#AUTHOR: Durre, Karl P.; Tuttle, Dean W. and Durre, Ingeborg

#TITLE: A Universal Computer Braille Code For Literary And Scientific Texts

#ORGANIZATION: International Technology Conference, December 1991

#CATEGORY: Braille, Braille Code

#PUBLICATION: Unpublished Paper

#ABSTRACT: With the advancement of computerized braille not only into the offices of braille transcribers and transcribing agencies, but also into braille users' homes and offices, the demand for a computer braille code that combines literary, scientific, and mathematical notation in a universal computer braille code is more stringent than ever.

A Universal Computer Braille Code For Literary And Scientific Texts

Karl P. Durre, Dean W. Tuttle, and Ingeborg K. Durre

With the advancement of computerized braille not only into the offices of braille transcribers and transcribing agencies, but also into braille users' homes and offices, the demand for a computer braille code that combines literary, scientific, and mathematical notation in a universal computer braille code is more stringent than ever.

Originally, braille was designed for use on paper. Soon, however, the bulkiness of paper braille contributed to the development of contracted braille. With computers around at the time, the demand for a contracted version of braille might not even have arisen. The often heard claim that contracted braille also increases reading rates might not have sufficed as a reason to develop the grade 2 braille code.

Another problematic aspect of braille is the limitation of the number of characters resulting from the braille system itself. The amount of existing print symbols by far exceeds that of the original six-dot braille system which is restricted to a maximum of 64 different combinations of dot distributions in the braille cell, including the blank symbol. This limitation has led to the development of different codes for literary, mathematical, music, and foreign language codes. Through an overlapping of symbols used in each special code, the integration of these special codes into a universal code is impossible. The extension from a 6-dot to an 8-dot code helped in some way byincreasing the number of possible representations to 256 different characters.

It is true, however, that the introduction of special codes for blind individuals has led to their segregation from the sighted population in the area of written communication. Today, with the computer as a mediator allowing direct communication between sighted and blind computer users, this main disadvantage of braille has lost much of its former significance. When using a computer, there even is no longer a need for contracted braille, since there is no difference in storing print or braille electronically, and the reason of bulkiness of uncontracted braille is no longer valid. Print spelling appears as a natural consequence for non-visual computer users, and should not be of particular concern for them, as typewriting has always been part of the curriculum for blind persons.

With the use of a computer, the internal representation of characters allows a simultaneous conversion to print on the monitor screen and braille on the refreshable braille display. This permits an immediate conversion of print to braille and braille to print, entailing an improvement of the direct communication between sighted and blind individuals as well as direct access for non-print readers to all electronic texts available to the sighted population [2,3,5].

The internal computer representation of characters, the ASCII code, is, in fact, being used by many blind computer users who access general application software non-visually.

However, the computer braille code that is widely used in the U.S., the so-called MIT code, is based exclusively on the internal representations without consideration of the formation of the respective braille symbols. For the extension of the six-dot braille code to an eight-dot code, its selection of the characters that use dots 7, 8, and 7 and 8 does not take into account the mnemonic correlation between semantically related characters. For instance, the 'less' and 'less or equal' signs are represented in the MIT braille code as 'dots 1,2,6' and 'dots 2,5,8'. It is obvious that these two braille signs do not show any resemblance whatsoever. The use of a pattern for the second sign that resembled the 'less' sign would permit the logic identification for the reader, thus increasing the ease of reading and memorizing the symbol. From that point of view, it would make much sense to assign 'dots 1,2,6,7' or '1,2,6,8' to the 'less or equal' sign.

The design of the universal computer braille code presented in this paper is based on the logic structure of the braille character set rather than on internal computer representations. Special consideration also went into the correlation of mathematical and foreign language characters. This has resulted in a user-friendly universal braille code that permits the combination of literary English texts with foreign language and/or mathematical texts.

Due to the limited amount of symbols available on the computer, there are still certain restrictions in the area of mathematical texts that exist for print and braille users alike. In those cases, therefore, consent has to be achieved as to the representation of mathematical formulas on the computer. Suggestions for representations [6,7] follow later in this text which have been in use since 1983 and have proved to be successful in that their easy readability was enjoyed by both blind and sighted users alike. Their design was based on the way of representation commonly used in higher programming languages. The introduction of this new universal computer braille code has major advantages over other braille codes in use at this time in the following four areas [5,6].

First, the supply of texts in the school and college environments is simplified considerably, since entering texts into a computer does not require special knowledge of braille.

Second, there is no need for braille transcriptions into print, since print is produced simultaneously and in the same process as writing braille. This makes text provision simple, cheap, and fast, compared to the traditional way of braille production and braille transcription.

Third, texts in different languages that require various foreign language symbols can be written in the same file so that it is possible for the blind user to keep multiple language texts conveniently together.

Finally, the mathematical symbols and formulas represented in this universal computer braille code, which includes all commonly used higher mathematics symbols and is not restricted to the symbols listed, directly follow the image of their print equivalents. They can also be combined with literary text without switching from one code to another. All braille composition signs are discarded, thus removing a major obstacle for blind persons to do mathematics as well as for the communication between print and braille users in general. Problems with time delay that exist with traditional transcribing of literary texts, and are increased exponentially with mathematical texts due to the special expert knowledge necessary for mathematical braille writing, are avoided with the universal computer braille code presented here. The code should be regarded as a basis for the standardization of a universal computer braille code.

The Design of the Universal Computer Braille Code

In 1987, the Braille Authority of North America (BANA) adopted a 'Code for Computer Braille Notation" [1] based on 6-dot braille. The basic 64 characters that are represented in the BANA code by one single braille cell each has formed the basis for the definition of the 8-dot code described here. With one exception only ( 'grave accent' instead of the 'at' sign), the suggested code uses the same characters that make up all those braille symbols in which neither dot 7 nor dot 8 occurs in raised position. Table 2 gives a list of those symbols in the third column. They comprise all 26 lower case letters, all punctuation symbols, several mathematical symbols, and the three special symbols #, $, and &.

Table 2

Table 2(Continued)

Explanations for table 2:

Column 1: Line number

Column 2: Basic six-dot braille pattern for all characters in same line; a raised dot is represented by its number in the given sequence.

Column 3: Basic characters (Lower case letters, digits, punctuation symbols, basic mathematical symbols, and some special symbols):

ASCII and print representation of the character assigned to the basic six-dot braille pattern in column 2 (i.e. where neither dot 7 nor dot B are raised).

Column 4: Shift (dot 7) characters (Upper case letters and additional mathematical and special symbols available in the MS-DOS system):ASCII and print representation of the character assigned to the braille pattern that consists of the six-dot pattern in column 2 with dot 7 raised additionally.

Column 5: Alternate (dot B) characters (special lower case letters from other languages including Greek letters): ASCII and print representation of the character assigned to the braille pattern that consists of the basic six-dot pattern in column 2 with an additional dot B raised.

Column B: Alternate-shift (dot 7&B) characters (special upper case letters from other languages): ASCII and print representation of the character assigned to the braille pattern that consists of the respective six-dot pattern in column 2 with additional dots 7 and 8 set.

Column 7: Comment for columns 5 and 6; letters D, G, F, and I stand for Danish, German, French, and Italian respectively. Each capital letter denotes that the six-dot braille cell in column 2 represents the special letter in column 5 and/or 6 in the braille code of the respective language. An '*' denotes that, due to collision, the respective cell is assigned to a different braille cell. Example: Row 28 - braille pattern 1256 represents letter u in German, French, and Italian braille, as well as letter u in Spanish braille; u has been assigned to row 41, i.e., it is represented by braille pattern 23456.

There are 111 more characters (including the 'hard space', ASCII 255), 48 graphics symbols, and 33 control characters available on IBM Personal Computers and compatibles that run with the MS-DOS system (see e.g. [8]). As we are only concerned with the access and manipulation of text data, control characters and graphics symbols are not being considered in our code design. For that reason, we are free to assign the braille patterns containing dot 7, dot 8, or both dots 7 and 8, with the following underlying systematic pattern of assignment.

Shift Symbols (Dot 7 Symbols)

The shift symbols represent all those braille cells where dot 7, not dot 8, is raised. This group comprises the capital letters and all science and mathematics symbols (including currency symbols and the @ sign) that are not represented by a basic six-dot cell. They are listed in the fourth column of table 2. Symbols that are not represented on the computer keyboard are identified by their ASCII number (decimal) followed by their print fonts. The assignment of braille cells to characters is done in the following order:

- Capital letters are represented by adding dot 7 to the six-dot pattern of the respective lower case letter, as is the case in the MIT code.

- Symbols 123 {, 124 }, and 125 | have the same basic braille pattern as [, ], and \ because they are similar by means of semantics, print pattern, as well as their assignment to the same keys on the keyboard. The two remaining symbols available directly on most QWERTY keyboards, 64 @ and 126 ~, correspond with 96 ` and 95 ^ , i.e. all 95 symbols of the keyboard are placed together in the third column and the upper part of the fourth of table 2.

- Symbols 239 fl (intersection), 249 •, 241 ±, 170 ¬ (negation), 246 ÷, 243 <, 240 _, and 242 _ have the same six-dot patterns as &, *, +, -, /, <, =, and > respectively and are only distinguished from those by an additional dot 7 (& being interpreted as the logical 'and', * as multiplication symbol, and / as replacement for the fraction bar in mathematics notation).

- The inverted exclamation mark, 173 ¡, and the inverted question mark, 168 ¿, correlate with ! and ? respectively.

- The degree sign, 248 °, has the same basic six-dot pattern as the double quote sign merely for the reason that this sign will be interpreted as 'minute' when measuring angles in geometry.

- Mainly for reasons of similarities of the print symbols, the symbols 159 ƒ (actually the symbol for currency 'Franc', but used also as integral sign), 251 (radical), 236 (infinity), 174 «, 175 », 169 _, and 250 · (small bullet) have the same basic six-dot braille patterns as #, $, %, (, ), comma, and period respectively.

- The remaining 12 symbols, 155 ¢, 156 £, 157 ¥, 158 Pt, 171 ½, 172 ¼, 244 , 245 , 247 , 252 n, 253 ²and 254 , are, more or less, assigned arbitrarily (see Table 1). The symbols 244 and 245 (top and bottom half of an integral sign) are not useful for our purposes as the resulting sign extends over two lines; we have, therefore, chosen the 'Franc' sign ƒ to represent the integral sign.

Table 1

Alternate Symbols (Dot 8 Symbols)

All letters different from the 26 letters of the English alphabet are represented by braille patterns where dot 8 is raised. Those where dot 7 is not raised represent lower case letters, those with dot 7 raised represent upper case letters. The former are listed in column 5 of table 2, the latter in column 6 of that table.

For special symbols of Danish, French, German, and Italian alphabets, the basic 6-dot pattern is chosen to be identical with it's 6-dot representation in the braille code of the respective language. However, in different languages the same braille representation is used for different symbols [10]. In table 1, the conflicting letters are listed.

The majority of special French and Italian letters and their respective native braille representations coincide. However, they conflict with specific Spanish letters. Therefore, we have a choice either to assign the native braille representations to specific French/Italian letters or to the specific Spanish letters, listed in table 1. The code presented in table 2 has given preference to French and Italian. All letters marked with '*' are represented by a braille pattern different from that in the respective native braille code. However, a code with 'Spanish preference' can easily be deducted by exchanging the assignments of the respective symbols.

Braille representations for upper case foreign language letters are selected similarly and written with dot 7 raised additionally. They are listed in column 6 of table 2.

Lower and uppercase Greek letters are written with dot 7 raised, or with dots 7 and 8 respectively. The 6-dot pattern is taken from the respective Latin letter, e.g., a,ß, etc., have the same basic 6-dot pattern as a, b, d, e, f, m, p, s, and t respectively (see column 5 of table 2). The capital Greek letters (symbols not available) have the same basic pattern as f, g, h, o, and s (see column 6 in table 2.

Writing Mathematics

Table 3

Table 3(Continued)

Most existing mathematical print symbols are included in the code given in table 3 and can be immediately interpreted by braille and print readers. Only the two-dimensional notations have to be written slightly differently so that they can be easily recognized by the braille reader [3,5,6]. Formulas are written in one line like in programming languages, e.g., FORTRAN. Only matrices and determinants may comprise several lines In daily school and college life, this method has proven to eliminate the barriers for blind students in mathematics and science. No knowledge of braille or - even more important as there are only very few braille transcribers for mathematics - no knowledge of a special mathematics braille code is necessary for a sighted teacher or aide to produce mathematical texts for a blind student, and the teacher can directly read any formula in print that a blind student writes in braille. Ordinary text and formulas may be mixed arbitrarily as they are written in one unique code.

Conclusion

With the acceptance of a universal computer braille code that includes representation of mathematical and foreign language texts as the code described above, we can finally take advantage of the manifold benefits of the computer for the production of literary, mathematical and foreign language texts, books, and magazines. Paper brailling mathematical texts in the Nemeth code [9] and foreign languages [10] is one of the most difficult tasks for braillists, and there are not many persons who are skilled enough to braille such texts. Furthermore, the procedure is extremely time- and money-consuming.

With a universal computer braille code as the one suggested here, we can finally make a step into the future in that we can use the existing computer programs for the production of braille literary, mathematical, and foreign language texts. This will mean an immense improvement for braille users as well as teachers in the mainstreaming scene, braille producing individuals and agencies, as well as for all individuals engaged in written communication between blind and sighted persons.

Due to the current lack of a universal computer braille code, the non-visual computer user so far was factually married to the company that had produced the computer access equipment of his or her choice. Access to little used texts now is made so much easier, cheaper, and faster, and, therefore, feasible for the individual. The argument of disproportionate costs and efforts will be a thing of the past. Furthermore, typing errors are very hard to correct on paper braille. With the universal code, anybody can key in mathematical texts on the computer keyboard, and braille typists can enter the braille version of a text on the braille keyboard of the braille terminal. Sighted persons now are able to follow what has been entered in braille, and can continue where the braillist has left off, and vice versa. Corrections no longer are a problem.

Thus, the full advantages of word processors can finally be used for the production of braille mathematical texts as well as for combined literary, scientific and foreign language texts.

It is the purpose of the code presented here to serve as an instigation towards the introduction of a standardized universal computer braille code.

References

[1] Braille Authority of North America (1987). Code for computer braille notation. Louisville, KY: American Printing House For The Blind.

[2] Durre, K.P. (1990). BrailleButler: A new approach to non-visual computer applications- Proceedings Third Annual IEEE Symposium on Computer-Based Medical Systems. University of North Carolina at Chapel Hill 1990. Los Alamitos, CA: IEEE Computer Society Press.

[3] Durre, K. P. (1986). Braille and advanced rnan-computer-interaction. In J. E. Ebersold, Th. Schwyter, & W. A. Slaby (Eds.). Computerized braille production. Proceedings of the 5th International Workshop, Winterthur 1985. Eichstatt, Germany: Katholische Universität.

[4] Durre, K. P., Durre, I. K., & Durre, I. (1989). The BrailleButler manual. Version for IBM-PC and compatibles. Fort Collins, CO: Preprint.

[5] Durre, K. P., & Durre, I. K. (1986). Electronic paper for blind children. Education and Computing, Special Issue: The Computer in the Home, 2, 101-106.

[6] Durre. K. P., & Durre, I. K. (1983). ASCII-kompatible Mathematikschrift unter Verwendung der Zuordnungstabelle nach DIN. Karlsruhe, Germany: Unpublished manuscript.

[7] Durre, K. P., Mehring genannt Friehoff, T., & Schmidt-Lademann, F. P. (1984). BrailleButler: A successful microcomputerbased aid for mainstreaming blind children. Proceedings Third Annual Workshop on Computers and the Handicapped, Wichita, KS.Silverspring, MD: IEEE Computer Society Press, 89-96.

[8] MS-DOS 3.2 for the Epson Equity+ System (1986). Microsoft Corporation and Seiko Epson Corporation.

[9] Nemeth, A. (1972). The Nemeth code of braille mathematics and scientific notation. Louisville, KY: American Printing House For The Blind.

[10] World Braille Usage. National Library Service for the Blind and Physically Handicapped. Washington, DC: Library of Congress.