65th IFLA Council and General
Bangkok, Thailand, August 20 - August 28, 1999
The Digital Library - Prospects for Developing Countries
The FORCE Foundation
The Hague, Netherlands
Much or even all of what is contained in this paper may already be familiar to you. In order to build up the picture of the potential of digital libraries it is necessary to go over some of the basics. If they are all too familiar, bear with me until we get to the main argument.
At a time when digital technologies are developing and converging at ever increasing rates, the thought of proposing that the libraries for the blind and visually impaired should adopt such technologies is met with the harsh realism of their situation.
How can one expect libraries in developing countries to "go digital" when they don't have stable electricity supplies, they don't have secure and air conditioned rooms, there are no technicians and no repair facilities? In whole or in part many countries have these deficits and they might appear to be fully justified in rejecting the idea of the digital library.
If some well-meaning benefactor were to parachute all the equipment necessary to run a digital library into a developing country, then in all probability the people receiving that equipment would not be able to succeed. What I hope to do in this short paper is to point to some of the advantages of taking the decision to "go digital" and to outline the steps that may have to be taken in order to secure success in the venture.
What do we mean by a digital library?
In the full definition a digital library would be one in which all the texts and the spoken books would be held as digital files. But that will take a long time to achieve. Let us start by considering text files. If you have a text as a word processor file on a floppy disc or the hard disk you have the basis of the digital library. What are the advantages of the digital file? It takes up very little room. One can make safety copies in seconds. From the digital files on can drive a printer - either for a standard printout or for large letter. One can even define the needs of particular clients and produce a copy that is of the size and typeface that the partially sighted can best read.
As you all know there are, in many languages, software programmes that can automatically translate from the original letters into braille - many of these also contracting the braille to fit the particular rules of a country or language. Once the text has been passed through this software one has yet another digital file, but now in a code which generates braille characters rather than letters. It is important to keep this in mind since we will be coming back to these digital braille files.
The digital braille file can be used to drive a braille printer. Suddenly we can break away from the former restriction of whether to print a whole book. If you are hand-making a book the tendency has been to make the whole book. Most libraries have rules about producing the integral text of a book. But if we regard the creation of the digital braille file as the act of production, so that we have the integral text should we need it, then we can save time and money by printing off in braille only those parts of a book that are needed. This does not apply to novels but may well apply to a study book or a reference work. Over time the savings can be considerable. Remember that the text and the braille files can both be used in this way.
Does this mean that all the customers have to have computers?
If you could guarantee that all your clients had computers adapted to their particular limitations and that they had printers at home (and supplies of paper) to braille out materials that they wanted to read directly - then you might set up a totally electronic service. Not even the richest and most advanced country in the world can guarantee this level. A less wealthy country may have few, if any, clients with computers. But that doesn't matter. What the argument presented here suggests is that the libraries should go digital in order to produce traditional reading materials at lower costs and faster.
Leaving aside audio books for the moment (though they will be dealt with in depth during the conference) let us stay with texts. How much time does it take to prepare a page of braille by hand? Any mechanical method is slower than using a computer (assuming a moderate level of operator skills). If libraries or the braille production centres start to use computerised methods then dramatic improvements can arise in the production of braille text.
To change the infrastructure within a country is a very different question. However desirable it might be to equip all visually impaired people with audio players and adapted computers this is just not feasible. In countries without a technological infrastructure it makes no sense in the short term to assume that one can change the main forms of materials and the methods of distribution.
Digital Production of texts
There are three main methods of producing digital texts. The simplest is to use a scanner and a good Optical Character Recognition program that turns the scanned text from an image into letters that can be handled by a word processing program. The accuracy of these OCR programs is now very high. The clearer the printing in the original the fewer mistakes there will be in the transfer. Of course one then has to run a spelling checker and proof read the resulting file. However if one uses a sheet feeder, whole books can be turned into digital files in tens of minutes rather than tens of hours. The availability of good software varies depending on the sort of alphabet that is used, of course.
The second and most common method is to copy-type the text. In libraries for the blind such a method is reasonably efficient. The text still needs to be proof read but a sighted typist can key in text using a conventional keyboard at speeds far in excess of traditional brailling.
The third method is to get the text directly from the publishers as a digital file. In theory this should be possible. After all, the printer has produced the book from a digital file. In reality these files are difficult to find, are often full of setting codes and in any case the publishers are frightened that somehow the files will escape onto the open market. Experience shows that it is often quicker to scan or re-key the work.
As indicated above, once one has a digital file then passing this through a braille translator to arrive at a braille file which can then be used to drive a braille printer.
Using other peoples resources - don't go it alone if you can help it
To produce the braille file the library will need a computer, the right software and a printer. There are countries that have hesitated to go towards digital production because there is no stable electricity or no technical back up. There are some counter arguments that will be looked at later. The main point here is that no library should think of itself as being isolated. There are many other braille libraries around the world who will be able to lend some help. It is also most likely that a library just getting into digital production will be in or near a major centre. There are banks and insurance companies, industrial concerns and shops that will already be using computers. They can be approached for help if needed. Some of them might even be prepared to pass on computers as they up-grade.
Join the world wide family of libraries
By deciding to change to digital production, a library makes it much easier to co-operate with other libraries around the world. The more that a library can communicate with others the more they will be able to learn about what has already been done. One of the ways we waste time and money is to try to invent everything ourselves. Whatever you are trying to do in developing your library you can guarantee that somebody else has already done something similar. It may not be in your language or in your cultural and political circumstances but we can all learn from each other (especially from the mistakes of others if they are prepared to talk about them). It has to be acknowledged that in the world of libraries for the visually impaired, the English language dominates in terms of numbers of libraries and the amount of production. One only has to think of the production of the UK, North America, and the countries of the British Commonwealth to see that these are some of the most developed and prolific. Spanish also covers a large part of the globe with ONCE as the driving force. There are more Francophone countries in Africa than Anglophone, but the later are much more advanced in terms of production of alternative format materials. It can be argued that in such a case the need for co-operation within the Francophone area is even greater if they are to maximise their scare resources. No matter what the working languages are of your library, if there is even one other library that produces the same sort of materials in that language then it makes sense to co-operate with them. If two libraries can agree which books they are going to produce then they can avoid the expense of making the same book twice. If they can then agree to share the braille files then they will effectively have doubled their production capacity. Think what this could mean for a language area in which there are tens of libraries
Build up a network to make traditional materials more efficient.
If we are working with traditional methods and the only means that we have of sending materials to other institutions is by the post (or camels as in Kenya) then it is understandable that libraries tend to concentrate on their own affairs and their own public. If it takes weeks to communicate with other libraries then the effort becomes too much and it is faster to produce material than to borrow it.
If however, libraries can start to use the Internet to exchange information and materials, the exchange can happen in seconds. Within a group of libraries it is possible to establish a simple system based on the sending of messages "Has anyone got a copy of book X?" or "I have this problem, has anyone got an answer?". But once libraries start to work together it makes sense to be able to look into the catalogues of the other libraries before making a decision. One can think of making a common union catalogue between the partner libraries or of setting up a distributed catalogue. The first will require putting all the entries into a standard form, preferable one accepted by the library world and based on a MARC format. The second entails having the server at each library permanently accessible and also that each catalogue is so constructed that it can be interrogated by a common search tool (e.g. Z39.50 compatible). One should neither underestimate the amount of effort this will take nor the great value that it can bring.
Once an item has been found in the catalogue it should be possible to download it or request it from the other library. Sending a braille file is the easiest method of exchange but carries with it some potential problems. There are many different formats for printing braille; numbers of lines per page, page size, number of characters per line can vary. In general libraries in one language area tend to use the same form of Grade 2 braille if contractions are used, but even this is not consistent. In an ideal situation a group of libraries would all be using the same braille format, but this world is rarely ideal. One way round the problem might be to agree to send the braille file in the version before it was formatted for printing. It is extremely unlikely that there will be any problem with copyright since only braille readers will be able to use the file and most laws allow for such exemptions. If libraries are to start exchanging digital texts then the publishers may take a much more serious interest.
Libraries may argue the need for exchanging text files where these are to be used to create large letter output. For the client the advantage is that the library can produce the text in a size, font and spacing which is best suited to the individual's visual impairment. In these cases care must be taken with copyright clearance if required by national or local laws.
Such networking of catalogue information and texts has the great advantage of speed and access to greater resources. However it does push the actual production down the line away from the central libraries or production houses, nearer to the end user.
This paper has been concerned with text to this point. Digital audio files can also be accessed in the same way, but one has to be aware that transfer of such massive amounts of data can take a long time. Once down loaded the file then has to be transferred to a different carrier. It might be more effective to treat audio recordings as a separate issue, whether dealing in traditional analogue carriers such as audio tape or more modern alternative such as DAISY encoded files on CD-ROM, DVD or Flash Cards.
The need for training
Though the gains for the libraries and clients can be great, it has to be recognised that the introduction of new technologies implies different methods of production and different tasks. People need to be trained not only to carry out the tasks but also to administer the systems. There will be a need to set up different levels of courses. Most of the work implied by what has been described above is for sighted operators. If an institution has a policy of employing visually impaired people then they will have to consider carefully where they fit into such a digital structure and devise appropriate training.
The need for technical support
The more machines that are brought into the production system the greater the need for good technical support. There have been too many cases of machinery being donated without care for support. The first time that the machine breaks down could be the end of the production process if there is no one trained to repair it. Mechanical devices such as braille embossers need mechanical skills. Since the suppliers are few and far away basic training is required. Computers are another matter. The machines and the software are developing rapidly. The adaptive software may not work easily with the standard packages. While help may be obtained on-line in some cases it is advisable to have someone on call to help. Few libraries can afford all the computer support that they require in the form of permanent staff. It is therefore essential that computers are purchased locally from companies which can guarantee back-up when it is needed. As indicated earlier it may be possible to lean on the specialists in large companies such as banks. You might even get them to understand and share the objectives of the library.
You can start small.
Despite the apparent complexity of the new technologies, cooperation between libraries can start at a relatively simple level. Just being able to send electronic mail to each other is a good start. Join some of the information exchanging structures such as special list-servers or bulletin boards. You would be amazed at how many other people working in the world of provision of support for visually impaired people there are who are willing to offer helpful advice. One Internet connection can put you in direct contact with many of the larger libraries for the visually impaired reader, around the world. Start at a level that you can sustain with local support, but share your burdens with other libraries. The increase in efficiency and the ability to provide more with the budget that you have will be worth the effort.