Special math alphabets

But simple formulas with one alphabet and a huge number of symbols are not sufficient for mathematicians to expose their thoughts properly. They tend to use every available typeface to denote special things.

To cope with this need for special alphabets in formulas,' we introduce the concept of 〈math alphabet identifiers〉. These constructs are special commands which switch to a specific typeface. They might correspond to different typefaces in different math versions but within one version they always select the same typeface regardless of surrounding conditions.

A 〈math alphabet identifier〉 can be defined according to the users' needs but standard L^ATEX already has a few of them built in. They are described in table .

Table: Predefined 〈math alphabet identifiers〉 in L^ATEX

When using such an 〈alphabet identifier〉 two syntax variants are available: one can understand a command like \cal as a switch to a different font, i.e. using a syntax {\cal ...} as the old L^ATEX does, but we prefer to view the 〈math alphabet identifier〉 as a command with one argument, i.e. to use a syntax of the form ..\cal{A}.. To select the first alternative a style option `nomargid' is provided. This option is automatically selected if the `oldlfont' option is used since this option is supposed to produce identical results for older documents.

New 〈math alphabet identifiers〉 are defined in two steps. First the identifier is made known to the system with the \newmathalphabet command. Then specific typefaces in some or all 〈math versions〉 are assigned by means of the \addtoversion command.

Let us discuss this process in detail. Suppose that you want to make a sans serif typeface available as a math alphabet. First we choose a new command name (e.g. \sfmath) and tell L^ATEX about it with the line

 \newmathalphabet{\sfmath}

Then we consult table  to find suitable fonts to assign to this alphabet identifier. As you find out, the computer modern sans serif family consists of three series, a medium, semi bold condensed and a bold extended one. The medium and the bold extended series both contain a normal shape typeface. So we add the line:

 \addtoversion{normal}{\sfmath}{cmss}{m}{n}
 \addtoversion{bold}{\sfmath}{cmss}{bx}{n}

A_i = tanα

\mathversion{normal}
\[  \sum \sfmath{A}_{i} = \tan \alpha \]

If we are interested in a slanted shape we have to face a problem: there is no slanted shape in the bold extended series of the Computer Modern sans serif family. So, if we make the identifier known only in the normal version then it would produce an error message when encountered in the bold (or any other) version. Of course we can get by using always the same typeface in all versions. To make this task a bit easier there is also a * variant of the \newmathalphabet command which takes three more arguments: the default values for family, series and shape for all math versions in which the alphabet identifier is not explicitly defined via an \addtoversion command. So our second example can be set up simply by stating

 \newmathalphabet*{\sfslmath}{cmss}{m}{sl}

Here we show the same formula as above, but this time in the Euler version and with \sfslmath instead of \sfmath:

A_i = tanα