| UNIBETACODE(5) | File Formats Manual | UNIBETACODE(5) |
unibetacode - Format for polytonic Greek Beta Code files
source_file.beta
Unibetacode is an implementation of Beta Code, as adopted by the University of California, Irvine Thesaurus Linguae Graecae (TLG) Program and the Tufts University Perseus Project, among others. Beta Code provides a way of encoding polytonic Greek characters using plain ASCII characters. The unibetacode package contains three utility programs: unibetaprep(1) converts TLG-unique numeric codes to Unicode code points, beta2uni(1) converts a Beta Code file to UTF-8 Unicode, and uni2beta(1) converts a UTF-8 Unicode file to Beta Code. These programs can also process Coptic and some Hebrew, but historically the focus of Beta Code documents has been classical Greek.
A Unicode code point is an assignment to a specific numeric value for glyphs and other entities in Unicode fonts. Throughout this document, Unicode code points are given by their Unicode numeric values in the form U+xxxx, where "xxxx" is a string of four hexadecimal digits representing a glyph in the Unicode Basic Multilingual Plane. This is how they are usually specified in The Unicode Standard and elsewhere.
Note: Thesaurus Linguae Graecae and TLG are registered trademarks of the University of California.
Regardless of the language mode (Greek, Latin, Coptic, or Hebrew), several punctuation marks are retained as is between the input file and the output file. They are as follows:
The TLG Beta Code specification supports nine different styles of quotation mark. While support of different styles is beneficial, this complicates round-trip conversion from Beta Code to Unicode and back. This is further complicated by the same Unicode character being used as an opening quotation mark in one style and as a closing quotation mark in another style.
Double quotes in the unibetacode package just use ASCII quotation marks in a Beta Code source file; the quotation style is determined by the language mode. This greatly simplifies round-trip conversion between Beta Code and a Unicode UTF-8 document. Double quotation marks must be balanced. The first quotation mark encountered will be interpreted as a left quotation mark for the chosen style, the second will be interpreted as a right quotation mark, and so on. Double quotation marks are converted as follows:
Single quotes are specified explicitly. For Latin and Hebrew, they are specified with a Grave Accent (U+0060) for an opening single quote and an apostrophe (U+0027) for a closing single quote. For Greek and Coptic, they are specified with a "<" for an opening single quote and a ">" for a closing single quote. The use of "<" and ">" for Greek and Coptic was a compromise so that an ASCII apostrophe in Greek mode would render as U+02BC, MODIFIER LETTER APOSTROPHE. As with double quotation marks, the rendering of these characters is dependent on the language mode. Single quotation marks are converted as follows:
The unibetacode package includes two extensions to TLG Beta Code: one for efficiently inserting an ASCII string into text when not in Latin mode, and the other for inserting a Unicode code point in any language mode. These are described in the following two sub-sections.
An ASCII string can be enclosed in curly brackets when in a non-Latin language mode. The string will be output verbatim. This can be useful if a Greek text uses ASCII symbols, in order to produce a Greek document with an ASCII colon (':') rather than a Unicode "GREEK ANO TELEIA" character, U+0387. The format is as follows:
This is an extension to standard Beta Code; the TLG specification assigns a different use to '{'. By itself, the use of '{' not followed by a decimal number is deprecated in the TLG specification, so this should avoid some conflict. Curly brackets are not allowed in the string.
The original Beta Code specification lists many numeric codes for producing special symbols that today have become part of The Unicode Standard. In the future, it will likely be most beneficial if any specialized numeric codes for characters use Unicode code points rather than the historical TLG Beta Code numeric assignments.
The unibetacode package allows any Unicode code point to be specified in hexadecimal (which is how The Unicode Standard provides them) as a string inside an ASCII '{'...'}' escape sequence. The Unicode hexadecimal code point of one to six digits is preceded by "\u", taking the form "\ux...x". For example, strings such as
can be used to insert the Unicode character U+03D8, GREEK LETTER ARCHAIC KOPPA, which does not have an associated Beta Code letter assignment. Such Unicode code point strings can be mixed with other characters in the same string, as long as any character that follows the Unicode code point is not a hexadecimal digit of '0'-'9', 'A'-'F', or 'a'-'f'.
The numeric digits '0' through '9' are simply entered as '0' through '9', respectively, in any language mode. The language modes are Coptic, Greek, Hebrew, and Latin. They are described in the following sections.
Capital and small letters can take the same set of accent marks, but the order in which these are specified differs between capital and small.
Small letters are given in Beta Code in this order: (1) letter, (2) breathing marks, (3) accents, and (4) iota subscript. This follows the traditional typed appearance of small polytonic Greek letters, where breathing marks and then accent marks appear on top of the small letter, and iota subscripts appear below small long vowels.
Capital letters are given in Beta Code in this order: (1) asterisk (which denotes a Capital letter), (2) breathing marks, (3) accents, (4) letter, and (5) iota subscript. This follows the traditional typed appearance of capital polytonic Greek letters, where breathing marks and then accent marks appear to the left of the capital letter, and iota subscripts appear to the right of capital long vowels.
The letter mapping is as follows, in Greek alphabetical order. Letters can be capital or small; generally speaking, small is easier to read, so it is the default output from uni2beta(1):
Example: "*to fws", "the light" (without accent marks). This could also be written as "*TO FWS"; both capital and small letters give the same conversion into UTF-8.
These are the encodings of breathing marks and accents. In Beta Code (as in written Greek), breathing marks appear before accents.
Example: "*to\ fw=s", "the light", with a grave accent, or varia, over the omicron and a circumflex accent, or perispomeni, over the omega. This could also be written as "*TO\ FW=S". N.B.: Note that the case of the Latin letter does not matter for accent placement; it is only the case of the Greek letter that matters. Greek capital letters are encoded with a preceding asterisk, so in this example, "O\" and "W=" will appear as small UTF-8.
The iota subscript is the last character written after a long vowel with which it appears, whether the letter is capital or small. It is denoted by a vertical bar:
These are the punctuation symbols that the unibetacode package supports:
The Greek Extended range of The Unicode Standard, U+1F00 - U+1FFF, contains 16 small and capital vowels that have identical representation in the Greek and Coptic range, U+0370 - U+03FF. These are vowels with an "oxia" (acute) accent in the Greek Extended range; they have equivalent glyphs with a "tonos" (acute) accent in the Greek and Coptic range. Because of this duplication, the use of these 16 Greek Extended glyphs is deprecated. uni2beta(1) will convert those 16 characters to Beta Code, but beta2uni(1) will convert the resulting Beta Code into characters in the Greek and Coptic range (U+0370 - U+03FF); it will not convert them back into Greek Extended glyphs.
Also in the Greek Extended Unicode range, the TLG Project considers U+1FBF to be the equivalent of a smooth breathing mark, and uni2beta(1) will convert it as such.
To display ASCII characters, including the Latin letters 'A' through 'Z' and 'a' through 'z', begin with an ampersand ('&') character. Switch back to Greek mode with a dollar sign ('$') character.
ASCII characters can also be surrounded with curly brackets; for example, "{Here is some ASCII!}". This is non-standard though; the TLG specification uses '&' and '$' to enter Latin and then switch back to Greek.
For efficiency, beta2uni(1) is conditioned to interpret sequences that look like accented Greek as accnted Greek. Curly brackets can also be useful for overriding such interpretations. For example, if a document contained the text
The "e)" could be interpreted as a small epsilon with a smooth breathing mark above it. To break this behavior, type
and the Unicode output from beta2uni(1) will appear as intended. This technique will appear familiar to TeX users.
To display Coptic letters, begin with the character sequence "&100". Switch back to Greek mode with a dollar sign ('$') character. As with Greek Beta Code, capital Coptic letters in Beta Code begin with an asterisk ('*') and small Coptic letters do not.
Note that unlike in Greek mode, the Coptic Beta Code letters are case-sensitive. In general, Coptic letters derived from Demotic use lowercase Beta Codes and map to the Greek and Coptic Unicode script in the range U+03E2 - U+03EF; the rest of the Coptic letters use uppercase Beta Codes and map to the separate Coptic Unicode script in the range U+2C80 - U+2C8D.
The encoding is as follows:
Switch back to Greek mode by ending with a dollar sign ('$') character.
The TLG specification only covers the basic Hebrew letters aleph through tav. These letters map to the Hebrew Unicode script in the range U+05D0 - U+05EA. Beta Codes are not defined in the specification for cantillation marks, Yiddish digraphs, etc.
To display Hebrew letters, begin with the character sequence "&300".
Note that unlike in Greek mode, the Hebrew Beta Codes are case-sensitive and they never begin with an asterisk ('*').
The encoding is as follows:
Switch back to Greek mode with a dollar sign ('$') character.
The directory test/reference contains samples with mappings from Beta Code to UTF-8 and vice versa. The "genesis-1-1.beta" and "genesis-1-1.utf8" files show the verse Genesis 1:1 in Koine Greek (from the Septuagint), Hebrew, and Bohairic Coptic in Beta Code and UTF-8, respectively.
The unibetacode package was created by Paul Hardy.
unibetacode is Copyright © 2018, 2019 Paul Hardy.
This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2 of the License, or (at your option) any later version.
The format is very straightforward and no known bugs exist. However, Beta Code has been evolving for almost 50 years, especially since the advent of Unicode. As a result, many Beta Code-encoded documents exist in versions of the standard much older than the current version. This version also does not implement many numbered codes that are contained in the TLG Beta Code specification. There are no plans to support the TLG Beta Code formatting codes, as that is beyond the scope of Unicode.
| 2019 Jan 26 |