🚀 go-pugleaf

Christian Marcotte writes
> In article <1992Apr8.021502.19201@sifon.cc.mcgill.ca>  
peterd@cc.mcgill.ca (Peter Deutsch) writes:
[* we plan to charge a license fee for archie follow-ons *]
> 
> 	Peter !!!
> 
> 	As you have written yourself sites are volunteering some machine
> 	resources to run archie clients.  This already shows great 
> 	generosity from those sites.  Now please let me know if I'm wrong
> 	but those sites pretty much loose usage to a machine when it is 
> 	an archie server (archie server keeps the machine very busy).  They
> 	don't get much pay back from offering such service.  
> 
> 	HOW DO YOU EXPECT PEOPLE TO PAY TO OFFER VOLUNTEER SERVICE ?????

First off, a number of sites are paying to offer volunteer services
now. All that I'm proposing is that they pay some of it to us!

Seriously, this is a question I've been trying to answer since this 
whole thing started. Certainly the current version of archie came into 
being because we were willing to develop it and make it available
at no cost to its operators. In a certain sense, we had to do this to 
convince people of the validity of the idea. A lot of great 
volunteers helped us bring it up and run it for the past year, for a 
variety of reasons, and I think we have now convinced a lot of people 
that the Internet is a lot more than bandwidth alone and that it is
actually feasible to gather, track and serve large collections of
information in this manner. We still haven't convinced everyone
that such services should be funded and we still haven't figured
out ourselves how they can be paid for. I believe that we are now at 
the point in the development of the Internet where this question has to 
be addressed.

The problem is that our "drug pusher" approach to getting
people interested in such projects doesn't scale, and we are rapidly 
wearing out our welcome in our "real" jobs as archie takes more and 
more of our time. We need to find a firm and on-going source of funding 
if we are to keep project development alive.	 This is not an isolated
case here. This is a problem that is being faced (or will be faced)
by any potential new Internet service developer.

One way to support such Internet services would be for the regional and 
midlevel service providers to fund development and deployment as part 
of their response to their users' needs and as part of their normal
services. The idea would be that since they are the people actually 
charging for net access, they would have to be the people to pay for 
service development and deployment.

The obvious problem, as we see more and more commercial service
providers coming on-line, is "why would one such provider fund a 
service for another such provider's users?" If we ask them to fund 
services, then we all have to accept the possibility that some
of them will be limiting access to those services to 
their own clients. It would be nice if we could avoid this, if 
possible. Still limiting acess to those who pay may yet be one way we 
do things in the end.

In the meantime, we've had a fair degree of success convincing people
to run the current archie as a free service. This is not to say that it
has been without cost. I understand SURAnet alone has spent some $30k 
setting up their machine, and ANS has allocated a good portion
of an RS-6000 to archie and allowed one of their programmers to handle
porting our code to their machine. These represent real costs to 
these groups to offer archie to the net.

Now, the question is whether we can convince them that
it is worthwhile supporting us so that they can have follow-on
versions and additional services from us. That is the heart of 
the question at this point.

Most of those "commercial" networks currently operating archie that  
we've asked for a license fee have agreed to pay for the next
release and I have some hope that others will license the software and 
offer additional services as a useful way of publicizing their 
activities. Whether they then restrict access is going to be up to  
them, although I obviously hope they choose not to.We are deliberately 
trying to keep our costs as low as possible so as to encourage people 
to pay us while wearing the cost to the users. In the long run I think 
this will lead to greater acceptance of our ideas and perhaps greater 
sales of our software as we offer additional services.

Another other alternative would be for groups 
to fund such useful services as a community service. This still 
translates to us charging a license fee for the software, since we 
still have to pay the bills up here. It's just that we are making a
different argument as to why people should want to do this.

From what I've heard so far I have every expectation that some 
universities and other such sites will also license the software and do 
this, but this is a hard sell to such sites in a time of shrinking 
resources. 


So why not just ask McGill to pay? Well, we can't really ask McGill to 
pay two salaries, buy us a $30k machine, allow us to take half of their 
bandwidth and also pay for such things as travel, etc just because we 
have a service that everyone agrees is useful. The Canadian and Quebec 
governments have not funded McGill to run services for the Internet, 
and McGill wont pay this out of pocket when they face a huge 
accumulated budget deficit. They tell me that they are just are not in 
the Internet services business.

I suppose I can't really blame them, either. Given the potentially huge 
scope for Internet services I believe that we really need to get away
from ad hoc volunteer funding of services and find practical
models for paying for new services. If we do that, we can expect
additional new service providers (and lots of nifty new services)
to be created. If we don't, you will still see interesting projects,
but you just wont realize the potential of the Internet as a
provider of services.

Right now, you have to be pretty crazy to try to deploy something 
on the Internet, since there is no money in the loop to pay for it
and this usually involves either hiding its true scope from the
powers-that-be or running things at a loss until it takes off. I 
would hope that we can find a way to start paying for things, 
because I think we will be better off for it.

Now this may mean users subscribe to useful services, but you do
that now with magazines. It may mean pay per service, but you
do that now when you go to the movies. The idea is that we accept
that such services don't "appear" out of thin air and we accept
that, somewhere along the way, those who consume resources accept that
they will be paying for them.

For what it is worth, a conservative estimate of the cost of
developing the next version of archie sets it at about $75k in 
forgone salaries, travel expenses, network resources, etc (this is NOT 
the cost of operating the 13 servers, just the cost to us of writing 
the code and making it available. Some of this has been paid by McGill 
in the form of time off for travel, access to the net, etc. but most 
has been donated time in designing, coding, support answering user 
questions, writing documentation, etc.)

Providing support services alone (answering mail, responding  to 
postings, publicizing the service) plus designing and coding the 
follow-on versions is really a full-time job now for several people.
We're just getting too big to continue offering further
development as a volunteer effort.

Now, I think that's a pretty good bang for the buck, but it's too high 
a price for Alan and I to continue offering on our own and McGill
has called a halt to our devoting so much time to the project unless
they can expect some return on any investment.

Because we believe in what we are doing, and because we wanted to
continue doing it, we have elected to form a company (Bunyip 
Information Services) and are now negotiating with McGill their  
potential involvement. At the same time, we are actively seeking
out other sources of funding, such as development contracts, etc.

If we can reach agreement, we could potentially remain associated
with McGill in some way. Alternatively, we might have to
move on, in which case the issue of charging takes on a whole 
significance. Whatever happens, we have to cover our costs if
we are to continue working on this.

Obviously, I would hope that our new versions have enough enhancements
that we attract people to operate additional archie servers, tracking 
additional kinds of information. If so, then the new company that owns 
the righst to archie will survive and additional services will be 
offered to the net. If we are wrong, then there will be no follow-ons 
and we'll be looking for new jobs. Wish us luck....


> 	My fear is that Archie-3 will NOT become a widely available  
service.

Well, that is a risk we take, because the issue really is whether
we can afford to charge for archie, but rather whether the
Internet can afford this service. If people are not willing to pay, 
then there will simply not be any ongoing support and future
releases simply wont get written and deployed.

I can't blame McGill for not being willing to fund us to
create and deploy services for the entire Internet. We 
should be grateful for what they have given in support so far
and ask ourselves how we expect them to get paid for in the
future.

BTW, McGill did NOT fund the previous versions of archie, which were in 
fact the result of student projects and volunteers working outside 
their normal duties. McGill has allowed us to use a machine borrowed 
from the School of Computer Science, as we have used machines at the 
other 12 volunteer sites, to provide the service after we left the
School but they did _not_ support any of the 
initial coding other than to provide resources available to every 
graduate student working on programming projects. Alan actually 
submitted his work on the code (which was the great majority of the 
current release) as part of his Masters' degree coding project.


> 	What *I* don't understand about your funding problem (and I sure
> 	you have problems with this too :->) is that McGill will offer 
> 	burseries and provide financial help to students and researchers
> 	to study, research annnnddd write papers cuz papers are GOOD 
> 	for Universities' image and visibility.  NOW, RIS *is* research,
> 	it *is* highly visible (so is archie) and you guys *can* (and
> 	have if I recall) write papers on this work.  Now why don't you
> 	have funding too ????  (And we ALL KNOW Mcgill is not the poorest
> 	university [I should know, I was there]).

Well, to be fair neither Alan nor I are academic staff and 
the McGill Computing Centre is not empowered to develop and deploy
new Internet services for several million users around the world.
(I have advocated that they should be, but that is not a decision that
either myself or the Director of the Computing Centre are in a 
position to make! :-(

After a great deal of agonizing over how such development really could 
and should be funded, I'm afraid I see no other mechanism than the open 
market as a model for what really will work as a means of funding such 
work. That is why we have gone with a company to license and 
distribute our work.

I am also pursuing the possibility of securing research funding from 
various places, and am also looking for various development contracts  
for the follow-ons and other neat ideas we'd like to try out. We have
every hope and expectation that we will survive, some how or the other.
In many ways it's quite exciting (if nerve-wracking!)


> 	Please shed some light on this matter. 

Well, I obviously can't go into too much detail about the company
at this point. We have several product ideas for which we are seeking
development funding, but I can't discuss unreleased products. If 
nothing else, commercial advertising on the net is still an hot-button
topic for many people! ;-)

I do think that the net as a whole will have to undergo a fundamental 
examination of how it intends to support the creation and operation of 
services.  Certainly, once developed we can expect universities, etc to 
_operate_ such services for their own users, but from my conversations 
with other people going through what we are going through most of the 
services that are out there now came about because of "guerilla action" 
by specific individuals, not an institutional endorsement to do free 
development work for millions of people you'll never meet and there is
still little incentive beyond the love of what we are doing to drive
developers.

If we are to continue to see _new_ services such as ours, then we have 
to confront the possibility that not everyone will get access, unless
they are willing to pay. This may be by subscription, it may be per
use, it may be through subsidies, but IMHO the net is simply getting
too big to continue creating and operating such services entirely 
through volunteer efforts. I'm sorry that this is true, but my 
experience in fighting to keep archie alive has led me to this 
conclusion.


> 	Keep up the good work !!!  We *ALL* need it.

Thanks for the encouragement. I might add that we have been somewhat 
fearful of people's reactions to our decision as it certainly
conflicts with the ideas of freedom of access, and openess that
we experience on the Internet. In our case, initially we would
certainly would have preferred to simply be paid to "play in the 
sandbox", but I just don't see this as being a realistic funding model 
as the net passes one million machines in the next couple of months.
What we are experiencing now will be played out again and again
until we agree, as users, that the net needs such services and we
all agree on how to pay for their creation, development and operation.


> BTW:  What is gopher exactly?  I know archie, prospero and WAIS but
>       not gopher.  I read about this because you crossposted in 
>       news.archive.admin


Briefly, Gopher is a distributed, system for serving
information in an Internet environment. Gopher servers offer menus
of iformation that can be accessed, and this information can itself
reside on other machines, or even be extracted from services such
as archie, anonymous FTP, etc dynamically.

If archie provides a useful indexing model in the Internet world, 
Gopher offers a successful example of a user front to the Internet's 
information. It was developed by a group of people at UMinnesota,
and there are currently some 200 Gopher servers, with the
number climbing weekly.

I urge you to check it out. Try an anonymous ftp to 
boombox.micro.umn.edu and look in the subdirectory "pub/gopher". I 
believe that is where the release lives.

				- peterd