bstempleton@watmath.UUCP (Brad Templeton) (07/24/83)
Wouldn't it be nice if the poster paid for the cost of transmitting the news rather than those who get it? This would be like the postal service or private publishing. What we have now is the receivers and forwarders paying the cost. If we had USENET Inc. this could be arranged. Then all the xxx@mit-eddies and Mad Programmers and people on Soap Boxes and warlocks in dungeons and users of overworked keywords and users of net.test and net.general would get a bill (no names, no suit) for what they posted that month. One sure effect would be that people would shut up. Essentially, what I think we want is a system where the poster has to pay but the receiver can accept collect messages. Thus most people will say, "I'll accept collect from the following people", where those people will be digest moderators (magazine editors) and wonderful people like myself ( 8-> ) They might accept certain groups collect and some not. Thus if you want to post, you can pay for it, in which case everybody gets it, or you can send it to a moderator who can moderate it, or you can just send it out and hope that enough people are willing to take it collect. Note that libertarians (like me, in my own fashion) can't complain about this idea because it includes the current system as a subset. If people love the current system so much, they can elect to get everything collect and get the same mess they have now. For myself, I know a fair number of people I would not accept collect.... -- Brad Templeton - Waterloo, Ont. (519) 886-7304
alb@alice.UUCP (07/24/83)
One sure effect would be that people would shut up. That's for sure! That's exactly what would happen: EVERYONE would shut up and the net would die. How many people do you actually think can pay for the uucp bill, cpu charges, and disk space for even one article for EVERY site on the net (and do it on a regular basis?)? It's obvious that the present scheme is not entirely fair, but a scheme in which the posters pays all the costs is not practical.
laura@utcsstat.UUCP (07/24/83)
Brad (& everybody), I have some problems with this scheme, allow me to outline them. What would you do about the people who ABSOLUTELY WILL NOT have net.jobs forwarded to their sites? They don't care who pays for it; they just don't want their staff reading it and leaving their company. I assure you that these folks exist. I gather that there are others who feel the same way about net.philosophy, net.flame and others. they are not worried about costs, they simply dont want them on their machine. What if I decide that I want only certain articles from net.whatever? Either I personally have to keep a huge list of which people I am willing to pay for, which I have to describe in some way (for example, all new submissions, all submissions that are less than 5 lines long and greater than 0 lines long, <not counting headers!>, all submissions from sites that are within 80 miles from here, all submissions from the sites unc and ucmp-cs, everything that mentions 16 and 32 bit computers, nothing that mentions 8 bit computers, nothing from watmath!bstempleton < :-) > anything from vortex!lauren and nothing from ucbvax). Also at what level do we decide to pay for this? If the "system manager" decides, then no net.jobs gets through, despite internal interest. If the individual decides... have you ever got a huge bill that you weren't expecting? there are other logistics problems. When utzoo first joined the net, many moons ago, the department of Zoology was in the middle of a huge crack down on long distance phone calls. 3 hour connections to duke didn't go over very well. In addition, we didnt have an auto-dialer so every call meant that either Henry Spencer or myself had to babysit the phone connection. I still chuckle when I hear the auto-dialer clicking away as it dials out! After one month, almost everything got the axe. FA.unix-wizards, NET.bugs.all (yes those were the days of the NET and FA, not the net and fa) remained. FA.space did as well (by Henry Spencer fiat) but FA.sf-lovers didn't, nor did FA.human-nets. I remember calculating the costs of both of these groups and staggering around the department and asking if anyone was willing to pay one Xth of the cost. The plan folded for several reasons. First of all, a share worked out to ~50 dollars a month, and I wasn't prepared to pay that much. Secondly, the atmosphere changed radically when people thought about paying for it out of their own pocket. Some people thought that Henry was being unfair and wanted to chew him out about FA.space. I cooled them out. other people wanted to pay, but they wanted the files read only by a group that they were in (at utzoo we don't use groups for anything, everyone is in their own group, with the group 'bin' and the group 'uucp' being the notable exceptions) so that nobody could read sf-lovers without paying for it. Some people wanted the lineprinter patrolled so that no one made illegal copies of sf-lovers to be used by non-payers. I received a lot of abuse when I mentioned that I intended to run off a line-printer copy for me to read in class as usual, and that even worse I was going to run off copies to take and post at csrg and at the physics department as had been my normal habit. What I found most frightening was that everybody was willing to appoint me as facist leader and patroller of the lineprinter. I wasn't interested in the job of fascist leader. Presumably there are people out there that can stomach "security checks" and "law enforcement" but I am not one of them. I might have been willing to pay something around $50 for sf-lovers, but I was not willing to be dictator for ANYBODY. I got out of the "money for sf-lovers" business, and did without sf-lovers until decvax offered to be utzoo's news feed (thanks decvax!) which was a shorter hop, (thus a phone saving) and utzoo began to make money for the department. if we go with payments on the site or individual level, how do we amortise the costs? If we go with Mike Lesk's scheme mentioned at USENIX (everybody sends their L.sys information to <say> ihnp4, in exchange for getting all the information back) we have gone to an ARPANET situation where every machine calls every other machine. (I know that I am oversimplifying, and ignoring the "super secure" machines which cannot talks to the world at large). This is fine for people with lots of money from the DoD to make phone calls with but poorer sites cannot do this. However, we would all know that ihnp4 could talk to everyone wouldn't we! I see ihnp4 becoming so swamped with traffic that a vax isn't big enough, especially if Lauren Weinstein starts selling his uucp to everyone with a micro. Say we do not make ihnp4 the central site. Suppose we just leave things as they are (I don't think we can do this forever either, but ignore this). Now what. If I decide to pay for net.sf-lovers from ANYWHERE, and you decide the same thing, do you subsidise my costs, since utzoo will be sending sf-lovers to watmath where you are if we continue as present? If this doesn't happen, and each site pays the full cost of transport to that site the situation for you and I out in the "middle of nowhere" will be even worse than our current phone bill arrangements. In addition, where does all the money go? If we distribute the costs equally over every node on the net, the folks far far away (Australia and England for instance) are in for a price break, unless people in North America decide not to cover the costs because they don't think that Australia is worth it. The site that has never had a long distance call in its life (lucky dog!) is not going to be thrilled at this new prospect, and once again, we have curtailed access to a resource to the rich alone. This might keep the micros out of usenet, which is a good or bad thing depending on your point of view. What I want to know is: HOW EXPENSIVE IS A SATELLITE CHANNEL? At some point, it is going to make sense for decvax to pay for a large chunk of a satellite, since they are already paying a fortune in phone bills. the same can be said for ucbvax, and Bell as a whole (though I gather from talking to Lesk Bell has some strange ideas about what is cheaper and better when it comes to telephone bills). Existing sites could pay for "their share" of the satellite. New sites could pay for their share of the new computers needed to do the crunching to handle their mail. What do you think, gang? laura creighton utzoo!laura utzoo!utcsstat!laura (preferred) ps Will someone see that Mike Lesk gets a copy of this, and Brad's original article? I don't have a mail address for him, and I would hate to be misrepresenting him. laura
bstempleton@watmath.UUCP (Brad Templeton) (07/25/83)
Well, let me expand upon several points. This kind of scheme can only be done in a usenet, inc. scenario like I and Mark Horton have proposed before, both on the net and at Usenix. This is a system where the usenet is run by a profit-making company that maintains a central node with all databases, routing information etc. [There has been some debate as to whether such a company would control the net against the wishes of its customers. I state again that no company goes against the wishes of the people who pay it money, and furthermore this company would work mainly as a store and forward station and would make no judgements on content unless lawyers said it would be forced to by law.] As for deciding who you will accept from? Laura just made a short list in a few minutes. It may not be complete, but the idea is that you add something to the list when it starts to annoy you or you decide you like it. Really not that much work, and not too much disk space when it is on a site by site basis. News distribution would not all be from the central site. Sites would in fact pay a premium to be connected to the central site. Otherwise they might redistribute the news on an "amway" or "pyramid" type basis, either taking a (small) profit on what they pass along, or if they have a non-commercial system, just taking their own news for free in exchange. This only refers to collect news. I maintain that in this area we have a system at least as good as the one now. Any debates about who pays for what in the collect news are no different for what we have now when it is essentially all collect. (ie. most of the cost, if not all, is payed by the forwarding and reading sites) If individuals start going crazy with postings like this month's 53 article winner, does anybody not think they should pay for this? Usenet inc. is just a germ of an idea right now. It's main goal would be to equalize and lower the cost of usenet to everybody. The main thing blocking its formation is the fact that many netters are fooled into thinking the net is currently free because rich companies handle their costs (or the cost gets absorbed into their big corporate phone bill) and they don't see why they should start paying for something that is free. (I don't know how much the net costs, my survey failed because nobody has the time to work it out. It's not cheap, though. Ask Armando or Bill Shannon!) -- Brad Templeton - Waterloo, Ont. (519) 886-7304
mjb@ukc.UUCP (M.J.Bayliss) (07/25/83)
For what it's worth, here is what we plan to do in the UK over
news costs:
i) the main trans-atlantic and european feeds will be via ukc
(with edcaad as a backup if ukc falls apart), thus the only
site getting the large costs will be ukc, and these costs
will be split equally among all sites on the network in the UK.
(problem, this is still a high cost for each site, we need to expand the
network here to four times its size to get a reasonable charge
for each site.)
ii) any other link is negotation between the sites involved,
with the expectation that charges will be paid by sites
receiving the news. (currently traffic flow is mainly
one way, when this changes, we will probably have to rethink
all our costing plans.)
iii) sites tend to cover their capital costs by charging connection
fees for each site that links into them, and this includes
making sure the new site has upto date s/w, and knows what they
are doing. Also if a site is handling a large load, there is
nothing stopping them making a monthly service charge to adjacent
sites which is how some sites cover their running costs.
Mike Bayliss University of Kent, UK
...!{vax135,mcvax}!ukc!mjblarry@grkermit.UUCP (Larry Kolodney) (07/25/83)
The whole beauty of the USENET is its inherent democracy. Any person is able to communicate to any other person or group of people on the net without regard to his/her financial status, agressiveness, or whatever other circumstance might ordinarily prevent him/her from getting an idea accross. To start charging people on a per item basis would turn USENET into nothing but a fancy telephone system. Most people are not going to pay to get mail from strangers, especially ones who disagree with them. And who can really afford to pay to send out personal opinion items. In a pay per item system, the personal cost becomes too great compared to the personal benefit. In the current system, the presumably the overall cost and benfit are roughly equal. This is sort of like the paradox of the farmers and the field. A number of farmers share a field, on which they can graze X sheep. For every sheep greater than X that grazes on the field, the field becomes slightly less productive the next year. Unfortunately, the marginal cost of a given farmer adding 1 more sheep, namely a slight deterioration of the field, is much less than the marginal gain, an extra sheep to sell at market. So, if each farmer acts in his self interest, he will continue to add sheep to the field until the cost of adding one sheep balances the benefit of an extra sheep. By that time, the field will be in ruins. The same is true of a pay per message USENET, only inverse. Since the marginal cost of one message is greater than the marginal gain, people will put less and less messages on the net until the costs balance out. By that time the net will be a wasteland where only vitally important messages are sent. In both of the above cases, the solution involves a communal setup. The farmers who must share the field must also share the sheep, thus no farmer makes any marginal gain by adding sheep. Like wise, if the cost of USENET is shared equally by all, no person takes a marginal loss by sending messages. <retroactive FLAME ON> <FLAME OFF> -- Larry Kolodney #8 (Moving up) (USENET) decvax!genrad!grkermit!larry allegra!linus!genrad!grkermit!larry (ARPA) rms.g.lkk@mit-ai
spaf@gatech.UUCP (07/26/83)
Time for my $.02 of soap...
I believe that if you set up some scheme of charging for news,
we would see the development of new news networks. There is
no reason to believe that Usenet would survive undiminished --
if at all. The first thing we would probably see would be
lots of cities developing their own isolated "islands" of
newsgroups, sort of like the regional newsgroups in use now.
Some of these "islands" might connect together by links from
machines whose operators/owners have access to 800 numbers or
the like. I mean, why pay for news when it is possible to
get any number of machines together for (nearly) free?
What would happen next is that smaller or isolated machines
would start dialing into machines in these "island" networks.
I'm not sure how the cost would be distributed, but I suspect that
it would be done close to the same as now. I suspect that we'd
see the net reduced to about 60% or so of current sites, and
these would be divided up into 4 to 20 separate or loosely
connected networks.
In parallel with this development we would see these isolated machines
and networks sending mail to each other. Does your proposal
include charging for mail transfer? If it doesn't, then
I'll just post all of my local net news articles to your
network "island" by mail and let you resubmit them there.
Since this will be quite a chore for me (and for you) we'll
write software to automate this process, and soon we'll have
another version of news software running -- which will require
more machine resources than the current system.
Suppose you charge for individual mail messages. How long do
you think the system would last? I suspect that any workable
system you would implement would cost so much extra in overhead
that it would be cheaper for each individual site to establish
their own connections. Thus, we'd end up with each site on
the network having huge L.sys files. Your new network would
be bypassed whenever possible and as a result would
probably not make enough to justify its continued existence.
CSNet is using a scheme similar to what has been proposed here.
You have to pay dues to join, all of the messages are routed
through 2 relay points which connect to other relays, and
message traffic is metered and charged for. However, $30K
per commercial site is a bit steep, don't you think? And for
me to send mail to Emory University, just a few miles away,
I have to incur the charges for a call to Delaware, and Emory
gets charged for another call from Delaware. Using uucp,
it is a local call. Needless to say, most CSNet sites will use
uucp if possible.
In conclusion, I think that any attempt to charge on a per-message
basis will only result in fracturing the network. Some might
argue that that could be desirable, but I won't. When you
buy a newspaper you buy the whole thing. You may not read the
sports section, you may not care about the stock prices, you
could probably not care less about the horoscopes -- but if
you want to read any of the news, you buy (and subsidize) the
whole thing. If we paid our newspaper reporters based on
how well we liked their style, or if reporters had to pay
to get their stories published, how many newspapers or magazines
do you think we'd have? If it was economically feasible, don't
you think someone would have tried it already?
--
The soapbox of Gene Spafford
CSNet: Spaf @ GATech ARPA: Spaf.GATech @ UDel-Relay
uucp: ...!{sb1,allegra,ut-ngp}!gatech!spaf
...!duke!mcnc!msdc!gatech!spafteus@mcvax.UUCP (07/27/83)
I'm in favour to let the author pay for his article in some way.
This to keep the number of articles low and the quality high.
But I know this is not the solution to the problem.
Newspapers let their subscribers pay for the transport costs
as well. Even there exist some special oversea pricing.
But if there is some way to do something to some high transport costs
it is ok to me.
In the US the uucp system seems to work. However the problem of
phone bills is coming up now.
Here in Europe we had already that transport cost problem.
To solve that problem mcvax created some structure in the net in Europe.
mcvax is calling two sites in the US (decvax and philabs).
For news mcvax knows exactly to whom it is forwarding.
The costs are split among those sites.
It is up to those backbone sites to split the costs to them as well.
(ukc is doing that for the UK, etc).
Only in this way Europe can afford receiving news.
Yes of course some site can play a trick by receiving and forwarding
news cheaper. So it withdraws itself as backbone site from mcvax.
And the price for transport costs to other sites increases, etc ...
The plan is to use X25 between the backbone sites in the different countries.
The transport costs will then not so depending on the amount of traffic
(which is the problem now in the US).
So you create a certain amount of highways between backbones for data.
Still there exists a need for a cheap communication to the US (Tymnet,
DABAS, EURONET). (And if you know please send mail).
The advantage of creating such a structure is:
- people are not posting as wildmen for they know they are billed.
- there is some site looking after the transports.
- there is some site acting as a corrective agent (which is a must!).
- also poor sites can receive news.
And there are disadvantages:
- not sending all groups (which is done by vote now here)
- depending on the willingness of some people to do work (mostly universities).
- it can be ran by dictators.
- some bureaucracy is needed to let everybody pay.
However most of the time to let money roll is more problematic, then
to let data flow. But perhaps it is the difference: from money you want
to have more, from data you want to get rid of ...
I thought the US was growing in this direction of structuring the net.
Teus Hagen, Math Centre, Amsterdam
...{decvax,philabs}!mcvax!teuslarry@grkermit.UUCP (Larry Kolodney) (07/29/83)
From: teus@mcvax.UUCP I'm in favour to let the author pay for his article in some way. This to keep the number of articles low and the quality high. ~~~~~~~~~~~~~~~~~~ In what way is the users ability to pay correlated with the quality of his articles? Do you think that rich people write better articles? -- Larry Kolodney #8 (Moving up) (USENET) decvax!genrad!grkermit!larry allegra!linus!genrad!grkermit!larry (ARPA) rms.g.lkk@mit-ai
john@genrad.UUCP (John Nelson) (07/31/83)
Here's an alternative system: The basic right to post articles probably ought to be free, but there is a lot of junk being transmitted to hundreds of sites. Here's the idea: anyone submitting a stupid or inappropriate article is FINED for mis-use of the net. :-) A stupid article is defined as anything that is prohibited in the nettiquite (sp?) article that is (supposed to be) required reading for anyone using USENET. The money from fines is collected into a general fund which can be used to help defray the cost of phone bills of certain backbone sites (for instance the usa <--> europe link) The amount of the fine could be based on the length of the article in question, and the newsgroup it is posted to (certain sites do not recieve some of the groups) and should be proportional to the actual cost of sending that article around the world. (By the way, has anyone attemted to figure out the actual cost of submitting a single news article? I'd be EXTREMELY interested in the results!) Maybe if the scheme catches on, we could also apply a penalty to boring, banal, or just plain dumb articles. I'm not sure how this would work, but I think some sort of scheme where people can "boo" an article they don't like, and any article which recieves more "boo"s than a certain threshhold is fined! :-) Not only would this scheme help maintain some of the net links about to fail because of phone costs, it should actually reduce the number of articles that we have to wade through to get at the interesting ones! John P. Nelson (decvax!genrad!john) GenRad MS 96 37 Great Road Bolton MA 01720