Computer Braille (August 8, 2012)

Executive SummaryThe Braille Authority of North America (BANA) is proposing to adopt a unified braille code to replace the translation capabilities of English Braille American Edition (EBAE) for ordinary text, of Computer Braille Code (CBC) for computer items, and of the Nemeth Code for mathematics and technical text. The two unified braille codes that are being considered are Unified English Braille (UEB) and Nemeth Uniform Braille System (NUBS), which are both transcribing codes. (Transcribing codes are ones with the main goal of efficiently representing general documents via the use of indicators, context, formats, and particular types of dot patterns.)

This article touches briefly on the different purposes of computer braille and of transcribing codes and points out that no transcribing code can provide an adequate substitute for all the technical purposes and functions of computer braille. It then compares the computer braille capabilities of UEB and NUBS with those of CBC and 8-dot braille. It concludes that both UEB and NUBS provide an adequate computer braille function within the context of a transcribing code but speculates that UEB's use of contracted braille in computer items may be harder to adapt to. (An earlier version of this article incorrectly stated that UEB's use of shortforms in computer items could lead to ambiguities. I apologize for this error, which resulted from my initial failure to understand UEB's special rules for the use of shortforms in computer items, and thank the reader who pointed this out.)

NUBS requires a leading one-cell notational mode indicator before all computer items but does not require additional mode indicators elsewhere in a computer item. UEB does not require a leading indicator to distinguish computer items but does require the use of its Numeric Indicator to set numeric mode each time a numerical sequence including one or more digits is present in a computer item. Also the UEB dots-56 grade 1 symbol indicator is required to terminate numeric mode if one of the small letters a-j occurs when the mode is still in force. Both codes require capitalization indicators for capital letters in computer items.

There are 68 ASCII characters not counting capital letters that can be present in computer items. Of these 68, NUBS represents 44 characters with one-cell symbols: the 26 letters, the 10 digits, and 8 special characters. NUBS represents the remaining 24 characters with two-cell symbols. Of these 68, UEB represents 33 characters with one-cell symbols: the 26 letters and 7 special characters. UEB represents 25 characters with two-cell symbols. Isolated digits and the leading digit in a numeric sequence are represented with the appropriate two-cell Numeric Indicator. Any remaining digits in a numeric sequence are represented with one-cell symbols.

Computer Braille

The general term "computer braille" refers to any six-dot or eight-dot braille code which provides convenient and unambiguous symbols for the 94 ASCII or Unicode Basic Latin characters found on a standard computer keyboard in the United States. The primary purpose of computer braille is to facilitate print-to-braille and braille-to-print conversion when it is important to know exactly which print characters are used.

Since there are only 63 braille cells, any six-dot computer braille code must use at least some braille symbols with more than one braille cell. However, an eight-dot computer braille code can be designed to provide one-cell symbols for all 94 ASCII characters. In this latter case conversion between braille and print isn't necessary as both braille and print can be directly entered from either a standard keyboard or a braille keyboard.

Some of the goals of a computer braille code are inconsistent with the goals of a transcribing code including a unified transcribing code. This issue has been discussed in detail in the article "The Importance of 'Computer Braille' to Braille Users".

Various Computer Braille Alternatives

The next section evaluates eight-dot computer braille. The following three sections evaluate the six-dot computer braille options in Computer Braille Code (CBC), Unified English Braille (UEB), and Nemeth Uniform Braille System (NUBS).

Most braille users with RBDs eventually prefer eight-dot computer braille in preference to any of the six-dot computer braille alternatives because of the significantly greater convenience of avoiding the need to backtranslate.

Eight-dot Computer Braille

A standard way of specifying an eight-dot computer braille adds a seventh dot to 31 of the related six-dot patterns. Typically a seventh dot is added to the 26 small letters to represent the corresponding capital letters and to five other six-dot patterns to represent five special characters. BANA does not have an official eight-dot computer braille code but eight-dot tables are built into Refreshable Braille Displays (RBDs). The eight-dot tables used in the U. S. are usually extensions of North American ASCII Braille and employ six-dot dropped numbers. EuroBraille, which is one of the standard eight-dot tables used in Europe, uses the six-dot dot-6 or Antoine numbers.

English Braille American Edition (EBAE) and Computer Braille Code (CBC)

CBC is an official BANA six-dot code which specifies one- and two-cell braille symbols for the 94 ASCII characters. CBC could be used as a stand-alone code but is typically used as a separate mode in EBAE literary braille for brailling "all e-mail, website and other Internet addresses, file names and the like." Embedded CBC is entered within EBAE via either the dots 4-5-6, 3-4-6 Begin CBC indicator or (when appropriate) the dots 4-5-6, 3-4-5 Caps Lock indicator. Embedded CBC is terminated by the dots 4-5-6, 1-5-6 End CBC indicator. Displayed CBC does not require the use of these indicators.

CBC is as similar as possible to six-dot North American ASCII Braille. CBC uses the standard one-cell symbols for the 26 small letters and the one-cell dropped numbers for the 10 decimal digits. It also uses one-cell symbols for all of the special characters except for Low Line, Grave Accent, Left Curly Bracket, Vertical Line, Right Curly Bracket, and Tilde. CBC specifies six indicators: the one-cell Shift Indicator, dots 4-5-6, and five two-cell indicators. The latter group is comprised of the three mentioned earlier; the dots 4-5-6, 1-2-6 Caps Release Indicator; and the dots 4-5-6, 1-2-3-4-6 Continuation indicator.

CBC represents single capital letters by preceding them with the Shift Indicator. It represents sequences of two or more capital letters by preceding them with the dots 4-5-6, 3-4-5 Caps Lock indicator and by terminating them with the appropriate one of the dots 4-5-6, 1-2-6 Caps Release Indicator, the dots 4-5-6, 1-5-6 End CBC indicator, or a space.

CBC has two advantages over other six-dot computer braille alternatives. First, it simplifies interpretation by using the maximum possible number of one-cell symbols. Second, it is very similar to North American ASCII Braille. The main disadvantage of CBC from the perspective of contracted braille use are the many symbols for special characters that differ from those used in contracted braille. This is, of course, a consequence of CBC's maximizing the use of one-cell symbols.

United English Braille (UEB)

UEB does not have a separate "computer braille" mode. The method used for brailling computer items is to some extent left to the individual transcriber although it is recommended that contracted braille be used for embedded computer items per these two rules from the UEB Rulebook.

5.9 Choice of indicators Note: An extended grade 1 mode, i.e. grade 1 word or grade 1 passage mode, may be used for non-literary expressions. This is especially useful in mathematics and computer programming texts. However, keeping in mind the general principle that the resulting braille should be as easy to read as possible, it is often appropriate to use contracted braille with a grade 1 indicator for just those symbols that can be misread as contractions. It is recommended that contracted braille is used for email addresses, filenames and web addresses.

10.12.3 Use contractions in computer material, such as email addresses, web sites, URLs, and filenames when it is embedded in regular text. Use uncontracted braille for computer material, such as computer program code which is displayed on separate lines, as well as any nearby excerpts from the program. [From UEB Rulebook]

It might seem that UEB's use of shortform contractions in computer material could lead to ambiguities. However, special rules limiting the use of these contractions in this context have been designed to eliminate this possibility.

Numbers in computer material, just like numbers in other material, must be preceded with one of the UEB Numeric Indicators. Also, a grade 1 indicator is required to terminate Numeric Grade 1 mode when a lowercase letter a–j follows a digit, a full stop/period or comma.

Capital letters in computer material are brailled just like capital letters in other material in UEB. Individual capital letters are preceded with the dot 6 capital sign. Sequences of two or more capital letters are preceded with the dot 6, 6 capitals word indicator and terminated by either a space or the dot 6, 3 capitals terminator. Passages of three or more word-like items where all letters are capitals are preceded by the dot 6, 6, 6 capitals passage indicator and terminated by the capitals terminator after the last item.

UEB has at least three disadvantages in comparison with the other two six-dot alternatives for a computer braille. First, the braille reader must be aware that computer items can include contractions and, in particular, must understand the UEB rules for the use and non-use of shortform contractions in computer items. (For example, an apparent shortform sequence, such as "brl", in a computer item is intended to be read as a letter sequence, not as a shortform.) Second, it requires extra indicators whenever there are embedded numbers. Finally, its smaller overlap with and less similarity to the symbols used in 8-dot computer braille, 6-dot ASCII braille, and CBC means that the many braille users who need to learn one of these other systems would have to remember a greater number of different braille symbols than if they were using a transcribing code with dropped numbers.

Nemeth Uniform Braille System (NUBS)

Computer items are brailled using the NUBS notational mode which is the same mode used for mathematics. Individual items must be preceded with the appropriate one-cell notational indicator, either the "number sign" for items beginning with a digit or the "letter sign" for items beginning with other characters. Phrases must be preceded with the appropriate two-cell notational indicator and terminated in a corresponding manner.

Individual capital letters in computer material are preceded with the dot 6 capital letter indicator. Sequences of two or more capital letters are preceded with the dot 6, 6 capital word indicator and terminated by either a space or the dot 6, 3 terminate capital indicator.

There are four characters used in computer items, the semicolon, colon, exclamation point, and question mark, which require a preceding dots 4-5-6 punctuation indicator because of the use of notational mode. Also there are two characters, the period and the comma, which have different one-cell braille symbols in notational mode than in the default narrative mode used for ordinary text.

NUBS does not share the disadvantages of UEB as a computer braille alternative. It is more print-like since it does not use contractions in computer items and it doesn't require extra indicators for embedded numbers since it uses dropped numbers. Moreover, there is a greater overlap with and greater similarity to the symbols used for the ASCII characters and the symbols used for the same characters in CBC and typical 8-dot computer braille in NUBS than there is in UEB. This may be a significant advantage since many braille users will eventually need to learn another computer braille code to complement their use of a transcribing code.

NUBS makes the opposite trade-off from CBC. It maximizes the use of the same braille symbols for computer items as for other items rather than maximizing the use of one-cell symbols in computer items. NUBS thus uses two-cell symbols for 24 special characters in contrast to CBC's use of two-cell symbols for only 6 special characters.

Computer Braille Symbols

The 94 ASCII characters are conveniently grouped into the 26 Latin Small Letters, the 26 Latin Capital Letters, the 10 (Decimal) Digits, and 32 special characters.

All four alternatives described earlier use similar approaches for the Small and Capital letters. The next two sections detail the various approaches for digits and for special characters.

Digits

In all the systems described here with the exception of UEB, the digits are represented using dropped numbers which are unique one-cell symbols. Since UEB uses upper numbers, which re-use the cells for small letters a-j, digits must be represented in Numeric Grade 1 Mode. A grade 1 indicator is required to terminate Numeric Grade 1 mode if a small letter a–j follows a digit, a full stop/period or comma.

Special Characters: UEB versus NUBS

Computer braille requires representing 32 special characters. UEB uses two-cell symbols for 25 special characters and one-cell symbols for the remaining seven: vertical apostrophe, comma, full stop, hyphen-minus, exclamation point, colon, and semicolon. NUBS uses two-cell symbols for 24 special characters and one-cell symbols for the remaining eight: vertical apostrophe, comma, full stop, hyphen-minus, left parenthesis, right parenthesis, plus, and equals. (The comma and full stop symbols differ from those in NUBS narrative mode.) In NUBS notational mode the dots 4-5-6 punctuation indicator must be inserted before the NUBS narrative mode symbols for the semicolon, colon, exclamation point, and question mark.

When brailling computer items UEB and NUBS use the same two-cell symbols for ten special characters: vertical quotation mark, dollar sign, percent sign, ampersand, slash, less-than sign, greater-than sign, commercial at, reverse solidus, and tilde. However, UEB and NUBS use different two-cell symbols for eleven special characters: number sign, asterisk, question mark, left square bracket, right square bracket, circumflex accent, low line, grave accent, left curly bracket, vertical line, and right curly bracket. [Note that there is some uncertainty as to the proper UEB symbol for the question mark in computer items. UEB uses dots 2-3-6 to represent both its opening nonspecific quotation mark and the question mark where the reader is presumed to distinguish them from context. However, UEB Rule 7.1.3 states, "Place a grade 1 symbol indicator before a punctuation mark which appears in a position where it would be read as a contraction."]

The remaining four UEB two-cell symbols are for special characters for which NUBS uses one-cell symbols: left parenthesis, right parenthesis, plus sign, and equals. The remaining three NUBS two-cell symbols are for special characters for which UEB uses one-cell symbols: exclamation point, colon, and semicolon.

It is arguable that the NUBS symbols for the number sign (hash mark) and the four brackets are more systematic than the corresponding UEB symbols. The NUBS number sign is the traditional dots 3-4-5-6 preceded by dots 4-6 whereas the UEB number sign is dots 1-4-5-6 preceded by dots 4-5-6. The NUBS brackets use the same symbols as for its parentheses but preceded by dot 4 for the square brackets and by dots 4-6 for the curly brackets. The UEB parentheses and brackets are all two-cell symbols where the second cells are dots 1-2-6 for the left symbols and dots 3-4-5 for the right symbols as are also used for the second cells of the UEB less-than and greater-than symbols. The first cells are dot 5, dots 4-6 and dots 4-5-6 for the parentheses, square brackets, and curly brackets, respectively.

NUBS also makes better use of tactile mnemonics than UEB does: the shapes of some NUBS symbols are more informative than the corresponding UEB ones since they resemble the corresponding print glyphs. See for example the symbols for the asterisk, circumflex accent (caret), low line (underscore) and grave accent.

A third simplification in NUBS is that its two-cell special character symbols only use dot 4, dot 6, dots 4-6, and dots 4-5-6 as prefix cells. UEB, by contrast, uses all seven cells with dots only in column two as prefix cells in its two-cell special character symbols.

UEB vs NUBS for Computer Items

It is difficult to use examples to compare UEB with NUBS for computer items since examples can be chosen so as to showcase the advantages or disadvantages of one or the other system. However, this example from p. 52 of the UEB Rulebook illustrates some typical differences between the two systems.

• Print: jim@take2.com
• UEB:  jim@atake#b4;com
• NUBS: ;jim@atake2]com

[Note that the UEB Numeric Indicator always sets Grade 1 mode. This means that if the email address had been jim@take2.edu, the UEB reader would be aware that the required dots 5-6 cell following the "dot" is a Grade 1 symbol indicator and not the first cell of the final-letter contraction for "ence."]

Revised and corrected version posted August 8, 2012.