Welcome once more to the Library of Babel.  At today's lunch we will be discussing some of the economic issues raised by our uniquely comprehensive acquisitions programme, and a possible solution.
First, the problem. This Library's goal, as those who have visited previously will know, is to acquire or otherwise collect and (eventually) to catalogue all expressions of human knowledge. We have previously discussed questions raised by the huge outpouring of grey and darker material through the new electronic media, and will have to return to these.  Today we will concentrate on the problem of commercially published material in the new media - notably the World Wide Web and its successors.
There seems to be a consensus that, whatever the rights and wrongs and ideals of the matter, increasing amounts of important reference and cultural material will be available only in electronic form. 
Some of this material will be available on subscription. What is a librarian to do? It is already increasingly difficult to justify paying subscriptions for journals on paper which may be consulted only a few times a year, or less often. How can we make choices about subscribing to electronic journals, when we will not even be able to show the accountants physical evidence that we have acquired material for future reference?
The problem of subscriptions is exacerbated by the fact that we cannot, as things stand, archive or acquire journals in any meaningful sense. Many of the proposals for business models for electronic publishing suggest that material will be delivered in encrypted form, with "keys" to read it a set number of times.  In a sense, what we are being offered here is a paid-for equivalent of the inter-library loan: rental direct from the publisher. Even where there is parallel publishing on paper and in electronic forms, we can expect that an increasing number of our users will be accustomed to the power of free-text searches and will demand electronic access. 
Clearly, we cannot afford to subscribe to all possible electronic books, journals and services. Nor does any library's budgets permit subscribing to all those which users might reasonably be expected to want or need. Nor, we fear, will it be possible to pay for subscriptions as users make requests.
I have previously argued against the subscription model of journal pricing, not least on the grounds that its dependence on restricted access renders our automatic cataloguing schemes difficult or impossible.  We do not, however, have control over the business models adopted in electronic publishing: subscriptions there will be, at least for a transitional period which may be decades long.
The preferred pay-per-read business model is not without its problems for this Library, nor for those with more restricted acquisitions policies. How can we exercise any form of budgetary control? For some years now, information services have been available on the internet which charge significant sums for small, sometimes very small, chunks of information.
The NewsPage service from Individual,  for example, charges up to £5 for some 100-word items, some of which are news-wire stories and some of which turn out to be press releases. They may, however, be exactly the source which our user requires. The US Dun & Bradstreet Business Background Reports cost $20.  We could quite legitimately charge the Dun & Bradstreet report through to a local business user; but what about A-level economics students, or tenants seeking information on their ultimate freeholder, or even squatters doing the same?
In short, as a smaller proportion of the information which users want and expect is held on paper on our library's physical premises, we're in a mess. The fact that this is partly to do with the increasing quantities of information to which users expect instant access does not allow us to evade the problem. If our wealthier users cease using our facilities because they can bear a £500-a-year information budget in their home offices, while our less-wealthy users are condemned to wait weeks for information which the rich can get in seconds, then the division of our society into the information-rich and the information-poor will be a reality. 
Nor is the argument restricted to reference information although, after 20 years in which The Market has been the official religion, force of habit leads me to cast arguments in these terms. Our libraries exist after all, not just to make data available, but to make recorded culture accessible to all, regardless of ability to pay.
Culture, cutting-edge culture in particular and High Culture even, increasingly means much more than books. Future historians who want to understand the last decade of the 20th century in the UK will (probably) need to understand the environmental protest movement and (almost certainly) the whole area of popular culture associated with dance music. Their understanding of either will be impaired if they do not have access to, say, a CD-ROM issued this September by Ninja Tunes  with pictures of one cut to the sounds of the other.
There I go again, justifying access to information on the grounds that there is a research and, hence, ultimately a financial interest. Libraries - not just public libraries - should make culture available to the public. It is time to make that argument clearly, to prepare for the time when The Market is seen for what it is, an intellectual fad serving particular interests and, on the time-scale on which libraries such as ours must operate, a passing fad at that.
What, then, is to be done?
To list some possibilities:
We can imagine some sort of electronic analogy to the copyright deposit law, under which owners of on-line information must grant free access to accredited libraries.
This has an instant intellectual appeal. "Information," notoriously, "wants to be free". Academics, in particular, having become inured to giving their work lock, stock and barrel to so-called "learned" publishers (or even paying page charges to persuade them to take it away), like to dream that all the important information in the world is free for the taking. Few such people, it can safely be assumed, have paid sustained attention to the budget or accounts of the library which offers them the illusion of free information.
But, remember, we are not talking about books, nor necessarily about fixed documents in any medium: some of the most useful "publications" will bear more resemblance to live radio than to a manuscript. Nor will we be able safely to draw a line, if we ever could, between "serious" and "entertaining" stuff. To illustrate this, readers in the UK should do the following: stay up until 2am; watch the "keywords" strand on BBC2. Note that its design cries out to be hypertextual: alongside the parallel moving pictures, text and voice there ought to be live links to movies, documentary newsreel, and scientific papers. Now decide where the entertainment ends and the serious stuff starts.
Consider, then, the rules and regulations to which libraries would almost certainly have to submit in order for such a "real-time copyright deposit" law to stand any chance. Imagine, first, Disney's lawyers' and market-droids' reaction on hearing that Snow White is to be available, free, in libraries - and, even worse, to be available whenever the library is open, foiling their plans to deliver carefully rationed doses to the market.
I see a nightmare outcome in which library users are expected to sign affidavits that they have Had No Fun, with periodic inspections by teams of corporate lawyers to ensure that no fun has, indeed, been had.
Imagine if we had done nothing when printing was introduced, and had comfortably remained repositories for manuscripts. ("Everything of real value is a manuscript," I hear some dry stick mutter.) The effects of ignoring the new media will be similarly marginalising.
Alternatively, libraries could respond as they have to video - and become, in effect, second-rate retail hire outlets for the new media. All the arguments about access to information, including cultural information, being independent of ability to pay would thus be lost.
Another economic model is readily available in the UK: the Public Lending Right. This recognises that the information which is libraries' lifeblood is produced by breathing humans, not corporations, and that these humans invest their lives in producing it in the hope of reward when - and whenever - it is consumed. Because of the limitations of book technology - the fatal lack of interactivity, in summary - Public Lending Right payments are distributed on the basis of statistical surveys of library loans.
Extend this to a Public Reading Right. ("Reading" sounds more worthy and less politically threatening than "Multimedia" or "Access to Information", and we can take it in the post-modern sense of "reading" television.)
Each time information - text, moving pictures, or as-yet-unnamed media - is accessed in a library, a credit is made to the author or authors. This will all happen automatically and will involve librarians in no day-to-day management work: various schemes are well under way to develop systems for collecting small payments to authors and re-distributing them economically.   These systems will have, before too long, to deal with the issues of distributing payments for access to collaborative works. It is likely that authors would agree to payment under the PRR at a reduced rate, compared to the "commercial" price of their work.
PRR payments would be met from central government funds. The most cynical argument for this is that it is the price of avoiding the growth of an information underclass. (Perhaps an even more cynical argument, based on nationalism, can be made. To be brutally honest, PRR is the price of ensuring that the information underclass forms elsewhere.)
PRR, like Public Lending Right, has the advantage of an appeal to fairness and the fostering of creativity. It will be a payment to authors, for an entirely new right which they cannot legally have signed away.
It does, of course, raise the possibility of budgeting problems for libraries. If it were funded by libraries themselves, it would in the near future lead to local newspaper stories about school pupils being prevented from handing their homework in on time, and students delaying their dissertations, because budgets had run dry. We will have to give careful thought to the implications of the obvious alternative, which is that PRR payments should be met from a central government fund.
PRR will also expose yet another conflict between the border-less nature of the new media and the archaicisms of national government. The solution, however, is simple: it should be introduced as soon as possible in the UK, as a model for the EU and then the rest of the world. Payments to authors who live in jurisdictions which do not yet have PRR or distribution mechanisms will be held on trust until such time as their governments and authors' organisations catch up. (The Norwegian authors' rights collection agency Kopinor,  for instance, successfully operates on this basis.)
The most difficult problem, however, will be to define what, or rather where, is a library. I have a solution, though it is not perfect. PRR should apply to information accessed by users on the physical premises of a library, elsewhere in a recognised educational institution, or in a health centre or GP's surgery.
On the one hand, this restricts PRR to its primary goal: ensuring that information is available, free at the point of use, to those who cannot afford to obtain it in the comfort of their own homes and for whom an office is probably somewhere where they claim benefits. On the other, it has disturbing resonances of a restricted role for the library. Why should libraries as institutions be restricted to physical premises? That, like many things, will have to be the subject of a further discussion.
 For a particularly fine example of this genre, see Disgruntled Housewives of the World Unite at http://www.urbekah.com/housewife/index.html. To digress: the nature of the cataloguing problem may be brought home by the fact that the Librarian's Assistant came across this while searching for a contact address for Ninja Tunes, a cutting-edge producer of dance music. Someone who also likes the Tunes also likes the Disgruntled Housewives. A proposal for a quantitative study of converging tastes in unrelated fields, with the goal of generating a Hilbert space of all possible personalities, is now being drawn up.
 See "Time to Shelve the Library?", published in New Scientist 5 December 1992 and archived (as submitted) at http://www.poptel.org.uk/nuj/mike/articles/ns-elib.htm
 As an example, recently the Librarian's Assistant paid a visit to the Science Reference Library for the first time in some years. Having spent a large part of the intervening time working on the internet, he was amused to find himself composing an AltaVista query as he approached the reference desk. So he asked, "err... how do you do a text search on this thing?" To two librarians the request appeared meaningless. The third, younger, librarian responded appropriately: "Over there. Subject catalogue. 'Aardvark'. Read."
 NewsPage is at http://www.poptel.org.uk/nuj/mike/ttp://www.newspage.com/
 See for example "What does it mean to be information-poor?", published in the New Statesman and archived at http://www.poptel.org.uk/nuj/mike/articles/infopoor.htm