There's a Library in your Television

Presented to IFLA/SLB Conference, Washington DC

Date: 15/08/2001
Venue: Washington DC

Three years ago I read two major reports on the digitisation of the public library services that managed not to mention television once. As I said at the IFLA/SLB meeting in Denmark, we seem to confuse the hardware with the value of the information. The computer is worthy and serious; the television is vulgar and frivolous. A lot can happen in three years. I don't think that attitudes have changed deep down but at least we are getting a much better feel for access to information through television. A substantial number of people whom we don't associate with the library sector are watching sport on digital television, choosing not only games but also camera angles. At the same time, we are beginning to see a ceiling on PC use, at somewhere between half and 2/3 of the population. A third factor, however, of immense importance is the growth of wireless devices such as personal digital assistants (PDAs) and generation upon generation of cellular/mobile phones. This talk is about the library in your television but that is simply because that is the medium that will provide the best access to magnification; but for totally blind people it might as well be the library in your telephone or digital radio.

I am sure you are all perfectly familiar now with the interoperability of technologies and the design of information that does not need to be re-purposed every time a new platform is developed but as technology changes so quickly, perhaps I should say just a few words about where we are and where the evidence says we are likely to go in the next few years.

Putting aside for a moment the problems particularly associated with metadata, the root problem with information design is that most designers come from an aesthetic background; they are, in essence teenagers with Hollywood brains and a nice sense of irony. They know that all that matters on a web home page is that you grab the viewer before he clicks through. This is not likely to change very quickly because assembling multi-sectoral teams to combine such skills as aesthetics, taxonomy, navigation, psychology and human/computer interface is not very easy. There has been precious little inter-sectoral collaboration so far. But I think the problem goes even deeper. The insistence, for example, on proportionate spacing and justified right margins in books has cost the fiction market about 15% of potential sales; but the aesthetic has such a strong grip that it even conquers profitability. I think we are going to have exactly the same problem with digital information design.

The second difficulty we face is that designers of information are still designing what I would call 'valve' information. It is essentially a one-way, stream. The interactivity I described earlier is simply a choice of channels from a supplier. It only offers a set of tools at an extremely basic level. Information suppliers are just about aware of metadata but they are not very interested in tools. You can largely blame this on Anglo Saxon, Protestant, work ethic culture which thinks that tools are somehow a cheat; this is why teachers don't like children to use pocket calculators and why there is still a silly insistence in many schools on writing with fountain pens; even if you're left handed! We are obsessed with process rather than with output; this is an important topic to which I will return later.

Thirdly, where convergence is going on it is not between commerce and government but between commerce and television, with governments trailing rather a long way behind. So far the chief perception of western Governments concerning the use of IT is that it might cut the cost of delivery of government services. It won't, of course. The more effective the transaction the more transaction people will want. Still, the Governments of the Western world are so tied up with access in its narrow, economic and physical sense, tied up with cost cutting and making their information systems attractive rather than accessible, and are so tied up with the baggage of worth and seriousness that there is currently no meaningful alliance, outside Scandinavia and Canada, between government services and television. This problem is, of course, complicated both in public and commercial television by the natural desire of independent media to keep well away from government lest they be tarnished with partisanship.

Having made these introductory remarks, I want to turn to four central issues for us all:

The Manufacture of Information, Metadata and Tools
Maintenance, Accretion and Updating
Mainstream versus Niche
The Limits of Provision.

I will then round off with some remarks on how we might approach a synthesis of many of the topics I have raised.

1. The Manufacture of Information, Metadata and Tools

As I implied a little earlier, making information isn't a simple matter. A really important conceptual breakthrough came with the development of Standard Generalised Mark-up Language (SGML) which significantly disentangled style from content and led the way in the development of the concept of the browser which can render information on different platforms with the same hierarchy but without precisely the same physical manifestation. In our field, for instance, somebody who uses large print may have body text which is the size of the designer's heading level three but it doesn't matter as long as the style sheet is properly established. Well, this was all very wonderful but it was far too elegant and picky, the kind of thing that experts really liked; but it wouldn't sell, so HTML was developed and that was a good deal simpler; now there's XML. Don't worry about all the acronyms, the essential message is that the metadata is interspersed in the text so that the text can be rendered for the user according to the user's requirements. From the point of view of large print and braille this is important because we will soon reach the stage when the output requirement is simply the function of a style sheet.

That's only the beginning. Any element of a piece of information can be given a variety of labels designating significance. In ordinary word processing you can designate a word not only for an alternative font but also to be incorporated into a table of contents or an index. So you could equally designate a whole set of labels to a piece of information. This is pointing the way towards a merger between what we think of as documents and a database; i.e. instead of just putting documents into a database, a document is a mini database. Similar functions are already carried out by précis and parsing engines, by indexing and other forms of metadata and by various kinds of intelligent agents and data mining programmes such as Autonomy. The problem is striking a balance between a narrow set of options which give a big, holistic picture, and a degree of granularity which may simply cause confusion and delay. We know hardly anything about this as it affects general information users but we know nothing about it in the case of blind and visually impaired people. There also has to be a trade off between the number of links per page and the number of clicks to reach a piece of information. This is an important topic if a website is to be accessed by telephone. In fact, one simple design rule to help people who can't see is to design a site and then try to access it by a telephone without a screen.

One thing we have learned from the World Wide Web Consortium Web Accessibility Initiative (www.w3.org/wai) is that writing manuals about these subjects is not all that helpful. Most information providers are not professionals relying on these skills for a living. Even full time information providers who are currently struggling with the digitisation of legacy content don't want a manual. What they want are tools in their information design toolkits that help them to design inclusively, or at least optimally. They also need to be prompted to provide users with tools within the information so that adjustability is not hardware specific but is embedded in the information itself.

That's quite enough of that. In practice what this means is that we need to develop information manufacturing processes which include our required accessibility amongst a whole set of accessibilities. It's my contention that this requires us to make good practice rather than writing manuals on how to make it or urging other people to make it; they won't. The National Centre for Accessible Media (NCam), an independent arm of WGBH Public Broadcasting in Boston, has made a very good start in this area with audio description and text subtitling for deaf people but there is much further to go; at least what it has started with is television; and anybody who can get there during the main IFLA conference should. What's interesting about NCAM from the marketing point of view is that audio description, initially designated for blind people, is now being used regularly by home workers who don't want to be tied to looking at a television screen while doing household chores. It also opens the possibility of re-purposing television soap opera soundtrack for use in car radios. This is a key message for us because although a lot of what we want may be difficult, some of what we want has definite commercial possibilities; there is an uncanny synergy between what blind people and automobile drivers need and between VI use and telephone use.

Finally on this point, although we will continue to concentrate on making information for our clients we ought to concentrate equally on productising our processes and selling them to the information production market.

2. Maintenance, Accretion and Updating

Another thing we all notice about information is how ill equipped we are to keep it up to date. It was all right with the book or the movie, the radio programme or the LP. We made it, we dated it, we finished it. For long periods we made do with information that had been manufactured decades ago. I remember during a dull moment in a presentation given in the braille library in Dublin that I found some braille volumes taken from plates made in the 19th Century. Of course, in a very central way we know that a novel of Dickens does not change but we do not look at information any more as a set of discrete artefacts. What we want is a set of relative commentaries, the ability to provide our own marginalia, for ourselves or for others. We see the text as a muscular phenomenon, as something that breathes. Of course there is a source file but that is simply like a piece of DNA; what counts for us now is the very different life forms which intellectual property creates. This is why document provision and serial updating are not enough. It's also why one-way, reception only systems aren't good enough.

You might ask why this is important in the library context to which my simple reply is that the library in your television can't anymore simply be a virtual bookshelf. The data on new publishing is awesome; the level of specialisation in our culture is daunting; the rate at which we are losing synthesis is alarming; but against that background people still want to relate their experiences to what we once regarded as involate, canonical intellectual property.

Much of what people want to do with intellectual property is neither improving nor even respectable. We cannot go back to my school days when even the Latin text of Catullus was heavily censored in the braille edition. I want to suggest two reasons why the library sector has to be central to the way in which blind and visually impaired people access lifelong information:

The first, based on my earlier discussion, is that libraries are not judgmental as to the merits of process and output. Librarians, who can be very fixed in their views in many ways, are eminently non judgmental about how you get something as long as you get it. That cannot be said about public sector education and training systems which are obsessed with how rather than what. For this reason alone I think that we have to back the library sector and see how close we can get it to television, resisting the general public sector ambition to try to infiltrate television with worthy educational material, almost educating by stealth.

The second reason for choosing libraries is that they are not judgmental with respect to content. In a very important way they are not pastoral; they are not responsible. You don't have to reach a conclusion, hit a target, stay respectable, eschew pornography. Blind people have had very little of that kind of freedom from the education and training sectors, not to mention social services. They have a very poor track record in most countries, running from the patronising to the indifferent. I see the recent emergence of the public library in the life of blind people as a massive plus, particularly in the new age of breathing, interactive information.

Of course, it goes without saying that what I am calling for is a change in the role of librarians but I get the impression that this is much more welcome to professional librarians than it is to governments. The sad paradox is that the more socially democratic a country is, the better its public library provision; but the more inclined it is to be worthy.

The chronic task in the area of information systems is the need for maintenance, accretion and updating. In many ways this will become the central task, along with facilitation which I will deal with in the next section. It is not so much that professional librarians do not recognise this but that it will mean a new kind of relationship between VI consumers and providers which is much more dynamic, two-way, even rather messy.

Finally, on this point, I want to recommend that we try to do some experiments in this area; looking at what our customers know about availability and information churn and looking at what librarians can reasonably provide.

3. Mainstream versus Niche

We struggle all of the time with the issue of where to draw the line between general, public provision and what not-for-profit organisations or charities should do for blind and visually impaired people. For perhaps a hundred years since the birth of braille this was not a serious problem but it is now. For decades special, private library services produced analogue material such as braille and audio books; a few went in for large print. Nobody published more than about 1% of book output per year and the production of magazines, journals, newspapers, brochures, manuals, etc was at an even lower level. Now that is all changing. We expect our disabled citizens to be accorded a service in exchange for the tax they pay and I think all of us are trying very hard to ensure that we copy the best features of the North American and Scandinavian models where alternative format production is seen as a part of the public obligation.

This is easy to maintain in theory but what I am seeking is a much more sophisticated discussion at two levels:

First, how far can general provision, through the way information is made, provide for the needs of our client group, particularly for those with some residual vision?
Secondly, what kind of intermediary and niche services should we concentrate on whether using donated funds or acting as paid agents of the state?

You will recall my emphasis in the first section on the importance of the way that information is made. If we can concentrate our resources on ensuring that general information providers make their information in such a way that it can be accessed by the majority of our clients then it frees resource which can then be concentrated on the smaller number of cases which require special niche marketing approaches: special equipment, special intermediary services, translation software applications, metadata adaptations.

If you analyse the market, a lot of what we want applies to a huge minority of the population, to people with poor vision, people who cannot read small screens. Millions of people who would not and would not wish to be classified as blind or visually impaired say that they just don't see as well as they used to. There is, as I have said, a market but in the disability world we have spent far too much energy defining who sits inside the disability pen with its benefits and special agencies and who's outside it. We would make a better case for what we were trying to do if we saw ourselves as simply solving an extreme manifestation of a widespread problem. You will often hear the mantra that what is good design for disabled people is good design for all. Well, up to a point; but we all know that good design for one kind of visual impairment may be very bad design indeed for another kind of visual impairment; so let's not get too carried away with our own rhetoric. What counts is flexibility in design to accommodate the optimal number of people.

I've used that word "optimal" more than once so perhaps this is a good time to define it. For me optimal means the balance you strike between how much resource something requires and how many people can use it; it's a simple way of describing cost/benefit ratios. So, for instance, to include synthetic speech and magnification in all public information systems is optimal because it wouldn't cost very much if engineered from the initial design and a great many people would benefit; but putting a braille display on the front of every public terminal would not be optimal because it would cost more than the system it fronted and very few people would use it. In this sense speech output and magnification are optimal and braille is niche. This is a practical, technical way of drawing some kind of line between public and niche service and it is a much better way than defining the difference by reference to epidemiology or ideology.

Of course, as you would expect by now, it's not that simple. There are three recent developments which make braille, for instance, much less niche:

The first is that digitisation radically lowers the price of producing braille text; if it's digital and manipulable it can be put into braille.
Secondly, it is now perfectly possible to think of braille as just another style sheet which can be used to render a stream of text. It will work much better in Braille Grade One where there are very few rules and most of the expression is 1 : 1; and that's a price I am prepared to pay for taking a huge amount of braille production out of the special agency remit.
Thirdly, with wireless communications protocols like WAP and communication conventions like Bluetooth it will soon be possible for all of us to have just two access devices; one we nail down at home and one we walk with; again, this will make braille access much more optimal because we will be carrying our personal access instead of relying on a braille display at a public terminal. It will be a while yet but ultimately all the software will be in the server and our braille displays will look like a ruler rather than a big piece of kit.

Still, when all is said and done, our clients will still need highly specialised intermediary and facilitation services. No matter how well the information is made for optimal access and no matter how cheap and flexible the access devices, there will still be problems in areas such as:

the ordered reporting of searches from databases
allocation of value to information
graphic description and interpolation.

I think the first of these is relatively simple to understand. Turning information distributed in more than two dimensions into a linear report can be quite challenging but we are all familiar with it. Even turning two-dimensional information into linear form is troublesome. Think of the technique for brailling a family tree.

Assigning value to information is perhaps a little less familiar and I want to illustrate this without using slides. Think of a square which frames nine elements in three rows and three columns. You could drop a graphic of a piece of fruit into each cell and this might simply be a display or it might be a menu for on-line shopping. It might, however, be the report of a daily lottery based on fruit machines where the order of fruit in the rows and columns has some significance.

Next, think of a strategy model based on the same nine elements or cells in three rows and three columns. All that you have is an indication that the horizontal axis represents good value and the vertical axis represents high quality; the first attribute is represented by red, the second by blue. As you move from right to left the red becomes paler, as you move from top to bottom the blue gets paler. So the top left cell, row 1 column 1 stands for the best fit of value for money and quality with the bottom right, row 3 column 3, being the worst. This framework is used to represent the attributes of an array of products which can then be grouped in each cell. Visually this is quite straightforward but representing it in text only is, as you've just heard, quite a challenge.

Now think of the same square with its nine cells as representing expenditure; the horizontal axis represents location and the vertical represents sector; so you can add up the expenditure across three sectors in one location by totalling the figures from left to right; and you can add up the figures for one sector in three different locations by totalling from top to bottom. Not only is this quite difficult to tabulate but you have the extra problem in a dynamic setting that when one figure changes the horizontal and vertical sub-totals change and the extreme bottom right figure for grand total also changes.

As for rendering graphical and pictorial information, you are all familiar with audio description and picture captioning using the Alt tag in Web design; but I never miss an opportunity to press all of us to think more carefully about teaching, examining and giving credit for description; just because people are good braille transcribers it does not mean that they can describe a Breughel or the quality of light in a Vermeer.

The other aspect of intermediary work we should look at carefully is the handling of legacy information and hybrid information systems. For a complex of reasons I won't go into now, our client group is usually a good couple of steps behind peers in technology, software and access. It will be a long time before everything is backward compatible and before new designs are so good that they are, to some extent, forward compatible. Just as we had that terrible few years between DOS and Windows, so we are bound to have the same problem over and over again; and we will need people to take material we want, feed it through systems and tidy it up before it is sent back to us in our required format. This may not be very glamorous work but it's essential if we are to get access to most of our information through mainstream sources. If there is going to be a library in our television it's going to be as big and untidy as the real world of our sighted peers, not tiny, comfortable, bespoke but out of date.

My conclusion is that the line between mainstream and niche may move slightly as technologies change but that there will always be a core function for the niche market in the areas of information interpretation and facilitation. Who pays for it is quite another matter.

4. The Limits of Provision

Which leads me into my final main section. I want to give a very strange, counter intuitive final major message at the end of this conference. We must not, and we must not allow our clients, to expect too much from us. Just as we don't want a tiny corpus of knowledge that's out of date, so we can't have absolutely everything in the format we want. It will, of course, be a lot easier if we are not so heavily bound by the analogue old guard who think that the processes of alternative format production are sacred. To listen to the prelates of the International conference on English Braille talking about a unified code you would think that what was at stake was not the future of the universe but the nature of eternal life. Only if we free ourselves from that, only if we are prepared to sacrifice that last 5% of processing which takes 95% of the time and whose only real achievement is elegance, only then can we rapidly expand our access but, even then, there will be severe limits. The library in our television needs to be thought of in practical terms, as a way of enjoying a right; we don't want to get this practical ability tangled up with a principle of universal accessibility to everything. It isn't, anyway, possible. If a congenitally blind person listens to a sports commentary where there is a close call by an umpire she can't ever say whether the umpire made the correct decision; she can only place weight on the reliability she puts on commentators. This is true in other spheres too. There is information which is more or less worth concentrating on. It won't be the same for every individual but we have to give ourselves a chance of being both just and fair by coming to recognise some reasonable limits to the special services we provide. These limits are not just a matter of money. We have to recognise, reluctantly but perhaps, that being blind or visually impaired presents such drawbacks that the cost of remedying some of its shortcomings is too high compared with what could be done in the same time, for the client as well as the service provider.

In our schools, for instance, we can't expect blind children, already severely handicapped by the time they get to school, to follow the general curriculum, the additional curriculum and the hidden curriculum; we are going to have to make some trade offs if they are to emerge as normal, balanced, adults. Going back to one of my earlier points, we have to concentrate much more on output and much less on process.

5. Synthesis/Conclusion

I would now, you will be relived to know, like to pull some of this together into a couple of major, general conclusions.

The first is that the key factor for us must be information design which allows us to take what is on the market and put it through special software programmes, if they are not part of the design itself, in order to render information in a variety of accessible formats. This is a tall order but it's actually easier than using our scarce resources to try to produce alternative format material as if we were living in the analogue age. If we continue in this way, the amount we produce may go up but the percentage of what is available will go down.
Secondly, we need to think of ourselves as intermediaries, facilitators and translators, mentors, guides, trusted agents. We will need a completely different set of skills if we are to act at one remove from the client, factoring and pumping information.
Thirdly, we will also have to stop being so medium bound; blind and visually impaired people don't have one medium, so if we start from the single file source to produce braille, large print, synthetic audio, we will be getting there.
Fourthly, we must stop the benefits and donations systems from imprisoning us inside a VI sector; all disabilities are on a spectrum from severe to mild, and the milder the manifestation, the more people have it. We need to make a market case for what we're doing and that depends on general usability, not on the narrow issue of disability access. I recognise the legitimacy of the United States rights-based approach and do not see this in conflict with the marketing approach but we must always be prepared to operate a twin strategy.

Finally, it's all a good deal too serious. It's bad enough being blind without all the people who seek to serve you wanting to improve you as well. All the matters I've discussed are extremely serious but our output does not necessarily need to be. I would prefer the qualities of television to spill over into the monasticism of the library rather than the other way round.

Let us bow to the vulgarity of the television in its own right as at least equal to his Lordship the computer.

Return to Information Design