

Version  1.0 
Authors  Barbara Beeton (bnb@ams.org), Asmus Freytag (asmus@unicode.org), Murray Sargent III (murrays@microsoft.com) 
Date  200213 
This Version  http://www.unicode.org/unicode/reports/tr25/tr253.html 
Previous Version  http://www.unicode.org/unicode/reports/tr25/tr252.html 
Latest Version  http://www.unicode.org/unicode/reports/tr25 
Tracking Number  3 
Summary
Starting with version 3.2, Unicode includes virtually all of the standard characters used in mathematics. This set supports a variety of math applications on computers, including document presentation languages like TeX, math markup languages like MathML, computer algebra languages like OpenMath, internal representations of mathematics in systems like Mathematica and MathCAD, computer programs, and plain text. This technical report describes the Unicode mathematics character groups and gives some of their default math properties.
Status
This document has been approved by the Unicode Technical Committee for public review as a Proposed Draft Unicode Technical Report. Publication does not imply endorsement by the Unicode Consortium. This is a draft document which may be updated, replaced, or superseded by other documents at any time. This is not a stable document; it is inappropriate to cite this document as other than a work in progress.
Please send comments to the authors. A list of current Unicode Technical Reports is found on http://www.unicode.org/unicode/reports/. For more information about versions of the Unicode Standard, see http://www.unicode.org/unicode/standard/versions/.
The References provide related information that is useful in understanding this document. Please mail corrigenda and other comments to the author(s).
This technical report starts with a discussion of the mathematics character repertoire incorporating the relevant block descriptions of the Unicode Standard [TUS]. Associated character properties are discussed next, including a number of properties that are not yet part of the Unicode Standard. Character classifications by usage, by typography, and by precedence are given. Some implementation guidelines for input methods and use of Unicode math characters in programming languages are presented next. The final section describes how many mathematical expressions can be rendered using a plain—or at least nearly plain—text format. Mathematical plain text can be handy for downlevel text copies, e.g., in email, math input methods, computer programs, and inline math display. Most mathematical expressions up through calculus can be represented unambiguously in Unicode plain text. Note that the discussion is only intended to show how mathematical plain text might be useful. It is not intended to be a complete specification or to be used for general information interchange at this stage in its development.
Unicode 3.2 provides a quite complete set of standard math characters to support publication of mathematics on and off the web. Specifically, Unicode 3.1 introduced 996 new alphanumeric symbols and Unicode 3.2 introduces 591 new symbols, in addition to the 340 mathspecific symbols already encoded in Unicode 3.0, for a total of 1927 mathematical symbols. This repertoire is the result of input from many sources, notably from the STIX Project (Scientific and Technical Information Exchange) [STIX], a cooperation of mathematical publishers. The STIX collection includes, but is not limited to, symbols gleaned from mathematical publications by experts from the American Mathematical Society (AMS) and symbol sets provided by Elsevier Publishing and by the American Physical Society. The new repertoire enables the display of virtually all standard mathematical symbols. Nevertheless this work must remain incomplete; mathematicians and other scientists are continually inventing new mathematical symbols and the plan is to add them as they become accepted in the scientific communities.
Mathematical Markup Language (MathML™) [MathML], an XML application [XML] , is a major beneficiary of the increased repertoire for mathematical symbols and the working group lobbied in favor of the inclusion of the new characters. In addition, the new characters lend themselves to a useful plain text encoding of mathematics (see Sec. 4) that is much more compact than MathML or T_{E}X, the typesetting language and program designed by Donald Knuth [TeX].
The Mathematical Alphanumeric Symbols block (U+1D400 – U+1D7FF) contains a large extension of letterlike symbols used in mathematical notation, typically for variables. The characters in this block are intended for use only in mathematical or technical notation; they are not intended for use in nontechnical text. When used with markup languages, for example with MathML the characters are expected to be used directly, instead of indirectly via entity references or by composing them from base letters and style markup.
Words Used as Variables. In some specialties, whole words are used as variables, not just single letters. For these cases, style markup is preferred because in ordinary mathematical notation the juxtaposition of variables generally implies multiplication, not word formation as in ordinary text. Markup not only provides the necessary scoping in these cases, it also allows the use of a more extended alphabet.
Basic Set of Alphanumeric Characters. Mathematical notation uses a basic set of mathematical alphanumeric characters which consists of:
Only unaccented forms of the letters are used for mathematical notation, because general accents such as the acute accent would interfere with common mathematical diacritics. Examples of common mathematical diacritics that can interfere with general accents are the circumflex, macron, or the single or double dot above, the latter two of which are used in physics to denote derivatives with respect to the time variable. Mathematical symbols with diacritics are always represented by combining character sequences, except as required by normalization. See Unicode Standard Annex #15, "Unicode Normalization Forms" [Normalization] for more information.
For some characters in the basic set of Greek characters, two variants of the same character are included. This is because they can appear in the same mathematical document with different meanings, even though they would have the same meaning in Greek text.
Additional Characters. In addition to this basic set, mathematical notation also uses the four Hebrewderived characters (U+2135 – U+2138). Occasional uses of other alphabetic and numeric characters are known. Examples include U+0428 cyrillic capital letter sha, U+306E hiragana letter no , and Eastern ArabicIndic digits (U+06F0 – U+06F9). However, these characters are used in only the basic form.
Semantic Distinctions. Mathematics has need for a number of Latin and Greek alphabets that on first thought appear to be mere font variations of one another. For example the letter H can appear as plain or upright (H), bold (H), italic (H), and script H. However in any given document, these characters have distinct, and usually unrelated mathematical semantics. For example, a normal H represents a different variable from a bold H, etc. If these attributes are dropped in plain text, the distinctions are lost and the meaning of the text is altered. Without the distinctions, the wellknown Hamiltonian formula:
,
turns into the integral equation in the variable H:
By encoding a separate set of alphabets, it is possible to preserve such distinctions in plain text.
Mathematical Alphabets. The alphanumeric symbols encountered in mathematics are given in the following table:
Math Style 
Characters from Basic Set 
Location 
plain (upright, serifed) 
Latin, Greek and digits 
BMP 
bold 
Latin, Greek and digits 
Plane 1 
italic 
Latin and Greek 
Plane 1* 
bold italic 
Latin and Greek 
Plane 1 
script (calligraphic) 
Latin 
Plane 1* 
bold script (calligraphic) 
Latin 
Plane 1 
Fraktur 
Latin 
Plane 1* 
bold Fraktur 
Latin 
Plane 1 
doublestruck 
Latin and digits 
Plane 1* 
sansserif 
Latin and digits 
Plane 1 
sansserif bold 
Latin, Greek and digits 
Plane 1 
sansserif italic 
Latin 
Plane 1 
sansserif bold italic 
Latin and Greek 
Plane 1 
monospace 
Latin and digits 
Plane 1 
* Some of these alphabets have characters in the BMP as noted in the following section.
The plain letters have been unified with the existing characters in the Basic Latin and Greek blocks. There are 25 doublestruck, italic, Fraktur and script characters that already exist in the Letterlike Symbols block (U+2100 – U+214F). These are explicitly unified with the characters in this block and corresponding holes have been left in the mathematical alphabets.
Compatibility Decompositions. All mathematical alphanumeric symbols have compatibility decompositions to the base Latin and Greek letters—folding away such distinctions, however, is usually not desirable as it loses the semantic distinctions for which these characters were encoded. See Unicode Standard Annex #15, "Unicode Normalization Forms" [Normalization] for more information.
Mathematicians place strict requirements on the specific fonts being used to represent mathematical variables. Readers of a mathematical text need to be able to distinguish single letter variables from each other, even when they do not appear in close proximity. They must be able to recognize the letter itself, whether it is part of the text or is a mathematical variable, and lastly which mathematical alphabet it is from.
Fraktur. The black letter style is often referred to as Fraktur or Gothic in various sources. Technically, Fraktur and Gothic typefaces are distinct designs from black letter, but any of several font styles similar in appearance to the forms shown in the charts can be used.
Math italics. Mathematical variables are most commonly set in a form of italics, but not all italic fonts can be used successfully. In common text fonts, the italic letter v and Greek letter nu are not very distinct. A rounded italic letter v is therefore preferred in a mathematical font. There are other characters, which sometimes have similar shapes and require special attention to avoid ambiguity. Examples are shown in the table below.
Theorems are commonly printed in a text italic font. A font intended for mathematical variables should support clear visual distinctions so that variables can be reliably separated from italic text in a theorem. Some languages have common single letter words (English a, Scandinavian i, etc.), which can otherwise be easily confused with common variables.
Hardtodistinguish Letters. Not all sansserif fonts allow an easy distinction between lowercase l, and uppercase I and not all monospaced (fixed width) fonts allow a distinction between the letter l and the digit 1. Such fonts are not usable for mathematics. In Fraktur, the letters I and J in particular must be made distinguishable. Overburdened Black Letter forms like I and J are inappropriate. Similarly, the digit zero must be distinct from the uppercase letter O for all mathematical alphanumeric sets. Some characters are so similar that even mathematical fonts do not attempt to provide distinguished glyphs for them, e.g. uppercase A and uppercase Alpha (A). Their use is normally avoided in mathematical notation unless no confusion is possible in a given context.
Font Support for Combining Diacritics. Mathematical equations require that characters be combined with diacritics (dots, tilde, circumflex, or arrows above are common), as well as followed or preceded by super or subscripted letters or numbers. This requirement leads to designs for italic styles that are less inclined, and script styles that have smaller overhangs and less slant than equivalent styles commonly used for text such as wedding invitations.
Typestyle for Script Characters. In some instances, a deliberate unification with a nonmathematical symbol has been undertaken; for example, U+2133 is unified with the pre1949 symbol for the German currency unit Mark and U+2113 is unified with the common nonSI symbol for the liter [SI]. This unification restricts the range of glyphs that can be used for this character in the charts. Therefore the font used for the reference glyphs in the code charts uses a simplified ‘English Script’ style, as per recommendation by the American Mathematical Society. For consistency, other script characters in the Letterlike Symbols block are now shown in the same typestyle.
Doublestruck Characters. The doublestruck glyphs shown in earlier editions of the standard attempted to match the design used for all the other Latin characters in the standard, which is based on Times. The current set of fonts was prepared in consultation with the American Mathematical Society and leading mathematical publishers, and shows much simpler forms that are derived from the forms written on a blackboard. However, both serifed and nonserifed forms can be used in mathematical texts, and inline fonts are found in works published by certain publishers. There is no intention to support such stylistic preference via character encoding, therefore only one set of double struck mathematical alphanumeric symbols have been encoded.
With Unicode 3.0 and the concurrent second edition of ISO/IEC 106461, the reference glyphs for U+03C6 GREEK LETTER SMALL PHI and U+03D5 GREEK PHI SYMBOL were swapped. In ordinary Greek text, the character U+03C6 is used exclusively, although this characters has considerably glyphic variation, sometimes represented with a glyph more like the representative glyph shown for U+03C6 (the "loopy" form) and less often with a glyph more like the representative glyph shown for U+03D5 (the “straight“ form).
For mathematical and technical use, the straight form of the small phi is an important symbol and needs to be consistently distinguishable from the loopy form. The straight form phi glyph is used as the representative glyph for the symbol phi at U+03D5 to satisfy this distinction.
The reversed assignment of representative glyphs in versions of the Unicode Standard prior to Unicode 3.0 had the problem that the character explicitly identified as the mathematical symbol did not have the straight form of the character that is the preferred glyph for that use. Furthermore, it made it unnecessarily difficult for general purpose fonts supporting ordinary Greek text to also add support for Greek letters used as mathematical symbols. This resulted from the fact that many of those fonts already used the loopy form glyph for U+03C6, as preferred for Greek body text; to support the phi symbol as well, they would have had to disrupt glyph choices already optimized for Greek text.
When mapping symbol sets or SGML entities to the Unicode Standard, it is important to make sure that codes or entities that require the straight form of the phi symbol be mapped to U+03D5 and not to U+03C6. Mapping to the latter should be reserved for codes or entities that represent the small phi as used in ordinary Greek text.
Fonts used primarily for Greek text may use either glyph form for U+03C6, but fonts that also intend to support technical use of the Greek letters should use the loopy form to ensure appropriate contrast with the straight form used for U+03D5.
Mathematical characters can be located by looking in the blocks that contain such characters or by checking the Unicode MATH property, which is assigned to characters that naturally appear in mathematical contexts (see Section 3 "Mathematical Character Properties"). Mathematical characters can be found in the following blocks:
Block Name 
Range 
Characters 
Basic Latin 
U+0021–U+007E 
Variables, operators, digits* 
Greek 
U+0370–U+03FF 
Variables* 
General Punctuation 
U+2000–U+206F 
Invisible operators* 
Letterlike Symbols 
U+2100–U+214F 
Variables* 
Arrows 
U+2190–U+21FF 
Arrows, arrowlike operators 
Mathematical Operators 
U+2200–U+22FF 
Operators 
Miscellaneous Technical Symbols 
U+2300–U+23FF 
Braces, operators* 
Geometrical Shapes 
U+25A0–U+25FF 
Symbols 
Misc. Mathematical SymbolsA 
U+27C0–U+27EF 
Symbols and operators 
Supplemental ArrowsA 
U+27F0–U+27FF 
Arrows, arrowlike operators 
Supplemental ArrowsB 
U+2900–U+297F 
Arrows, arrowlike operators 
Misc. Mathematical SymbolsB 
U+2980–U+29FF 
Braces, symbols 
Suppl. Mathematical Operators 
U+2A00–U+2AFF 
Operators 
Mathematical Alphanumeric Symbols 
U+1D400–U+1D7FF 
Variables and digits 
Other blocks 
… 
Characters for occasional use 
*This block contains nonmathematical characters as well.
Some Greek letters are reencoded as technical symbols. These include U+00B5 µ micro sign, U+2126 Ω ohm sign, and several characters among the APL functional symbols in the Miscellaneous Technical block. U+03A9 greek letter capital omega is the canonical equivalent of U+2126 and its use is preferred. Latin letters duplicated include 212A kelvin sign and U+212B angstrom sign. As in the case of the ohm sign, the corresponding regular Latin letters are the canonical equivalents and therefore their use is preferred.
The left and right angle brackets at U+2328 and U+2329 have long been canonically equivalent with the CJK punctuation characters at U+3008 and U+3009, which implies that the use of the latter code points is preferred and that the characters are ‘wide’ characters. See Unicode Standard Annex #11, "East Asian Width" [EAW]. Unicode 3.2 adds two new mathematical angle bracket characters (U+27E8 and U+27E9) that are unequivocally intended for mathematical use.
Mathematical characters are often enhanced via use of combining marks in the ranges U+0300 – U+036F and the combining marks for symbols in the range U+20D0 – U+20FF. These characters follow the base characters as in nonmathematical Unicode text. This section discusses these characters and preferred ways of representing accented characters in mathematical expressions. If a span of characters is enhanced by a combining mark, e.g., a tilde over AB, typically some kind of higherlevel markup is needed as is done in MathML. Unicode does include some combining marks that are designed to be used for pairs of characters, e.g., U+0360 – U+0362. However, their use for mathematical text is not encouraged.
For some mathematical characters there are multiple ways of expressing the character: as precomposed or as a sequence of base character and combining mark. It would be nice to have a single way to represent any given character, since this would simplify recognizing the character in searches and other manipulations. Selecting a unique representation among multiple equivalent representations is called normalization. Unicode Standard Annex #15 "Unicode Normalization Forms" [Normalization] discusses the subject in detail; however, due to requirements of nonmathematical software, the normalization forms presented there are not ideal from the perspective of mathematics.
Ideally, one always uses the shortest form of a math operator symbol wherever possible. So U+2260 should be used for the not equal sign instead of the combining sequence U+003D U+0338. This rule concurs with Normalization Form C (NFC) used on the web. If a negated operator is needed that does not have a precomposed form, the character U+0338 combining long solidus overlay can be used to indicate negation.
On the other hand, for accented alphabetic characters used as variables, ideally only decomposed sequences are used since there are no precomposed math alphanumerical symbols.
Mathematics uses a multitude of combining marks that greatly exceeds the predefined composed characters in Unicode. Accordingly, it is better to have the math display facility handle all of these cases uniformly to give a consistent look between characters that happen to have a fully composed Unicode character and those that do not. The combining character sequences also typically have semantics as a group, so it is handy to be able to manipulate and search for them individually without having to have special tables to decompose characters for this purpose. Note that this approach does not concur with Normalization Form C for the upright alphabetic characters (ASCII letters). To facilitate interchange on the web, accented characters should conform to NFC when interchanged.
However, to achieve consistent results, a mathematical display system should transiently decompose such letters when used in mathematical expressions and use a single algorithm to place embellishments.
The Unicode blocks U+2200 – U+22FF and U+2A00 – U+2AFF contain many mathematical operators, relations, geometric symbols and other symbols with special usages confined largely to mathematical contexts. In addition to the characters in these blocks, mathematical operators are also found in the Basic Latin (ASCII) and Latin1 Supplement Blocks. A few of the symbols from the Miscellaneous Technical block and characters from General Punctuation are also used in mathematical notation.
Semantics. Mathematical operators often have more than one meaning different subdisciplines or different contexts. For example, the "+" symbol normally denotes addition in a mathematical context, but might refer to concatenation in a computer science context dealing with strings, or incrementation, or have any number of other functions in given contexts. Therefore The Unicode Standard only encodes a single character for a single symbolic form. There are numerous other instances in which several semantic values can be attributed to the same Unicode value. For example, U+2218 ring operator may be the equivalent of white small circle or composite function or apl jot. The Unicode Standard does not attempt to distinguish all possible semantic values that may be applied to mathematical operators or relational symbols. It is up to the application or user to distinguish such meanings according to the appropriate context. Where information is available about the usage (or usages) of particular symbols, it has been indicated in the character annotations in Chapter 14, Code Charts in The Unicode Standard, Version 3.0 [TUS] and in the online code charts [Charts].
Similar glyphs. The Standard does include many characters that appear to be quite similar to one another, but that may well convey different meaning in a given context. On the other hand, mathematical operators, and especially relation symbols, may appear in various standards, handbooks, and fonts with a large number of purely graphical variants. Where variants were recognizable as such from the sources, they were not encoded separately.
For relation symbols, the choice of a vertical or forwardslanting stroke typically seems to be an aesthetic one, but both slants might appear in a given context. However, a backslanted stroke almost always has a distinct meaning compared to the forwardslanted stroke. See Section 2.16 "Variation Selector" for more information on some particular variants.
Unifications. Mathematical operators such as implies⇔ and if and only if ↔ have been unified with the corresponding arrows (U+21D2 rightwards double arrow and U+2194 left right arrow, respectively) in the Arrows block.
The operator U+2208 element of is occasionally rendered with a taller shape than shown in the code charts. Mathematical handbooks and standards consulted treat these characters as variants of the same glyph. U+220A small element of is a distinctively small version of the element of that originates in mathematical pi fonts.
The operators U+226B much greaterthan and U+226A much lessthan are sometimes rendered in a nested shape. Because no semantic distinction applies, the Unicode Standard provides a single encoding for each operator.
A large class of unifications applies to variants of relation symbols involving equality, similarity, and/or negation. Variants involving one or twobarred equal signs, one or twotilde similarity signs, and vertical or slanted negation slashes and negation slashes of different lengths are not separately encoded. Thus, for example, U+2288 neither a subset of nor equal to, is the archetype for at least six different glyph variants noted in various collections.
In two instances, essentially stylistic variants are separately encoded: U+2265 greaterthan or equal to is distinguished from U+2267 greaterthan over equal to; the same distinction applies to U+2264 lessthan or equal to and U+2266 lessthan over equal to. This exception to the general rule regarding variation results from requirements for character mapping to some Asian standards that distinguish the two forms.
Several mathematical operators derived from Greek characters have been given separate encodings since they are used differently than the corresponding letters. These operators may occasionally occur in context with Greekletter variables. They include U+2206 increment, U+220F nary product, and U+2211 nary summation. The latter two are large operators that take limits. Some typographical aspects of operators are discussed in Section 3.2 "Classification by Typographical Behavior". For example, the nary operators are distinguished from letter variables by their larger size and the fact that they take limit expressions.
The unary and binary minus sign is preferably represented by U+2212 minus sign rather than by the ASCIIderived U+002D hyphenminus, both because the former is unambiguous and because it is rendered with a more desirable length. (For a complete list of dashes in the Unicode Standard, see Table 62 in [TUS]).
Miscellaneous Symbols. U+22EE – U+22F1 are a set of ellipses used in matrix notation.
The Unicode block U+2070 – U+209F plus U+00B2, U+00B3, and U+00B9 contain sequences of superscript and subscript digits and punctuation that can be useful in mathematics. If they are used, it is recommended that they be displayed with the same font size as other subscripts and superscripts at the corresponding nested script level. For example, a² and a<super>2</super> should be displayed the same. However, these subscript/superscript characters are not used in MathML or T_{E}X and their use with XML documents is discouraged, see Unicode Technical Report #20, "Unicode in XML and other Markup Languages" [UXML].
Arrows are used for a variety of purposes in mathematics and elsewhere, such as to imply directional relation, to show logical derivation or implication, and to represent the cursor control keys. Accordingly Unicode includes a fairly extensive set of arrows (U+2190 – U+21FF and U+2900 – U+297F), many of which appear in mathematics. It does not attempt to encode every possible stylistic variant of arrows separately, especially where their use is mainly decorative. For most arrow variants, the Unicode Standard provides encodings in the two horizontal directions, often in the four cardinal directions. For the single and double arrows, the Unicode Standard provides encodings in eight directions.
Unifications. Arrows expressing mathematical relations have been encoded in the arrows block as well as in Supplemental ArrowsA and Supplemental ArrowsB. An example is U+21D2 rightwards double arrow, which may be used to denote implies. Where available, such usage information is indicated in the annotations to individual characters in the Unicode Standard, Chapter 14, Code Charts.
Long Arrows. The long arrows encoded in the range U+27F5..U+27FF map to standard SGML entity sets supported by MathML. Long arrows represent distinct semantics from their short counterparts, rather than mere stylistic glyph differences. For example, the shorter forms of arrows are often used in connection with limits, whereas the longer ones are associated with mappings. The use of the long arrows is so common that they were assigned entity names in the ISOAMSA entity set, one of the suite of mathematical symbol entity sets covered by the Unicode Standard.
The basic geometric shapes (circle, square, triangle, diamond, and lozenge) are used for a variety of purposes in mathematical texts. Because their shapes are distinct and they are easily available in multiple sizes from a variety of widely available fonts, they are also often used in an adhoc manner.
Ideal sizes. Mathematical usage requires at least four distinct sizes of simple shapes, and sometimes more. The size gradation must allow each size to be recognized, even when it occurs in isolation. In other words shapes of the same size should ideally have roughly the same visual "impact" as opposed to same nominal height or width. For mathematical usage simple shapes ideally share a common center. The following diagram shows which size relationship across shapes of the same nominal size is considered ideal.
Please note that neither the current set of glyphs in the standard nor the glyphs from many commonly available nonmathematical fonts show this kind of size relation.
Actual sizes. The sizes of existing characters and their names are not always consistent. For mathematical usage, therefore, the MEDIUM SMALL SQUARE should be used together with the MEDIUM size of the other basic shapes, and correspondingly for the other sizes. (The basic shapes from the Zapf Dingbats font match the unmarked size for triangle, diamond and circle and the MEDIUM size for the square.) To achieve the correct size relation, mathematical fonts may need to deviate in minor amounts from the sizes shown in the character charts. [ED: TBD: summary picture]
Sizes of derived shapes. Circled and squared operators and similar derived shapes are more constrained in their usage than "plain" geometric shapes. They tend to occur in two generic sizes based on function: a smaller size for operators and large size for nary operators.
Positioning. For a mathematical font, the centerline should go through the middle of a parenthesis, which should go from bottom of descender to top of ascender. This is the same level as the minus or the middle of the plus and equal signs. For correct positioning, the glyph will descend below the baseline for the larger sizes of the basic shapes as in the following schematic diagram:
The standard triangles used for mathematics are also center aligned. This is different from the positioning for the reference glyphs of existing characters shown in the charts. Mathematical fonts may need to deviate in positioning of these triangles.
Other symbols of use in mathematics are contained in the Miscellaneous Technical block (U+2300 – U+23FF), the Geometric Shapes block (U+25A0 – U+25FF), the Miscellaneous Symbols block (U+2600 – U+267F), and the General Punctuation block (U+2000 – U+206F).
Generally any easily recognized and distinct symbol is fair game for mathematicians faced with the need of creating notations for new fields of mathematics. For example, the card suits, ♥, ♠, etc. can be found as operators as well as subscripts.
The characters from the Miscellaneous Technical block in the range U+239B – U+23B3, plus U+23B7, comprise a set of bracket and other symbol fragments for use in mathematical typesetting. These pieces originated in older font standards, but have been used in past mathematical processing as characters in their own right to make up extratall glyphs for enclosing multiline mathematical formulae. Mathematical fences are ordinarily sized to the content that they enclose. However, in creating a large fence, the glyph is not scaled proportionally; in particular the displayed stem weights must remain compatible with the accompanying smaller characters. Thus, simple scaling of font outlines cannot be used to create tall brackets. Instead, a common technique is to build up the symbol from pieces. In particular, the characters U+239B LEFT PARENTHESIS UPPER HOOK through U+23B3 SUMMATION BOTTOM represent a set of glyph pieces for building up large versions of the fences (, ), [, ], {, and }, and of the large operators ∑ and ∫. These brace and operator pieces are compatibility characters. They should not be used in stored mathematical text, but are often used in the data stream created by display and print drivers.
The following table shows which pieces are intended to be used together to create specific symbols.

2row 
3row 
5row 
Summation 
23B2, 23B3 


Integral 
2320, 2321 
2320, 23AE, 2321 
2320, 3×23AE, 2321 
Left Parenthesis 
239B, 239D 
239B, 239D 
239B, 3×239C, 239D 
Right Parenthesis 
239E, 23A0 
239E, 239F, 23A0 
239E, 3×239F, 23A0 
Left Bracket 
23A1, 23A3 
23A1, 23A2, 23A4 
23A1, 3×23A2, 23A3 
Right Bracket 
23A4, 23A6 
23A4, 23A5, 23A6 
23A4, 3×23A5, 23A6 
Left Brace 
23B0, 23B1 
23A7, 23A8, 2389 
23A7, 23AA, 23A8, 23AA, 2389 
Right Brace 
23B1, 23B0 
23AB, 23AC, 23AD 
23AB, 23AA, 23AC, 23AA, 23AD 
For example, an instance of U+239B can be positioned relative to instances of U+239C and U+239D to form an extratall (three or more line) leftparenthesis. The center sections encoded here are meant to be used only with the top and bottom pieces encoded adjacent to them, since the segments are usually graphically constructed within the fonts so that they match perfectly when positioned at the same x coordinates.
In mathematics some operators or punctuation are often implied, but not displayed. U+2063 invisible separator or invisible comma is intended for use in index expressions and other mathematical notation where two adjacent variables form a list and are not implicitly multiplied. In mathematical notation, commas are not always explicitly present, but need to be indicated for symbolic calculation software to help it disambiguate a sequence from a multiplication. For example, the double _{ij} subscript in the variable a_{ij} means a_{i, j} — that is, the i and j are separate indices and not a single variable with the name ij or even the product of i and j. Accordingly to represent the implied list separation in the subscript _{ij} one can insert a nondisplaying invisible separator between the i and the j. In addition, use of the invisible comma would hint to a math layout program to typeset a small space between the variables.
Similarly an expression like mc^{2} implies that the mass m multiplies the square of the speed c. To represent the implied multiplication in mc^{2}, one inserts a nondisplaying U+2061invisible times between the m and the c. A related case is the use of U+2062 function application for an implied function dependence as in f(x + y). To indicate that this is the function f of the quantity x + y and not the expression fx + fy, one can insert the nondisplaying function application symbol between the f and the left parenthesis.
Another example is the expression f^{i j}(cos(ab)), which means the same as f^{i,j}(cos(a×b)), where × represents multiplication, not the cross product. Note that the spacing between characters may also depend on whether the adjacent variables are part of a list or are to be concatenated, that is, multiplied.
These include all remaining Unicode characters. They may appear in mathematical expressions, typically in spelledout names for variables in fractions or simple formulae, but they most commonly appear in ordinary text. An English example is the equation
distance = rate × time,
which uses ordinary ASCII letters to aid in recognizing sequences of letters as words instead of products of individual symbols. Such usage corresponds to identifiers, discussed elsewhere.
The variation selector VS1 is used to represent welldefined variants of particular math symbols. The variations include: different slope of cancellation element in some negated symbols, changed orientation of an equating or tilde operator element, and some welldefined different shapes. These mathematical variants are all produced with the addition of Variation Selector 1 (VS1 or U+FE00) to mathematical operator base characters. To select one of the predefined variations, follow the base character with the variation selector. Only the valid, recognized combinations are listed in the table of standardized variants. All combinations not listed here are unspecified and are reserved for future standardization; no conformant process may interpret them as standardized variants. For more information, see Section 13.7, Variation Selectors, in Unicode 3.2 [U3.2].
Using a variation selector allows users and font designers to make a distinction between alternate glyphs shapes both of which are ordinarily acceptable glyphs for generic, nondistinguishing usage of the standalone character code. This situation is somewhat analogous to the variants of Greek letterforms used as symbols. See Section 2.31, "Reference Glyphs for Greek phi".
It is important to further note that the variation selector only selects a different appearance of an already encoded character. It is not intended as a general code extension mechanism. At this time the variations encoded with the variation selector are thought to be primarily glyphic variations. Should their usage or interpretation change—over time, or because of better evidence about how these shapes are actually used in mathematical notation—it is likely that another character would be coded so that the distinction in meaning can be kept directly in the character code.
In extremis, the Unicode Standard considers the variation selector somewhat optional. Processes or fonts that cannot support it should yield acceptable results by ignoring the variation selector.
2268 + VS1 
LESSTHAN BUT NOT EQUAL TO  with vertical stroke 
2269 + VS1 
GREATERTHAN BUT NOT EQUAL TO  with vertical stroke 
22DA + VS1 
LESSTHAN slanted EQUAL TO OR GREATERTHAN 
22DB + VS1 
GREATERTHAN slanted EQUAL TO OR LESSTHAN 
2272 + VS1 
LESSTHAN OR EQUIVALENT TO  following the slant of the lower leg 
2273 + VS1 
GREATERTHAN OR EQUIVALENT TO  following the slant of the lower leg 
2A9D + VS1 
SIMILAR OR LESSTHAN  following the slant of the upper leg  or lessthan 
2A9E + VS1 
SIMILAR OR GREATERTHAN  following the slant of the upper leg  or greaterthan 
2AAC + VS1 
SMALLER THAN OR slanted EQUAL 
2AAD + VS1 
LARGER THAN OR slanted EQUAL 
228A + VS1 
SUBSET OF WITH NOT EQUAL TO  variant with stroke through bottom members 
228B + VS1 
SUPERSET OF WITH NOT EQUAL TO  variant with stroke through bottom members 
2ACB + VS1 
SUBSET OF ABOVE NOT EQUAL TO  variant with stroke through bottom members 
2ACC + VS1 
SUPERSET OF ABOVE NOT EQUAL TO  variant with stroke through bottom members 
2A3B + VS1 
INTERIOR PRODUCT  tall variant with narrow foot 
2A3C + VS1 
RIGHTHAND INTERIOR PRODUCT  tall variant with narrow foot 
2278 + VS1  NEITHER LESSTHAN NOR GREATERTHAN with vertical stroke (*) 
2279 + VS1  NEITHER GREATERTHAN NOR LESSTHAN with vertical stroke (*) 
2295 + VS1 
CIRCLED PLUS with white rim 
2297 + VS1 
CIRCLED TIMES with white rim 
229C + VS1 
CIRCLED EQUALS  equal sign inside and touching the circle 
2225 + VS1 
Slanted PARALLEL TO 
2225 + VS1 + 20E5 
Slanted PARALLEL TO with reverse slash 
2229 + VS1 
INTERSECTION with serifs 
222A + VS1 
UNION with serifs 
2293 + VS1 
SQUARE CAP with serifs 
2294 + VS1 
SQUARE CUP with serifs 
* The reference glyphs shown in the code charts [Charts] have been revised to show the slanted forms  this matches their existing decomposition using U+0338 combining long solidus overlay. Unless fonts intend to add support for both forms (via VS1 for the upright forms) they do not need to change.
Mathematicians are by their nature inventive people and will continue to invent new symbols to express their theories. Until these symbols are used by a number of people, they should not be standardized. Nevertheless, one needs a way to handle these novel symbols even before they are standardized.
The Private Use Area (U+E000 – U+F8FF) can be used for such nonstandard symbols. It is a tricky business, since the Private Use Area (PUA) is used for many purposes. Hence when using the PUA, it is a good idea to have higherlevel backup to define what kind of characters are involved. If they are used as math symbols, it would be good to assign them a math attribute that is maintained in a richtext layer parallel to the plain text.
Unicode assigns a number of mathematical character properties to aid in the default interpretation and rendering of these characters. Such properties include the classification of characters into operator, digit, delimiter, and variable. These properties may be overridden, or explicitly specified in some environments, such as MathML [MathML], which uses specific tags to indicate how Unicode characters are used, such as <mo> for operator, <md> for one or more digits comprising a number, and <mi> for identifier. TeX [TeX] is a higherlevel composition system that uses implicit character semantics. In the following, these properties are described in greater detail.
In particular, many Unicode characters nearly always appear in mathematical expressions and are given the generic mathematics property. For example, they include the math operators in the ranges U+2200 – U+22FF and U+29B0 – U+2AFF, the math combining marks U+20D0 – U+20FF, the math alphanumeric characters (some of the Letterlike Symbols and the mathematics alphanumerics range U+1D400 – U+1D7FF). Other characters may occur in mathematical usage depending on context. The math property is useful in heuristics that seek to identify mathematical expressions in plain text.
[ED: This classification is a work in progress.]
Strong mathematical characters are all characters that are primarily used for mathematical notation. This includes all characters with the math property [Sec. 4.9 of The Unicode Standard] [ED: Check that this is true after extension of the properties to the new characters.] with the following exceptions:
002D HYPHENMINUS
and the following additions [ED: any?]
These characters often appear in mathematical expressions, but they also appear naturally in ordinary text. They include the ASCII letters, punctuation, as well as the arrows and many of the geometric and technical shapes. The ASCII hyphen minus (U+002D) is a weakly mathematical character that may be used for the subtraction operator, but U+2212 is preferred for this purpose and looks better. Geometric shapes are frequently used as mathematical operators.
All other Unicode characters. Many of these may occur in mathematical texts, though often not as part of the mathematical expressions themselves.
Math characters fall into a number of subcategories, such as operators, digits, delimiters, and identifiers (constants and variables). This section discusses some of the typographical characteristics of these subcategories. These characteristics and classifications are useful in the absence of overriding information. For example, there is at least one document that uses the letter P as a relational operator.
In general italic Latin characters are used to represent singlecharacter Latin variables. In contrast, mathematical function names like sin, cos, tan, tanh, etc., are represented by upright serifed text to distinguish them from products of variables. Such names should not use the math alphanumeric characters. The upright uppercase Greek are favored over the italic ones. In Europe, upright d, D, e, and i are used for the two differential, exponential, and imaginary part functionalities, respectively. In common American mathematical practice, these quantities are represented by italic quantities. Products of italicized variables have slightly wider spacing than the letters in italicized words in ordinary text.
Operators fall into one or more categories. These include:
Category  Notes 
binary  some spacing around binary operators 
unary  closer to modified character than binary operators 
nary  often called "large" operators, take limits ordinarily above/below when displayed outofline and right top/bottom when displayed inline 
arithmetic  arithmetic includes binary and unary operators 
logical  unary not and binary and, or, exclusive or in a host of guises 
settheoretic  inclusion, exclusion, in a variety of guises 
relational  binary operators like less/greater than in many forms 
These include nary operators like summation and integration. These may expand in size to fit their associated expressions. They generally also take limits. The placement of the limits on an operator is different when it is used inline compared to its use in displayed formulae. For example _{ } versus _{ }.
Selection of a particular layout for limit expressions is outside the scope of the Unicode Standard.
Digits include 09 in various styles. All digits of a particular style have the same width.
Delimiters include punctuation, opening/closing delimiters such as parentheses and brackets, braces, and fences. Opening and closing delimiters and fences may expand in size to fit their associated expressions. Some bracket expressions do not appear to be "logical" to readers unfamiliar with the notation, e.g., ]x,y[.
Fences are similar to opening and closing delimiters, but are not paired. In addition, they include 'mid' delimiters, which are not opening or closing in character.
Combining marks are used with mathematical alphabetic characters (see Section 2.6 "Accented Characters"), instead of precomposed characters. Use U+0061 U+0308 for the second derivative of acceleration with respect to time, not the precomposed letter ä. On the other hand, precomposed characters are used for operators whenever they exist. Combining slash (solidus) or vertical overlays can be used to indicate negation for operators that do not have precomposed negated forms.
Where both long and short combining marks exist, use the long, e.g., use U+0338, not U+0337 and use U+20D2, not U+20D3. The actual shape or position of a combining mark is a typesetting problem and not specified in plain text. When using combining marks, the composite characters have the same typesetting class as the base character.
Operator precedence reduces the notational complexity of expressions and is commonly used for this purpose in computer programming languages, calculus, and algebra. A simple precedence table is used in Section 5.2 to convert the Unicode plaintext notation into a prefix notation used in twodimensional display code. Although that table has some unusual precedence assignments, it shares with ordinary algebra the concept that addition and subtraction have lower precedence than multiplication and division. Some display engines, e.g., TeX’s and MathML’s, do not use precedence and instead rely on complete specification of operator order via explicit bracketing, either with {} as in TeX or XML tags as in MathML.
[TBD: property files that specify the actual classification]
If Normalization Form C is applied to mathematical text, some accents or overlays used with BMP alphabetic characters may be incorrectly composed with their base character. Parsers should allow for this. Normalization forms KC or KD remove the distinction between different mathematical alphabets. These forms cannot be used with mathematical texts. For more details on Normalization see Unicode Standard Annex #15, "Unicode Normalization Forms" [Normalization] and the discussion in Section 2.6 "Accented Characters".
In view of the large number of characters used in mathematics, it is useful to give some discussion of input methods. The ASCII math symbols are easy to find, e.g., +  / * [ ] ( ) { }, but often need to be used as themselves.
Postentry correction. From a syntax point of view, the official Unicode minus sign (U+2212) is certainly preferable to the ASCII hyphenminus (U+002D) and the prime (U+2032) is preferable to the ASCII apostrophe (U+0027), but users may locate the ASCII characters more easily. Similarly it is easier to type ASCII letters than italic letters, but when used as mathematical variables, such letters are traditionally italicized in print. Accordingly a user might want to make italic the default alphabet in a math context, reserving the right to overrule this default when necessary. Other postentry enhancements include automaticligature and leftright quote substitutions, which can be done automatically by some word processors. Suffice it to say that intelligent input algorithms can dramatically simplify the entry of mathematical symbols.
Math keyboards. A special math shift facility for keyboard entry could bring up proper math symbols. The values chosen can be displayed on an onscreen keyboard. For example, the left Alt key could access the most common mathematical characters and Greek letters, the right Alt key could access italic characters plus a variety of arrows, and the right Ctrl key could access script characters and other mathematical symbols. The numeric keypad offers locations for a variety of symbols, such as sub/superscript digits using the left Alt key. Left Alt CapsLock could lock into the leftAlt symbol set, etc. This approach yields what one might call a "sticky" shift. Other possibilities involve the NumLock and ScrollLock keys in combinations with the left/right Ctrl/Alt keys. Pretty soon one realizes that this approach rapidly approaches literally billions of combinations, that is, several orders of magnitude more than Unicode can handle!
Macros. The autocorrect and keyboard macro features of some word processing systems provide other ways of entering mathematical characters for people familiar with TeX. For example, typing \alpha inserts α if the appropriate autocorrect entry is present. This approach is noticeably faster than using menus.
Hexadecimal input. A handy hextoUnicode entry method works with recent Microsoft text software (similar approaches are available on other systems) to insert Unicode characters in general and math characters in particular. Basically one types a character’s hexadecimal code (in ASCII), making corrections as need be, and then types Alt+x. The hexadecimal code is replaced by the corresponding Unicode character. The Alt+x can be a toggle, that is, type it once to convert a hex code to a character and type it again to convert the character back to a hex code. If the hex code is preceded by one or more hexadecimal digits, one needs to "select" the code so that the preceding hexadecimal characters aren’t included in the code. The code can range up to the value 0x10FFFF, which is the highest character in the 17 planes of Unicode.
Pulldown menus. Pulldown menus are a popular method for handling large character sets, but they are slow. A better approach is the symbol box, which is an array of symbols either chosen by the user or displaying the characters in a font. Symbols in symbol boxes can be dragged and dropped onto key combinations on the onscreen keyboard(s), or directly into applications. Onscreen keyboards and symbol boxes are valuable for entry of mathematical expressions and of Unicode text in general.
Unicode plaintext mathematics. One use of the plaintext format is as a math input method, both for search text and for general editing.
It can be very useful to have typical mathematical symbols available in computer programs (see Section 5.3 "Using PlainText Mathematics in Programming Languages" for a more detailed discussion). A key point is that the compiler should display the desired characters in both edit and debug windows. A preprocessor can translate MathML, for example, into C++, but it will not be able to make the debug windows use the mathoriented characters unless it can handle the underlying Unicode characters. Java has made an important step in this direction by allowing Unicode variable names. The mathematical alphanumeric symbols allow this approach to go further with relatively little effort for compilers.
The advantages of using the Unicode plain text in computer programs are at least threefold: 1) many formulas in document files can be programmed simply by copying them into a program file and inserting appropriate multiplication dots. This dramatically reduces coding time and errors. 2) The use of the same notation in programs and the associated journal articles and books leads to an unprecedented level of selfdocumentation. 3) In addition to providing useful tools for the present, these proposed initial steps should help one figure out how to accomplish the ultimate goal of teaching computers to understand and use arbitrary mathematical expressions.
[ED: This section may be moved to a separate document in the future.]
Getting computers to understand human languages is important in increasing the utility of computers. Naturallanguage translation, speech recognition and generation, and programming are typical ways in which such machine comprehension plays a role. The better this comprehension, the more useful the computer, and hence there has been considerable current effort devoted to these areas since the early 1960s.
Ironically one truly international human language that tends to be neglected in this connection is mathematics itself.
With a few conventions, Unicode can encode most mathematical expressions in readable (near)plain text. The format is linear, but can be displayed in builtup form. The approach uses heuristics based on the Unicode math properties to recognize mathematical expressions without the aid of explicit mathon/off commands. This is facilitated by Unicode’s strong support for mathematical symbols. This plaintext representation is substantially more compact and easy to read compared to the LaTeX [LaTeX] dialect of TeX, "Unicode TeX", or MathML. Most mathematical expressions up through calculus can be represented unambiguously in Unicode plain text. The plaintext encoding can be easily exported to (La)TeX, MathML, C++, and symbolic manipulation programs.
Note that this discussion is intended to illustrate how mathematical plain text might be useful, for example, in math input and computer programs. This is intended to be a preliminary draft of a specification of how to encode mathematical expressions in plain text. Readers are encouraged to comment on the proposed scheme.
Given the power of Unicode relative to ASCII, how much better can a plaintext encoding of mathematical expressions look using Unicode? The most wellknown plaintext ASCII encoding of such expressions is that of TeX, so one uses it for comparison. MathML is considerably more verbose than TeX, so some of the comparisons apply to it as well. Notwithstanding TeX’s phenomenal success in the science and engineering communities, a casual glance at its representation of mathematical expressions reveals that they do not look very much like the expressions they represent. It is certainly not easy to make algebraic calculations directly using TeX’s notation. With Unicode, one can represent mathematical expressions more readably, and the resulting plain text can be used directly for such calculations.
For example, one way to specify a TeX fraction numerator consists of the expression \frac{numerator}{denominator}. In both the fraction and subscript/superscript cases, the { } are not printed. These simple rules immediately give a "plain text" that is unambiguous, but looks quite different from the corresponding mathematical notation, thereby making it hard to read.
Instead, suppose one defines a simple operand to consist of all consecutive letters and digits, that is, a span of characters belong to the Lx and Nd General Categories. As such, a simple numerator or denominator is terminated by any operator, including, for example, arithmetic operators, the blank, all Unicode characters with codes U+22xx, and a special argument "break" operator consisting of a small raised dot. The fraction operator is given by the Unicode fraction slash operator U+2044, which is depicted here with the glyph /. So the simple builtup fraction
appears in plain text as abc /d.
For more complicated operands (such as those that include operators), parentheses ( ), brackets [ ], or braces { } can be used to enclose the desired character combinations. If parentheses are used and the outermost parenthesis set is preceded or followed by operators, that set is not displayed in builtup form, since usually one does not want to see such parentheses. So the plain text (a + c)/d displays as
.
In practice, this approach leads to plain text that is significantly easier to read than TeX’s, e.g., \frac{a + c}{d}, since in many cases, outermost parentheses are not needed, while TeX requires { }’s. To force the display of an outermost parenthesis set, one encloses the set, in turn, within parentheses, which then become the outermost set. For example, ((a + c))/d displays as
.
A really neat feature of this notation is that the plain text is, in fact, a legitimate mathematical notation in its own right, so it is relatively easy to read. In MathML, this fraction reads as
<mfrac> <mrow> <mi>a</mi> <mo>+</mo> <mi>c</mi> </mrow> <mrow> <mi>d</mi> </mrow> </mfrac>
Subscripts and superscripts are a bit trickier, but they are still quite readable. Specifically, one can introduce a subscript by a subscript operator, which is displayed as the subscripted down arrow _{↓}. A simple subscript operand consists of the string of one or more characters with the General Categories Lx (alphabetic) and Nd (decimal digits), as well as the invisible comma. For example, a pair of subscripts, such as δ_{μν }is written as γ_{↓}mn.Similarly, superscripts are introduced by a superscript operator, which is displayed here as the superscripted up arrow ^{↑}. So a^{↑}b means a^{b}.
Compound subscripts and superscripts include expressions within parentheses, square brackets, and curly braces. So δ_{μ+ν }is written as δ_{↓}(μ+ν). In addition it is worthwhile to treat two more operators, the comma and the period, in special ways. Specifically, if a subscript operand is followed directly by a comma or a period that is, in turn, followed by white space, then the comma or period appears on line, that is, is treated as the operator that terminates the subscript. However a comma or period followed by a nonoperator is treated as part of the subscript. This refinement obviates the need for many overriding parentheses, thereby yielding a more readable plain text.
Another kind of compound subscript is a subscripted subscript, which works using righttoleft associativity, e.g., a_{↓}b_{↓}c means a_{bc}. Similarly a^{↑}b^{↑}c means a^{bc}.
Parentheses are needed for constructs such as a subscripted superscript like a^{bc}, which is given by a^{↑}(b_{↓}c).
As an example of a slightly more complicated example, consider the expression
which has the plaintext format W^{↑}3β_{↓}㬣_{1}ρ_{1}ς_{2}. In contrast, for TeX, one types
$W^{3\beta}_{\delta_1\rho_1\sigma_2}$ ,
which is hard to read. The TeX version looks distinctly better using Unicode for the symbols, namely $W^{3β}_{γ_1ρ_1ς_2}$ or $W^{3β}_{γ_{1}ρ_{1}ς_{2}}$, since Unicode has a full set of decimal subscripts and superscripts. However the need to use the {}, not to mention the $’s, makes even the last of these harder to read than the plaintext version W^{↑}3β_{↓}㬣_{1}ρ _{1}ς_{2}.
For the ratio
,
the Unicode plain text reads α_{2}^{3}/ (β_{2}^{3} + δ_{2}^{3}), while the standard TeX version reads as
${\alpha^3_2 \over \beta^3_2 + \gamma^3_2}$·
The Unicode plain text is a legitimate mathematical expression, while the TeX version bears no resemblance to a mathematical expression.
TeX becomes very cumbersome for longer equations such as:
.
The Unicode plaintext version of this reads as
W^{↑}3β_{↓}δ_{1}ρ_{1}ς_{2}= U^{↑}3β_{↓}δ_{1}ρ_{1}+ 1/8π^{2}∫_{↓}α_{1}^{↑}α_{2}dα_{2}'[(U^{↑}2β_{↓}㬣_{1}ρ_{1} α_{2}'U^{↑}1β_{↓}ρ_{1}ς_{2})/U^{ ↑}0β_{↑}ρ_{1}ς_{2}]
while the standard TeX version reads as
${W^{3\beta}_{\delta_1\rho_1\sigma_2} = U^{3\beta}_{\delta_1\rho_1} + {1 \over 8\pi^2} \int_{\alpha_1}^{\alpha_2} d\alpha_2’ \left[ {U^{2\beta}_{\delta_1\rho_1}  \alpha_2’ U^{1\beta}_{\rho_1\sigma_2} \over U^{0\beta}_{\rho_1\sigma_2}} \right] }$ .
In a "Unicoded TeX", it could read as
${W^{3β}_{γ_{1}ρ_{1}ς_{2}} = U^{3β}_{γ_{1}ρ_{1}} + {1 / 8π^{2}}
∫_{α_{1}}^{α_{2}} dα_{2}' \left[{U^{2β}_{γ_{1}ρ_{1}}  α_{2}'U^{1β}_{ρ_{1}ς_{2}}
/ U^{0β}_{ρ_{1}ς_{2}}} \right] }$ ,
which is significantly easier to read than the ASCII TeX version, although still much harder to read than the Unicode plaintext version.
Brackets [ ], braces { }, and parentheses ( ) represent themselves in the Unicode plain text, and a word processing system capable of displaying builtup formulas should expand them to fit around what’s inside them. Here U+2032 is used for \prime and U+2044 for \over.
Certain operators like brackets, braces, parentheses, superscript, subscript, integral, etc. have special meaning in this notation. To treat them as normal characters, one precedes them by the "literal operator", for which the percent sign % is handy. So %[ is displayed as an ordinary left square bracket, with no attempt by the software to match a corresponding right square bracket. Table 5.1 lists operators that have special meanings.
Unicode plaintext encoded mathematical expressions can be used as is for simple documentation purposes. Use in more elegant documentation and in programming languages requires knowledge of the underlying mathematical structure. This section describes some of the heuristics that can distill the structure out of the plain text.
Many mathematical expressions identify themselves as mathematical, obviating the need to declare them explicitly as such. One wellknown TeX problem is TeX’s inability to detect expressions that are clearly mathematical, but that are not enclosed within $’s. If one leaves out a $ by mistake, one gets a slew of error messages because TeX interprets subsequent text in the wrong mode.
An advantage of recognizing mathematical expressions without mathon/mathoff syntax is that it is much more tolerant to user errors of this sort. Resyncing is automatic, while in TeX one basically has to start up again from the omission in question. Furthermore, this approach should be useful in an important related endeavor, namely in recognizing and converting the mathematical literature that is not yet available in an objectoriented machinereadable form, into that form.
It is possible to use a number of heuristics for identifying mathematical expressions and treating them accordingly. These heuristics are not foolproof, but they lead to the most popular choices. Special commands discussed at the end of this section can be used to overrule these choices. Ultimately the approach could be used as an autoformat style wizard that tags expressions with a richtext math style. The user could then override cases that were tagged incorrectly. A math style would connect in a straightforward way to appropriate MathML tags.
The basic idea is that math characters identify themselves as such and potentially identify their surrounding characters as math characters as well. For example, the fraction (U+2044) and ASCII slashes, symbols in the range U+2200 through U+22FF, the symbol combining marks (U+20D0  U+20FF), and in general, Unicode characters with the mathematics property, identify the characters immediately surrounding them as parts of math expressions.
If English letter mathematical variables are already given in one of the math alphabets, they are considered parts of math expressions. If they are not, one can still have some recognition heuristics as well as the opportunity to italicize appropriate variables. Specifically ASCII letter pairs surrounded by whitespace are often mathematical expressions, and as such should be italicized in print. If a letter pair fails to appear in a list of common English and European twoletter words, it is treated as a mathematical expression and italicized. Many Unicode characters are not mathematical in nature and suggest that their neighbors are not parts of mathematical expressions.
Strings of characters containing no white space but containing one or more unambiguous mathematical characters are generally treated as mathematical expressions. Certain two, three, and fourletter words inside such expressions should not be italicized. These include trigonometric function names like sin and cos, as well as ln, cosh, etc. Words or abbreviations that are often used as subscripts (see the program in Sec. 5.3 below), also should not be italicized, even when they clearly appear inside mathematical expressions.
Special cases will always be needed, such as in documenting the syntax itself. The literal operator introduced earlier (%) causes the character that follows it to be treated as an ordinary character. This allows the printing of characters without modification that by default are considered to be mathematical and thereby subject to a changed display. Similarly, mathematical expressions that the algorithms treat as ordinary text can be sandwiched between mathon and mathoff symbols. For this purpose, it is tempting to use Asian brackets such as the white lenticular brackets (U+3016 and U+3017), a matched pair that is not common in mathematical expressions. Such "overhead" symbols clutter up the text and hopefully will be rarely needed in Unicode plain text. This does complicate the preparation of technical documents and although one can get very good at it, it is not the most userfriendly way of doing things. On the other hand, identifying the beginning and end of math expressions using $’s or use of extensive markup tags is not user friendly either.
Operands in subscripts, superscripts, fractions, roots, boxes, etc. are defined in part in terms of operators and operator precedence. While such notions are very familiar to mathematically oriented people, some of the symbols that are defined here as operators might surprise one at first. Most notably, the space (ASCII 32) is an important operator in the plaintext encoding of mathematics.
Operators By Precedence 
FF CR \ 
( [ { 
) ] }  
Space " . , =  + LF Tab 
/ * × · • · ½ 
n √ 
∫ ∑ ∏ 
_{↓} ^{↑} 
Here Tab = U+0009, LF = U+000A, FF = U+000C, and CR = U+000D.
As in arithmetic, operators have precedence, which streamlines the interpretation of operands. The operators are grouped above in order of increasing precedence, with equal precedence values on the same line. For example, in arithmetic, 3+1/2 = 3.5, not 2. Similarly the plaintext expression α+ β/γ means
not .
As in arithmetic, precedence can be overruled, so (α+ β)/γ gives the latter.
The following gives a list of the syntax for a variety of mathematical constructs.
exp_{1}/exp_{2}  Create a builtup fraction with numerator exp_{1}and denominator exp_{2}. Numerator and denominator expressions are terminated by operators such as / *] ) ^{↑}_{↓ }· and blank (can be overruled by enclosing in parentheses). The "/" is given by U+2044. 
^{↑}exp_{1}  Superscript expression exp_{1}. The
superscripts ^{0  9 +  ( )} exist as Unicode symbols.
Sub/superscripts expressions are terminated by / * ] )^{ ↑}_{↓
}· and blank. Sub/superscript operators associate right to left. 
_{↓}exp_{1}  Subscript expression exp_{1}. The subscripts _{0  9 +  ( )} exist as Unicode symbols. 
[exp_{1}]  Surround exp_{1} with builtup brackets. Similarly for { } and ( ). 
[exp_{1}]^{↑}exp_{2}  Surround exp_{1} with builtup brackets
followed by superscripted exp_{2} (moved up high enough).
Similarly for { } and ( ). 
√exp_{1}  Square root of exp_{1}. 
·  Small raised dot that is not intended to print. It is used to terminate an operand, such as in a subscript, superscript, numerator, or denominator, when other operators cannot be used for this purpose. Similar raised dots like • and · also terminate operands, but they are intended to print. 
∑_{↓}exp_{1}^{↑}exp_{2 }  Summation from exp_{1} to exp_{2}. _{↓}exp_{1} and ^{↑}exp_{2} are optional. 
∏_{↑}exp_{1}^{↑}exp_{2 }  Product from exp_{1}to exp_{2}. 
∫_{↑}exp_{1}^{↑}exp_{2}  Integral from exp_{1} to exp_{2}. 
exp_{1}½exp_{2}  Align exp_{1} over exp_{2} (like fraction without bar). Useful for building up matrices as a set of columns. 
Diacritics are handled using Unicode combining marks (U+0300 – U+036F, U+20D0 – U+20FF). Note that many more operators can be added to fill out the capabilities of the approach in representing mathematical expressions in Unicode plain (or almost plain) text.
In the middle 1950’s, the authors of FORTRAN named their computer language after FORmula TRANslation, but they only went part way. Arithmetic expressions in Fortran and other current highlevel languages still do not look like mathematical formulas and considerable human coding effort is needed to translate formulas into their machinecomprehensible counterparts. For example, Fortran’s superscript a**k isn’t as readable as a^{k} and Fortran’s subscript a(k) isn’t as readable as a_{k}. Bertrand Russell once said that notation is a great teacher and that a perfect notation would be a substitute for thought. From this point of view, modern computer languages are badly lacking. Specialized mathematical applications such Mathematica are substantially better in this regard.
Using real mathematical expressions in computer programs would be far superior in terms of readability, reduced coding times, program maintenance, and streamlined documentation. In studying computers we have been taught that this ideal is unattainable, and that one must be content with the arithmetic expression as it is or some other nonmathematical notation such as TeX’s. It is time to reexamine this premise. Whereas true mathematical notation clearly used to be beyond the capabilities of machine recognition, we feel it no longer is.
In general, mathematics has a very wide variety of notations, none of which look like the arithmetic expressions of programming languages. Although ultimately it would be desirable to be able to teach computers how to understand all mathematical expressions, it is useful to start with the Unicode plaintext format presented here.
In raw form, these expressions look very like traditional mathematical expressions. With use of the heuristics described above, they can be printed or displayed in traditional builtup form. On disk, they can be stored in pureASCII program files accepted by standard compilers and symbolic manipulation programs like Derive, Mathematica, and Macsyma. The translation between Unicode symbols and the ASCII names needed by ASCIIbased compilers and symbolic manipulation programs is carried out via tablelookup (on writing to disk) and hashing (on reading from disk) techniques.
Hence formulas can be at once printable in manuscripts and computable, either numerically or analytically. Note that this is a goal of MathML as well, but attained in a relatively complex way using specialized tools. The idea here is that regular programming languages can have expressions containing standard arithmetic operations and special characters, such as Greek, italics, script, and various mathematical symbols like the square root. Two levels of implementation are envisaged: scalar and vector. Scalar operations can be performed on traditional compilers such as those for C and Fortran. The scalar multiply operator is represented by a raised dot, a legitimate mathematical symbol, instead of the asterisk. To keep auxiliary code to a minimum, the vector implementation requires an objectoriented language such as C++.
The advantages of using the Unicode plain text are at least threefold:
1) many formulas in document files can be programmed simply by copying them into a program file and inserting appropriate multiplication dots. This dramatically reduces coding time and errors.
2) The use of the same notation in programs and the associated journal articles and books leads to an unprecedented level of self documentation. In fact, since many programmers document their programs poorly or not at all, this enlightened choice of notation can immediately change nearly useless or nonexistent documentation into excellent documentation.
3) In addition to providing useful tools for the present, these proposed initial steps should help one figure out how to accomplish the ultimate goal of teaching computers to understand and use arbitrary mathematical expressions. Such machine comprehension would greatly facilitate future computations as well as the conversion of the existing paper literature and PenWindows input into machine usable form.
The concept is portable to any environment that supports a large character set, preferably Unicode, and it takes advantage of the fact that highlevel languages like C and Fortran accept an "escape" character ("_" and "$", respectively) that can be used to access extended symbol sets in a fashion similar to TeX. In addition, the builtin C preprocessor allows niceties such as aliasing the asterisk with a raised dot, which is a legitimate mathematical symbol for multiplication. The Java and C# languages allow direct use of Unicode variable names, which is a major step in the right direction. Compatibility with unenlightened ASCIIonly compilers can be done via an ASCII representation of Unicode characters.
To get an idea as to the differences between the standard way of programming mathematical formulas and the proposed way, compare the following versions of a C++ routine entitled IHBMWM (inhomogeneously broadened multiwave mixing)
void IHBMWM(void) { gammap = gamma*sqrt(1 + I2); upsilon = cmplx(gamma+gamma1, Delta); alphainc = alpha0*(1(gamma*gamma*I2/gammap)/(gammap + upsilon)); if (!gamma1 && fabs(Delta*T1) < 0.01) alphacoh = half*alpha0*I2*pow(gamma/gammap, 3); else { Gamma = 1/T1 + gamma1; I2sF = (I2/T1)/cmplx(Gamma, Delta); betap2 = upsilon*(upsilon + gamma*I2sF); beta = sqrt(betap2); alphacoh = 0.5*gamma*alpha0*(I2sF*(gamma + upsilon) /(gammap*gammap  betap2)) *((1+gamma/beta)*(beta  upsilon)/(beta + upsilon)  (1+gamma/gammap)*(gammap  upsilon)/ (gammap + upsilon)); } alpha1 = alphainc + alphacoh; }
[TBD: Discuss the interaction of C operators like  with math operators]
The above function runs fine with current C++ compilers, but C++ does impose some serious restrictions based on its limited operator table. For example, vectors can be multiplied together using dot, cross, and outer products, but there’s only one asterisk to overload in C++. In builtup form, the function looks even more like mathematics, namely
The ability to use the second and third versions of the function was built into the PS Technical Word Processor circa 1988. With it we already came much closer to true formula translation on input, and the output is displayed in standard mathematical notation. Lines of code can be previewed in builtup format, complete with fraction bars, square roots, and large parentheses. To code a formula, one copies (cut and paste) it from a technical document into a program file, inserts the appropriate raised dots for multiplication and compiles. No change of variable names is needed. Call that 70% of true formula translation! In this way, the C++ function on the preceding page compiles without modification. The code appears nearly the same as the formulas in print [see Chaps. 5 and 8 of Meystre and Sargent].
Open issues and future. Questions remain, such as whether subscript expressions in the Unicode plain text should be treated as part of programvariable names, or whether they should be translated to subscript expressions in the target programming language. Similarly, it would be straightforward to automatically insert an asterisk (indicating multiplication) between adjacent symbols, rather than have the user do it. However here there is a major difference between mathematics and computation: symbolically, multiplication is infinitely precise and infinitely fast, while numerically, it takes time and is restricted to a binary subset of the rationals with very limited (although often adequate) precision. Consequently for the moment, at least, it seems wiser to consider adjacent symbols as part of a single variable name, just as adjacent ASCII letters are part of a variable name in current programming languages. Perhaps intelligent algorithms will be developed that decide when multiplication should be performed and insert the asterisks optimally.
Export to TeX is similar to that to programming languages, but has a modified set of requirements. With current programs, comments are distilled out with distinct syntax. This same syntax can be used in the Unicode plaintext encoding, although it is interesting to think about submitting a mathematical document to a preprocessor that can recognize and separate out programs for a compiler. In this connection, compiler comment syntax is not particularly pretty; ruled boxes around comments and vertical dividing lines between code and comments are noticeably more readable. So some refinement of the ways that comments are handled would be very desirable. For example, it would be nice to have a vertical windowpane facility with synchronous windowpane scrolling and the ability to display C code in the left pane and the corresponding // comments in the right pane. Then if one wants to see the comments, one widens the right pane accordingly. On the other hand, to view lines with many characters of code, the // comments need not get in the way. Such a dualpane facility would also be great for working with assemblylanguage programs.
Export to TeX. With TeX, the text surrounding the mathematics is part and parcel of the technical document, and TeX needs its $’s to distinguish the two. These can be included in the plain text, but we have repeatedly pointed out how ugly this solution is. The heuristics described above go a long way in determining what is mathematics and what is natural language. Accordingly, the export method consists of identifying the mathematical expressions and enclosing them in $’s. The special symbols are translated to and from the standard TeX ASCII macros via table lookup and hashing, as for the program translations. Better yet, successors of TeX should consistently use Unicode.
This section has shown how, with a few additions to Unicode, mathematical expressions can usually be represented with a readable Unicode plaintext format. The text consists of combinations of operators and operands. A simple operand consists of a span of nonoperators, a definition that dramatically reduces the number of parenthesisoverride pairs and thereby increases the readability of the plain text. The only disadvantage to this approach versus TeX’s ubiquitous { } pairs is that the user needs to know what characters are operators. To reveal the operators, operatoraware editors could be instructed to display operators with a different color or some other attribute. To simplify the notation, operators have precedence values that control the association of operands with operators unless overruled by parentheses. Heuristics can be applied to the Unicode plain text to recognize what parts of a document are mathematical expressions. This allows the Unicode plain text to be used in a variety of ways, including in technical document preparation, symbolic manipulation, and numerical computation.
The heuristics given for recognizing mathematical expressions work well, but they are not infallible. An effective use of the heuristics would be as an autoformatting wizard that delimits what it thinks is mathematics with mathematics on/off codes. The user could then overrule incorrect choices. Once marked unequivocally as mathematics (an alternative to TeX’s $’s), export to MathML, compilers, and other consumers of mathematical expressions is straightforward. We have a workable plaintext encoding of mathematics that looks very much like mathematics even with the most limited display capabilities. Appropriate display software can make it look like the real thing.
[Charts]  The online code charts can be found at http://www.unicode.org/charts/ An index to characters names with links to the corresponding chart is found at http://www.unicode.org/charts/charindex.html 
[Data]  <Placeholder for reference to math specific property files> 
[EAW]  Unicode Standard Annex #11, East Asian Width. http://www.unicode.org/unicode/reports/tr11 For a definition of East Asian Width 
[FAQ]  Unicode Frequently Asked Questions http://www.unicode.org/unicode/faq/ For answers to common questions on technical issues. 
[Glossary]  Unicode Glossary http://www.unicode.org/glossary/ For explanations of terminology used in this and other documents. 
[LaTeX]  LaTeX: A Document Preparation System, User's Guide & Reference Manual, 2nd edition, by Leslie Lamport (AddisonWesley, 1994; ISBN 1201529831) 
[MathML]  Mathematical Markup Language (MathML™) 1.01
Specification. (W3C Recommendation, revision of 7 July 1999.)
Editors: Patrick Ion and Robert Miner. http://www.w3.org/TR/RECMathML/ 
[Meystre]  P. Meystre and M. Sargent III (1991), Elements of Quantum Optics, SpringerVerlag 
[Normalization]  Unicode Standard Annex #15: Unicode Normalization
Forms http://www.unicode.org/unicode/reports/tr15/ 
[Reports]  Unicode Technical Reports http://www.unicode.org/unicode/reports/ For information on the status and development process for technical reports, and for a list of technical reports. 
[SI]  International System of Units (SI)  Système Internationale d' Unites. The metric system of weights and measures based on the meter, kilogram, second and ampere, Kelvin and candela. 
[STIX]  STIX Project Home Page: http://www.ams.org/STIX 
[TeX] 
Donald E. Knuth,The T_{E}Xbook,
(Reading, Massachusetts: AddisonWesley 1984) Donald E. Knuth, T_{E}X, the Program, Volume B of Computers & Typesetting, (Reading, Massachusetts: AddisonWesley 1986) 
[TUS]  The Unicode Standard, Version 3.0, (Reading, Massachusetts: AddisonWesley Developers Press 2000) or online as http://www.unicode.org/unicode/uni2book/u2.html 
[U3.1]  Unicode Standard Annex #27: Unicode 3.1 http://www.unicode.org/unicode/reports/tr27/ 
[U3.2]  Unicode Standard Annex #28: Unicode 3.2 http://www.unicode.org/unicode/reports/tr28/ 
[UCD]  Unicode Character Database. http://www.unicode.org/Public/UNIDATA/UnicodeCharacterDatabase.html For and overview of the Unicode Character Database and a list of its associated files 
[UXML]  Unicode Technical Report #20: Unicode in XML and other
Markup Languages http://www.unicode.org/unicode/reports/tr20/ 
[Versions]  Versions of the Unicode Standard http://www.unicode.org/unicode/standard/versions/ For details on the precise contents of each version of the Unicode Standard, and how to cite them. 
[XML]  Tim Bray, Jean Paoli, C. M. SperbergMcQueen, Eve Maler, Eds., Extensible Markup Language (XML) 1.0 (Second Edition), W3C Recommendation 6October2000, <http://www.w3.org/TR/RECxml> 
The following four books are entirely about the composition of mathematics  
[Chaundy]  T.W. Chaundy, P.R. Barrett and Charles Batey, The Printing of Mathematics, (London: Oxford University Press 1954, third impression, 1965) [out of print] 
[Wick]  Karel Wick, Rules for Typesetting Mathematics, (Prague: Publishing House of the Czechoslovak Academy of Sciences 1965) [out of print] 
[Swanson] 
Ellen Swanson, Mathematics into Type, (Providence, RI:
American Mathematical Society, 1971, revised 1979, updated 1999 by
Arlene O'Sean and Antoinette Schleyer) 
[Byrd]  Mathematics in Type, (Richmond, VA: The William Byrd Press 1954) [out of print] 
The following books contain material on mathematical composition, but it is not the principal topic covered  
[Maple] 
The Maple Press Company Style Book, (York, PA: 1931)
(reprinted 1942) 
[Manual]  A Manual of Style, Twelfth Edition, Revised
(Chicago: The University of Chicago Press 1969) A chapter "Mathematics in Type" was produced using the Penta (computer) system. 
Changes from Tracking Number 2
Changed many special symbols to NCRs. Fixed an HTML glitch affecting table formatting and fixed contents of Table 2.4. A number of additional typographical mistakes and inconsistencies in the original proposed draft have been corrected. Merged duplicated text in section 2.7 and made additional revisions to further align the text with Unicode 3.2. Minor wording changes for clarity or consistency throughout. (bnb/AF).
Changes from Tracking Number 1
A large number of minor, but annoying typographical and HTML mistakes in the original proposed draft have been corrected. This includes the occasional mistaken character name or code point. Additional entries were made to the references section and new bookmarks and internal links have been added to refer to them from the text. Other minor improvements to the text and formatting have been carried out. Added section 2.10 and revised the first paragraph of section 2 to bring the text inline with Unicode 3.2 (bnb/AF)
Copyright © 20012002 Unicode, Inc. All Rights Reserved. The Unicode Consortium makes no expressed or implied warranty of any kind, and assumes no liability for errors or omissions. No liability is assumed for incidental and consequential damages in connection with or arising out of the use of the information or programs contained or accompanying this technical report.
Unicode and the Unicode logo are trademarks of Unicode, Inc., and are registered in some jurisdictions.