Let me start, though, by saying it's no surprise that for communicative efficiency the visual system is really a marvelous system for acquiring information about space and for acquiring information about the spacial distribution of forms. It is a system that is simply designed to do that. There are millions of examples that could be cited. For instance, one can ofte n pick up at a glance the city street plan that is shown on a map. I guess it d epends on the city. A glance would probably not be enough for Boston for ins ta nce, but for many cities a glance would suffice.
You can look at a graph with curves on it that depict the relationship between two interacting independent variables, and how they a ffect a dependent variable. This would be rather hard to describe verbally, but a well gra phed display of these variables is commun icated very effectively. And, as I said, there are many other examples that could easily be cited. Now, the reaso ns for this, I think, are pretty obvious.
The capacity of the visual system for pattern resolution is excellent. T he size of the field of observati on, in comparison to the other systems that pe rceive space is very large. There are stimulus dimensions, such as color and br ightne ss, that are available to vision, but not to the other perceptual system s. And so, all of these things account for the fact that the visual perceptual system is the system par excellence for comprehending space - the things in space, the shapes of things in space, and where they are in space.
It's been said, many times, and perhaps too many times, that a picture i s worth a thousand words. Well, in regard to touch, Tim Cranmer tells me that a feel is worth a thousand pictures. I don't understand that statement myself b ut he assures me that it is true. When I heard his story about Eve I began to understand it, but I have to get him to explain it to me more carefully. Be that as it may, our other spacial system, our other system for acquiring information about the spatial extension of objects, is t he tactile or t he tactile-haptic system, or, to put it more simply, the touch system. In compariso n to vision, touch is really much less efficient. Obviously, the capacity for pattern resolution is very much poorer, the field of view is very much smaller, and there are fewer stim ulus dimensions that can contribute information for the comprehension of space. Furthermore, there is no tactile counterpart to the optic chiasma, the function of which is to route neural impulses from each retina to both hemispheres, and no lateral geniculate nuclei which, in t he case of vision, compare the neural impulses generated by stimulating two reti nae simultaneously and extract dept h information. For all of these reasons, it simply isn't as good a spacial system as the visual system. Nevertheless , it is good e nough to provide a good deal of useful information if we learn ho w to take advantage of it. Unfortunately, many blind people do not take full advantage of this ability, and consequently, do not develop much skill i n acquirin g spatial information by touch. A number of reasons can easily be cited. These things begin in childhood, as most things do, even senility. The sighted child has an abundance of experience with graphic displays of many kinds, and gradually learns to understand them. A blind child doesn't have that experience, and to make it worse, the blind child typically gets little or no training in making sense out of graphic displays . What happens as often as not is that a kid gets into high school, and, all of a sudden, somebody gives him a scholastic aptitude test with a question the answering of which requires him to interpret a graphic of a sort that he has never examined before in his l ife. Well, guess what? The kid doesn't do very well.
These are some of the problems, and although we can't redesign perceptual systems, training is a problem that we could do something about. There are other problems. TGDs have, in the past, been fairly expensive, and many school systems have felt that they couldn't affor d them. Teachers have found it difficult to get graphic materials. They observe that blind children don't seem to interpret such m aterials very well, and without thinking very deeply about the reasons for this inability, they begin to feel frustrated, and are at tracted to specious alternatives that seem less troublesome. They can just describe the map or graph, or they can get a tape that d escribes it, and since it appears that blind kids can't understand information presented graphically very well anyway, it seems to t hem that a verbal description is probably the best solution. There are many things we must do to address this problem. We clearly need to know more about tactual perception. There is much to be learned. We need to put much more effort into the development of effective training methods, and we need to begin training at the right age - in preschool, not high school or college. But, we're here today to talk about the prod uction methods that are availabl e for making TGDs.
One point I would like to make before I begin the review of methods. It is that the value of the display produced by a particular m ethod depends upon what you're going to use it for. There are some methods that have rather limited capability in terms of what you can display with them, but properly used, they can b e entirely satisfactory. So, in regard to all of these methods, you have to ask yourself, what are you going to use them for, where , when, under what circumstances?
Embossed Paper Displays The embossed paper displays are the ones that have probably been around the longest. These are made by metal masters t hat are used to emboss the elements of graphic display on paper in the same way that metal plates are used to emboss braille. The Braille printin g houses have made displays in this way for a good many years. They've made some pretty good maps and other displays that have serv ed their purpose. However, paper is not the ideal medium. When you emboss paper, you don't get nice firm edges, and for tactual pe rception, firm edges are quite important. Also, the availability of elevation of form above the surface background as a codable dimension is limited, because the range within which elevation c an be varied when paper is used is very narrow. It's less durable than some ot her materials that might be used, and as elevation is increased, it soon begins to tear. But if you are aware of these thing s and if you are content to work within the limitations of the medium you can definitely make useful displays. A related method, though even more limited, is nevertheless interesting. This method produces graphic displays of the kind that are made possible by t he ability to emboss dots anywhere in a matrix of continuous rows and columns. Most of the braille embossing mac hines sold today have this capability.
The computer- controlled, refreshable displays discussed this morning would be displays of th is sort. The dot graphics produced by Braille embossers or refreshable Braille devices have serious limitations. If the only graphic elements at your disposal are dots, and if they have to be placed on the display surface at only those locations permitted by a r ow by column matrix, many of the features of good TGDs are unavailable. If a graph contains several curves, and if they are close t ogether, or cross one another, the result is likely to be confusion. Dots are the only graphic elements available for graphic compo sition, and so curves cannot be differentiated by the use of different graphic elements. Because dots can only be placed at the locations provided by a row by column matrix, and because resolution is very low, curves exhi bit a quite noticeable stair step effect. Although there is little that could be done to eliminate this problem when a refreshable braille display is used, it should be possible to control the movement of the embossing hea ds on braillers so that the dots they emb oss fall on the curves they are forming. The resulting curves would retain the disadvantages of curves formed only with dots, but t heir shapes would be accurate. However, within those limitations there are some graphic displays that can be quite effective.
Some simple curves may be effective. Some graphs, such as histograms or bar graphs, may communicate as effectively when formed with dots alone as when they are formed with continuous lines or bars. Of course, graphics composed of dots are inexpensive and easy to produce. They are amenable to large scale production, but what really makes them useful is the ability of individuals to produce dot graphics . A person who has a brailler and a computer can generate graphs and other figures to meet personal needs that may arise from time t o time. The graphics just discussed are embossed on paper, and the paper medium imposes further limitations. Dots embossed on paper are les s durable than dots formed on sheets of plastic, and the e levation of dots embossed on paper cannot be varied enough to make elevati on a useful coding dimension, as when graphic elements with different elevati ons can be given different meanings.
Graphics on Plastic Sheets Graphic displays formed on plastic sheets are now widely available. Typically, a master is produced by one of several methods, and a vacuum forming machine, such as the ThermoForm, is used to make copies of the master on plastic sheets. of course, the copy is no better than the master, But if the master is good, the plastic copy has definite advantages over paper. Because plastic is more malleable than paper, much higher elevation of the elements of graphic display is possible. more accurate and finer detail is possible. The medium is more durable. Equipment for making plastic copies of masters is available that is suit able for use by individuals whose needs are met by small production runs. Though I'm not familiar with it, I'm sure there is equipm ent that is suitable for very high production output. The graphic displays of this type most commonly seen are copies of forms that are impressed in aluminum sheets. Judy Tamburlin and Chuck Severin, for instance, have done a good deal of work on the preparation of anatomical drawings. They're ideally qualified to do this work because they're both professors of anatomy. They know their anatomy, and they know when their drawings are correct. As a result, they have made some excellent tangible drawings. At present, they form anatomical drawings by impressing negatives of the drawings that will be observed by the student on the backsi des of thin aluminum sheets. The positives of these negatives appear on the other sides of the aluminum sheets, and these sheets ar e the masters from which plastic copies are made. Recording for the Blind used a variant of this method for many years. The back sides of the aluminum sheets they used had a painted surface on which an image was projected by a projector that reversed the image. Thus, the negative could be formed directly from t his image, instead of having to reverse it mentally in order for it to be correct when the sheet was turned over. RFB also developed a set of tools for making the impressions needed for various points, lines, and textures. The American Printing House for the Blin d sells a graphics kit that includes aluminum sheets, tools like those developed by RFB, and an instruction manual. How good are the masters made by this method? It depends partly on the skill of the person making the master, partly on the tools t hat are available, and partly on the aluminum sheets, themselves. these sheets must be thin enough and soft enough so that images c an be formed. As a result, they are very delicate. They bend and wrinkle easily. A degree of skill rarely encountered would be required to form a perfect positive on an aluminum shee t, and once formed, the masters have a limited lifetime. As duplicates are made, the masters are gradually deformed. And, when the master wears out, you can't recreate it by some cunning method that gives you an exact replica. You have to start all over again. So, although for many of the things we want to do, this method appears to be the least expensive and most practical method at our d isposal, but there is a clear need for a better way.
The Collage Method I should just mention one other variant of the method in passing. One of the things you can do is to glue bits and pieces of material of various kinds to a surface, use that as the master, and make a vacuum fo rmed plastic copy of it. How good is this approach? It is possible to make superb displays this way if you have a great deal of patience and a lot of skill. People who are good at this are the people who like to make model airplanes, boats, cars, and the like. They get satisfaction from cutting out little pieces of material in very precise shapes and placing and gluing them on the surface accurately. The materials that can be used include balsa of various thicknesses, sandpaper, screen wire, sheet cork, or wh atever your imagination suggests to you, and if your imagination is fertile, and you know enough about tactile perception, you can g et some excellent results. But, it is not exactly a production process. It takes a long time. if you're going to make a master fr om which many copies will be made, it may very well be worth the effort.
Deposition Methods Another category includes those methods that depend upon depositing something or other on the surface of paper: wax, plastic, puff i nk, etc. We've seen some examples of these methods here today. The wax images that are deposited by the Dataproducts equipment emp loy a process similar to the process employed by the PixelMaster, which has received s ome attention as a method for making tangible displays. The Data Products machine seems to me to make significantly better images on paper, but the low elevation of the graphic display above its surface is still a serious drawback. How good it will be depends upon how much it gets developed. So we'll have to wait and see on that, but I agree with Tim that it's pr omising. Again, I want to make the point that how good a method is depends on what you use it for. As it stands, the method exemplified by t he deposited w ax method is quite adequate for conveying to the sense of touch the shapes of simple forms, such as print letter shape s, two-dimensional geometric forms such as circles and triangles, etc. If one is willing to respect the lim its of this method, its products can be very useful. The method does have the disadvantage that its displays are produced one at a time, rather slowly, on equipment that is quite expensive.
Another method that has received a good deal of attention makes use of s well or capsule paper. This paper is coated with a layer of plastic that consists of tiny capsules, each of which contains a minuscule quantity of alcohol. An electrostatic copier is used to print an image on the coated surface. The printed sheet is then run through a machine that applies heat evenly over its entire surfa ce. Where the black toner has been deposited by the electrostatic copier, the infrared waves emitted by the heat source are absorbe d, and where no toner has been deposited, they are reflected. The capsules beneath the toner are heated by the absorbed infrared en ergy, and the alcohol t rapped in each capsule expands. This causes the capsules to swell, and the part of the coating on which tone r has been deposited rises above the remaining surface. The result is a tactual approximation of the visual image formed by the toner. The edges of the tangible image are not anywhere near as good as we would like them to be. They tend to be soft and rounded, but the sense of touch is served better by edges that are firm and angular. The surface texture i s sometimes a little bit tacky, which interferes with the examination of t he display by touch. Some people have complained about du rability, although if they are handled with care, they should have a long life. At the risk of being tedious, let me again make the point that, as with the other methods discussed so far, how good a capsule paper display is depends on what you want to use it for. If you use it appropriately, it can really be quite effective. It has one real a dvantage. You can make one on the spot. For instance, a mobility instructor who would like to explain the mobility problem presente d by a complicated intersection, could draw the intersection on a sheet of capsule paper, run the paper through the heater, and show the raised drawing of the intersection to the mobility student. A really serious problem with the capsule paper method is that it tempts people to make raised drawings that are uninterpretable by the sense of touch. A sighted preparer of a graphic display who doesn't know very much about touch will see a beautiful picture, ta ke it to the machine, make a tangible copy that is visually pleasing, and show it with great expectation to the tactual observer, wh o reports that it makes no sense at all.
A little later in today's program, Maurine Edens will be telling you about the puff-ink, silk screening method under development at the American Printing House for the Blind. Since she will be describing it, I will confine myself to a few observations. The metho d is still under development, and it's getting better, but, at least at present, puff-ink drawings have some of the problems of caps ule paper drawings. For instance, like the edges of forms on capsule paper, the edges of forms produced with puff-ink are soft and rounded. They lack the angularity and firmness that would make them easier to interpret by touch. However, the method has the adv antage over some of the other methods I have discussed that it is amenable to large-scale production.
Etching Methods There are also the etching methods. Quite some time ago I tried to collect some information about chemical etching methods and, at that time, I found several that were available, some in the United States and some in Germany. It was possible to get tangible imag es by etching, but the equipment was quite expensive and it was not possible to obtain enough elevation of the elements of graphic d isplay above the display surface to make it available as a codable dimension. Also, the equipment was diffic ult to use. However, j ust recently, I have observed some very good tangible displays formed by the photo etching method. This method makes use of a transparency that is laid on a surface coated with a photo-sensitive material. When the transparency is illuminated, the light strikes the photo-sensitive surface, but it does not reach the photo-sensitive surface where there is image. The illuminated part of the photo-sensitive surface is etched away, and the surface can be removed to a considerable depth. The sp ecimen I observed had good definition (firm, angular edges), and plenty of elevation. I believe that Dupont is the developer of thi s process, and there are now companies where the process has been implemented, and they will use this process to make tangible displ ays from customer-sup-plied transparencies. The etching process may now have reached a developmental stage at which it can make an important contribution to the production of TGDs, and I think we need to investigate its possibilities more thoroughly.
The Milling Method The next method on my list, a method that holds considerable promise, is the method originally developed by Dr. John Gill at Warwick University. The implementation and demonstration of this process was his dissertation project, and what he accomplished was essent ially a proof of concept. To prove the concept, he employed a shop-grade, numerically controlled milling machine, and a PDP8 minicom puter. Thus, although He achieved excellent results, the setup was too expensive to be practical, and the inexpensive equipment that could have made it practical was not then available. Fortunately, the equipment is now available, and there are some promising efforts to take advantage of its capability. A desktop, numerically controlled, milling machine is used to cut a negative image in a sheet of plastic, and the plastic sheet serv es as a mold from which an epoxy positive is made. A computer-aided design package is used to form the display that is to be made t angible, and the display is then translated into the numerical code that controls the milling machine. the positive made from the mo ld produced by the milling machine serves as a master from which duplicates can be made on a ThermoForm machine, or some other therm oplastic forming machine. Gill demonstrated that, using this method, it was possible to produce very precisely formed symbols of various kinds. One of the pr oblems with the graphic displays produced by the commonly used methods is the variability among symbols that are supposed to be iden tical, and this sometimes results in confusion. Consequently, in order for symbols to be discriminable, it is sometimes necessary to make the difference between them larger than it would have to be if they were formed more precisely, and tactual observers would not have to wonder whether an apparent difference was a real difference. In any case, Gill achieved very good results and, of course, with an apparatus of that sort he could produce almost any kind of symbol one could imagine. Many years would pass before it was possible to take advantage of the method proved by Gill, but Karen Luxton and her colleagues at Baruch College have put together a system with which they have achieved some very impressive results. Though it took longer than it should have to follow up on Gill's demonstration, the good news is that the equipment needed for the i mplementation is now readily available. There are commercial sources for the desk top milling machines that can make accurate molds , and inexpensive desktop computers capable of running sophisticated computer-aided design packages are readily available.
As in the case of all of the methods I have discussed, the usefulness of Gill's method depends on the purpose for which it is used. It is capable of producing results of high quality, but appreciable time and effort is required to achieve those results. It would, therefore, not be the method of choice if a mobility instructor wanted to make a one-of-a-kind display to explain the prob lem offered by a particular street intersection. On the other hand, it would be quite suitable for making geometric drawings or sta tistical curves that could be used by many tactual observers for many years. One way in which Gill's method could be simplified would be to eliminate the intermediate step of making a positive from the mold. Later today, a representative of Hi-Tech Systems will describe a cold forming method that makes it possible to create duplicates directly from the negative mold. Instead of making a epoxy positive, you use very high pressure to force the plastic sheet on which th e duplicate is to be formed into the negative image or mold made by the desktop milling machine. At present, the equipment used for cold forming is quite expensive, but it's a beginning, And I think we're getting pretty close to having a practical system capable of making TGDs of high quality.
Refreshable Displays The only other methods that I need to mention are the methods embodied in the computer-controlled machines that produce refreshable displays. When the machines that produce such displays have become good enough and cheap enough to be practical, these displays will definitely be useful. They will be ideal for quick access to many of the displays that appear on monitor screens, but they will no t be very good for presenting elegantly contoured surfaces of livers, uteruses, and mountain ranges. They will retain all of the pr oblems inherent in dot graphics.
Problems Not Solved by Production Methods The method used to produce a TGD is an important factor in determining its utility. Some methods produce better results than others . However, it must be remembered that other factors are also important determiners of utility. The task that often confronts those who are available for the design of TGDs is the task of designing tangible analogues of graphic displays they do not understand because they have h ad no training in the discipline to which the subject matter of the display belon gs, and no appreciation of the requirements that must be met for effective perception by touch. TGDs are usually not exact replicas of the visible graphic displays they replace. Because the haptic system has much less capacity than the visual system for resolving patterns, some of the details of visible displ ays must often be omitted from tangible displays in order to reduce clutter and confusion, even if omission occasionally means the l oss of some information. The scales of drawings must often be changed, and sometimes distorted, so that visible detail will be dis cernible by touch. Because there is no analogy in haptic perception to binocular perception, making tangible the lines that are dra wn on flat surfaces to depict forms with extension in three dimensions produces a tangible display which, in all likelihood, will no t be meaningful to the sense of touch. In order to preserve the information in such displays, TGDs must often be completely redesig ned. The problem is that, to be successful, those who design TGDs must be skilled technicians, must have training in the discipline that includes the subject matter of the TGDs they make, and must have a working knowledge of both visual and haptic perception. This is a formidable set of requirements to meet, and it is not surprising that many of those who design TGDs fail, in some degree, to achie ve optimal results. The same requirements must be met by anyone who undertakes the design and production of TGDs, regardless of the method that is used to make graphic displays tangible. For this reason, an improved method can never be more than a partial soluti on to the problems that are encountered when information is presented graphically for perce ption by touch.