George MW writes:

I played around with MN for a while, one of the things that disapointed me was that you tend to lose mojo just by running the client. It seemed to me that the idea that everything you do should involve a cost one way or another was a mistake, it seemed to me that the housekeeping/say hello kind of stuff shuld be free. Publishing also, since 1) there's no guarentee that the site you publish blocks to will keep them for any nonzero length of time and 2) Theoretically the people you publish to will expect to make their "profit" from other people requesting those blocks.

This has been a long debate among hard-core capitalist types who want to see everything have a price. In a voluntary system for exchanging and publishing information, who should pay and who should be paid? Of course the easy answer is that the market will decide. But when you're trying to design the mechanisms then you have to anticipate the market's needs. You have to predict what the market will decide. Unfortunately, the realistic answer seems to be that many possible patterns of payment will be desired by the market. In some cases providers of data will pay others to read it: advertising, political or persuasive arguments, reputation-building essays. In other cases people who want the data will pay those who provide it: music and other arts, entertaining and informative works. MN identifies three roles: author, consumer, and publisher (they don't necessarily use these names). The author creates the data (or supplies data authored by someone else). The publisher provides the resources to make it available: disk space, network access. The consumer downloads the data and uses it. In some cases the author and publisher may be the same. It seems reasonable that the consumer would pay the publisher for most data. Presuming that the consumer is going to the trouble of downloading the data, he must want it and so he will be willing to pay for it. There may be some examples as noted above where the author wants to subsidize the consumer's access. In the MN model, the consumer pays the publisher, and the author pays the publisher as well. This latter aspect is the most controversial. Why should an author pay to get his data distributed? Well, as noted above there are some cases where it makes sense, but in those cases the author also wants to encourage consumers to download his work. Just putting it on servers is not enough. MN does not provide for a subsidy or payment to downloaders. Ideally the publisher should be able to pay the author for the work. This corresponds most closely to the traditional publishing model. Talented authors get paid by their publishers, who then receive payment themselves by consumers. For this to work though the publisher must be able to make judgements about which authors are talented. In the traditional publishing world this is handled by reputation and human judgement. It might be possible to automate this to a considerable extent using cryptographic reputations, but it is a new area that would need much infrastructure development. A mailing list such as this one is a microcosm of an information exchange, one where no payment is possible. Each person who posts puts some effort into creating his message (some more than others). Some people's messages are worth reading and subscribers might even be willing to pay to get them. Others serve more to gratify the ego of the writer and he might have to pay to get his contributions read. Imagine a cash-based mailing list (perhaps using play money). Each contributer would either pay or get paid for his submissions, depending on his current reputation. Subscribers would have a quantity of cash flow that they could funnel towards the reputation of the authors they enjoyed the most. Slashdot's karma system is similar to this and it does a pretty good job of separating the quality from the crap, although there are complaints that it promotes groupthink. A system based more openly on cash payments could be a model for future information markets.