Currently, distribution costs for small (approximately one megabyte) entropic items are roughly a dollar. For example, all the text in a paperback book adds up to about a half a megabyte; the artwork on the cover might require another megabyte; and such books can be purchased, in large volume, for a few dollars apiece.
Voice telephone connections in many countries are slightly cheaper than paperback books. A telephone connection can carry a quarter of a megabyte of digitized data per minute and costs tens of cents per minute, that is, about a dollar per megabyte.
Full-colour, medium-resolution imagery can now be printed on $300 printers for tens of cents per page, that is, for about ten cents per megabyte. People don't want to read text with a magnifying glass, so we can print at most a few kilobytes of text per page, at a per-page printing cost of a few cents. This works out to about a dollar per megabyte.
It should be clear from the foregoing analysis that people residing in New Zealand will not participate in as many small entropic transactions as will people in the United States, given the noticeably higher prices for such entropic distributions for some entropic media (such as books and point-to-point telephony) in this country. I strongly suspect, however, that we will soon enjoy ``world'' pricing levels here, even on books and long-distance telephone connections. Indeed, books are now readily available from overseas vendors, at prices close to world levels, as long as orders are placed in ten-book lots to amortise shipping costs. Long-distance telephony may soon be priced here in New Zealand at approximately ten cents per megabyte, after we have paid off the capital costs of establishing enough high-bandwidth transoceanic connections.
Despite our temporarily high charges for one-megabyte transactions in some media, New Zealanders are not disadvantaged for entropic transactions involving tiny (one kilobyte) items in electronic mail format. In this case, the distribution costs are negligible in comparison to the capital costs of setting up and maintaining an email communications system on, typically, a personal computer at home.
When we browse the world-wide web on a personal computer at home, we obtain a mixture of tiny (one kilobyte) textual items and small (one megabyte) graphical ones. As noted above, the latter are rather expensive here, by world standards.
For slightly larger entropic transactions, of approximately a gigabyte, CD-ROM disks are an attractive format. Two disks can hold a gigabyte, and these can be produced and distributed anywhere in the world for a few tens of dollars. CD-ROM disks, then, are almost one hundred times cheaper per byte than a telephone connection or a book. However each transaction with a CD-ROM involves a gigabyte and perhaps a week of time for airmail delivery. So they are only suitable for distribution of moderately-sized entropic items such as encyclopedias without full graphic content, small collections of high-quality imagery, and ten-minute video clips of moderate (television) quality.
Ten-gigabyte transactions can now be sent by DVD-ROM disk to the few people who have replaced their CD-ROM drives with units that handle both CD-ROM and DVD-ROM formats. DVD-ROM distributions are ten times cheaper per byte than are CD-ROM distributions, and their capacity is appropriate for feature-length video and encyclopedias with imagery similar to what is available in a typical printed 20-volume library set.
Within a decade, large (one terabyte) transactions should be routine, as they should fit on a single high-density magnetic tape cartridge costing about ten dollars.
Within two decades, huge (one thousand terabyte) transactions may well become feasible, worldwide, sending magnetic tape by airmail for tens of dollars. This may approach some fundamental limits on information density in magnetic media, so further reductions in entropic distribution costs cannot be projected with any certainty. Note that a single thousand-terabyte tape would hold all the information in a hundred university-sized libraries, and all the information in a hundred thousand feature-length videos. So it may be that we find few uses for such huge distributions. It seems to me more likely, however, that we will merely find commonplace uses for huge entropic distributions; such uses may be as inconceivable today as was our current world-wide web browsing only two decades ago.
Calculations of entropic distribution costs, such as the above, are commonplace in analyses of running costs for future libraries. The cost of producing, shelving, and distributing paper documents is not dropping anywhere near as rapidly as the cost of handling analogous items in electronic format, strongly suggesting that libraries should attempt to use electronic distributions whereever possible.
This tendency toward the ``electronic library'' is countervailed by most people's preference for reading from paper rather than from a screen. In other words, the value of a paper document, to most individuals, is much greater than that of a similar entropic item available for viewing on a computer screen. Thus we must remember to consider the ``syntactic carrier'' (that is, the structure of the entropic medium carrying the information) when evaluating an entropic item.
The projected drop in entropic production and distribution charges will, it seems clear to me, put huge stress on our current systems for rewarding authorship. Currently, we can pay authors satisfactorily out of a small percentage markup on production and distribution costs, for most media. For example, a paperback book with a production and distribution cost of a few dollars can easily support a per-sale royalty payment of a few tens of cents.
Already this cost-plus system of compensating authors is failing, in the case of distributing textual material by CD-ROM. As noted above, a single CD-ROM disk can contain the text from hundreds of books, at a production and distribution cost of ten or twenty dollars per unit, worldwide. If textual authors were paid a royalty of ten percent of these costs, they'd get about one cent per sale, and they'd have to sell a million ``books'' in CD-ROM format to get a royalty of ten thousand dollars. If books were distributed in (two thousand-volume lots) in the DVD-ROM format, each book's author would get only a thousand dollars for every million in sales. If this isn't worrisome enough, consider that by my projections, a textual author in the year 2017 would get only a pittance for a billion-selling title in high-density (one thousand terabyte) magnetic tape distributions, if these were priced near to the cost of production and distribution.
Clearly we must find some ways to charge by entropic value, rather than by production and distribution cost, in the near future.