On Mon, 11 Sep 89 11:01:22 EDT Michael R. Gettes said:
>On Sun, 10 Sep 89 21:20:01 GMT Eric Thomas said:
>>Michael, the present release of LISTSERV has no future, for technical
>>reasons which I have probably already explained. The new functionality I
>>have introduced has been, most of the time, bug fixes and things that
>>were really needed to help the network stay alive, like the BSMTP header
>>stuff. Yes, from time to time I have made enhancements that were neither
>>fixes nor absolutely needed functions, but these were either minor things
>>(like change "Ack=" to default to "No") or things which were a good
>>investment because they actually save me time, like the
>>single-nastygram-error-report function.
>>
>> Eric
>
>Ok, here are my concerns...
>
>Will a new listserv be developed? Will it be compatible with EARN? Will it
>be compatible with older listservs? Will it be documented? Will it be
>coordinated in its development so that certain needed functions can be
>addressed? Will this new listserv be free for use? Will paying people to
>develop a new listserv without having to have listserv sites pay for the
>product be sufficient? Who is going to develop the new listserv? When
>will this new listserv development commence?
>
>I do not expect answers to any of these questions. I have to go on the
>assumption that what we have now is all we have to go on. That is why
>I bring up the subject of documentation so that people who have time
>to work on the program might put in some of the other needed functionality
>and keep it in step with EARN (and EARN in step with non-EARN). I would
>also like to think that BITNET is still largely a VOLUNTEER effort and
>that financial solutions to problems should not always be thought of
>as solutions -- only last possible solutions. Am I making myself clear?
>If not, please tell me... It is true that there is some work that really
>must be paid for when it goes above and beyond regular volunteer efforts.
>It is my personal belief that the current listserv has not been funded
>and should not be funded at this time. If a new one were to be developed,
>one that is faster with even more functionality (not just the needed stuff),
>then I could seeing paying (not a huge amount) for its development as long
>as it met the needs of the entire network (that means all cooperating factions
>of the network). This is, of course, my opinion.
>
>/mrg
I decided to push this discussion off to the Policy-L list in addition to
LSTSRV-L, because we're really heading off into the realm of network policy, I
believe.
As I've taken on more managerial responsibilities, I've begun to take a
different point of view about BITNET than Michael's, even though a few
years ago I'd have agreed with him more.
First, when I say BITNET in the following, I'm really talking about the
combined network minus EARN. That's easier than listing the cooperating
country networks every time.
Second, every time I read Michael and others talk about volunteers, I begin
to cringe a little. 99% of those volunteers are actually paid
professionals of institutions that allow them to work on things that
benefit the entire network, for reasons that run everwhere from keeping
the staff member happy to realizing the benefits that the institution
gets from a better functioning network, and usually some of the whole
spectrum. Yes, some of those people do a lot of their "volunteer" work
during off hours, at home, etc. (I should know, I'm one of them :-), but
we're not dealing with volunteers in the traditional sense of someone
working at one place and then doing unpaid (or low-paid) work at some
civic or charitable organization.
Now to the meat of my harangue.
I believe that one of the real problems that BITNET faces today, outside
of the sheer question of continued relevance in the world of NSF regional
networks with high speed connectivity, is whether the anarchistic world that
it grew up in can continue, with dependance on small numbers of people
around the network providing much of the network's software outside of the
basic transport mechanism's provided by the computer vendors. This is not
to be taken as a flame against anyone. The fact is that to some extent
BITNET is at the mercy of people not under its control to any large measure
at all.
Example number 1. Myself. (Always good to go first, right?)
Suppose I (or the university which employs me) decided that we could not
afford to spend so much time working on the Mail code I distribute without
some recompense, (for the moment we'll leave aside the issue about whether
I or Rice would get any of this recompense.), and sold the code to company
which proceeded to take it off the market in order to protect their own
commercial product. (No, that's not a threat.) (Yes, this has happened
before.) Where would the people inside BITNET who depended on my code
be?
Example number 2. The Columbia Mailer
Remember when we all called this the Crosswell Mailer and Alan Crosswell
was actively working on it? (Again, no flames against Alan; he had other
things to do with his life, how much demanded by Columbia, I don't know.)
The mailer stagnated with well-known problems until John Wagner and Ross
Patterson stepped in and took the development over. And now Ross is
leaving BITNET, perhaps for good, who knows? Suppose Columbia had refused
the rights to the code after Alan quit working on it, and prohibited
anyone else from doing the development necessary to make it a reasonable
mailer for today's environment?
Example number 3. LISTSERV
We've all see then turmoil over the past year (?) as Eric and EARN had
their disagreements over ownership, control, (and just about every thing
under the sun, as far as I can see :-). So what we have now is one version
for BITNET, one version for EARN, some of the EARN people using the
BITNET version, the EARN board still having difficulties deciding how to
deal with the situation, and the whole question of where LISTSERV goes
from here up in the air. (Again, no flames here, against anyone. I'm
not privy to the EARN discussions in full so feel quite unwilling to make
any statement on the rightness of anyone's position.) Are we going to have
two LISTSERVs developing in their different ways until they can't talk
to each other? And what happens when Eric, who now is committed 350%
of the time to other things can't deal with the BITNET version any more.
What happens if Eric should (God forbid) get hit with a truck tomorrow? Who
would own the software and would be either willing or capable of continuing
to support of develop it? Is BITNET facing the same kind of ugly
political struggle that EARN is having in a few years, once more over
LISTSERV?
Example the last: The NSFNET.
Does anyone seriously think that the NSFNET could have been put together
by a group of "volunteers", and even if it had, could continue to function
without some network management responsibility by a central organization?
I don't have the answers to these rhetorical questions. Nor are these
the only examples I could come up with; they're just the ones that immediately
came to mind.
However, I'm concerned that we not complacently ignore the problems of
the dependency BITNET currently has on people not under the control
of the network itself. (One more example: suppose Chris Thomas had
followed through on his insistence of getting out of the routing table
generation business. I know Ed Zawacki is helping with this load, but
is BITNET really ready to handle the time when either or both of these guys
*really* has to stop? We talked about this in Los Angeles, and I don't
think we have a workable structure *still* if this happens.) We've
always worked this way, and it's worked pretty well in the past; I just don't
think BITNET can survive a changing future without more as it were
network ownership of the services being provided.
Right now there is no way for BITNET to fund the development of anything --
there's nothing anywhere setup for even trying to get grants from anyone
to help fund such development. We've got to come up with some sort of
way to do this.
Right now there is no way for BITNET to direct the development of software
that is developed for BITNET members to use. There's no one at BITNIC
(nor will there be after CREN is established, as far as I can tell) who
can direct any kind of design process for network software based upon
the needs of the network as a whole. (If I told you that I knew what
Mail functions CMS needed for BITNET as a whole, I'd probably be wrong.
What I know is what a few intelligent, outspoken, but limited people tell
me that they think, plus my own ideas.) When BITNET was 75 to 100 sites,
maybe that was good enough. Now? I don't believe it. We've got to come up
with some sort of officially recognized technical directions group with
support by the BITNET trustees, board, or whatever they'll be called after
CREN is formed, to give BITNET technical guidance. How many of you have
given up attending the BITNET technical meetings because of lack of
attention being given to their recommendations by the board; how many times
has the board said, well, we'd like to accept your recommendations, but the
technical meeting people aren't really capable of speaking for the network
as a whole, since they have no mandate from the membership for those
decisions? Without some officially recognized, sanctioned, and supported
technical steering committee, technical solutions to BITNET problems will
drag on forever. (Any one willing to guess whether the proposed
restructuring of the BITEARN NODES data will be officially adopted this
century?)
What official status does anyone from BITNET have with EARN, NETNORTH,
or any of the other cooperating networks? If EARN suddenly (well, alright,
not suddenly, knowing EARN it would take too long to call it sudden)
really gets its act together regarding OSI and no longer is an NJE protocol
network, will BITNET still be able to work with it? Is there anyone
at EARN (or BITNIC) whose *responsibility* it is to make sure that the
networks can interoperate?
Remember, we haven't ever had the ability to get BITNET to declare that
RFC822 is how you're supposed to send mail to other BITNET sites, so that
seemingly every year some poor soul gets told by his or her management that
they don't need to use anything but the IBM NOTE command or the VMS whatever
command because BITNET says you can use whatever the vendor provides you and
therefore they don't need to acquire anything else. All we've been able to
do is insist on 9600bps leased lines, and then argue whether T1 lines
would be acceptable or whether you really have to have exactly 9600 bps. :-(
So consider this as a proposal:
1. BITNET should appoint (in some fashion) a technical steering committee
from its member sites. This steering committee would have the authorization
to designate standard software functions (for the different environments
which make up BITNET), which would become official BITNET standards.
2. BITNET should investigate some method of acquiring funding for the
development of software for the benefit of the network as a whole. In
some cases, this might be already existing "volunteer" software, which
BITNET would attempt to guide the development of to make sure that the
software would be directed at meeting the needs of the consensus of the
network. In other cases, such as those environments which have
traditionally been unable to find "volunteers" to write such software,
BITNET would develop and own this software, or at least encourage
its development by providing grants for such effort. In any case, BITNET's
position should be that development of network tools should be guided by
some technical group charged with looking after the network as a whole.
3. An officially recognized group should be established (perhaps with
some funding) for the administration of the network itself. The current
BITNIC functions such as routing table maintenance would be handled by
this group.
The funding for these activities has to come from somewhere, and that means
either by raising the funds from the membership, searching for outside funds
such as the IBM grant which helped BITNET grow at the beginning, or both.
Look, these ideas aren't all that new, or that radical (I think). I don't
even claim that they're that well thought out or that their panaceas. Look
at the problems EARN has had with its officially recognized EARN Technical
Committee (or whatever it's call; I know it as EARNTECH). And one of the
problems with BITNET today is that there's no way for me or any group of
us to officially ask the board on behalf of the membership to do anything
like this, even if it *is* a good idea.
However, I had to get the stuff on the floor again, one more time. Perhaps
eventually someone will listen (or tell me to shut up).
These suggestions don't mean that "volunteer" work is no longer needed.
Probably many of the good ideas will come from people like Eric or newcomers
we've never even heard of yet. My insistence is that when those ideas
come, BITNET had better be capable of guiding them for the benefit of us all.
Richard
|