LISTSERV.
 
Even if there were an RFC, it would probably specify only the end-user mail
interface. and not the peer interface.  And given the different natures of the
BITNET and Internet architectures, I think it is pointless to try and find one
peering interface which works for both of them.  Since most users don't ever
see the peering interface, I don't see this as a problem.  It might be nice to
have an RFC for the Internet peering mechanism, though.
 
> I have trouble thinking of Alex as an end user rather than a CS-type
> militant.
 
I like this!  Thanks for the compliment!  I do favor strongly typed languages.
 
> But you really don't need to worry about me: since unix is unanimously
> considered to be the system of the future and my code only works on
> dinosaurs, you can safely ignore it and develop your own protocols. Which I
> am sure you will do no matter what my advice might be.
 
Which is what is happening, especially with other Unix mail-based subscription
managers.  I find it ironic that Tasos and I are the ones arguing for as much
subscriber-interface compatibility as possible between the BITNET and Internet
worlds, while you and most of the rest of the Internet list-managers seem
content to ignore each other's work, rather than trying to create a more
powerful synthesis of the two (and no, this doesn't mean I want to force
everybody's mail servers to be running the same implementation).
 
@alex