This is a simple question about the terminology.
C11, 6.4.2.1p1:
nondigit: one of
_ a b c d e f g h i j k l m
n o p q r s t u v w x y z
A B C D E F G H I J K L M
N O P Q R S T U V W X Y Z
The "nondigit" name is confusing because, for example, the $ is also a non-digit.
Does anyone know why (i.e the rationale) was it named "nondigit" instead of "letter-or-underscore"?
For naming identifiers, an implementation is required to support digits 0 to 9 and non-digits A to Z (upper + lower case) as well as the underscore _. Since _ is not a letter, this group can't be named letters: or similar. The only reason why two groups of symbols are required in the formal syntax is because an identifier cannot start with a digit, for historical reasons.
$ is not a valid symbol to use as part of an identifier in the C language. You might be confused by non-conforming compilers such as gcc allowing $ even in it's supposed conforming mode -pedantic. ISO C lists this as a "common extension" in Annex J.5.
The use of the term in the B programming language, which influenced the development of C, suggests that it was chosen for historical reasons.
The B language had a similar syntax to C and used nondigit to refer to the set of characters that could be used in identifiers, as seen in this snippet of B code:
main( ) {
auto a, b, c, sum;
a = 1; b = 2; c = 3;
sum = a+b+c;
putnumb(sum);
}
(For additional information auto is used to define 36-bit variables in B)
It is likely that the term nondigit was carried over from B into C, as C was developed as an extension of B. The use of nondigit has persisted in subsequent versions of the C standard, including the current C18 standard.
Furthermore, the use of concise and unambiguous language is a common practice in programming language specifications. This can be seen in the use of terms like whitespace and newline to describe specific sets of characters, rather than more descriptive terms like blank space or line break. The use of nondigit in C follows this practice, providing a clear and specific definition of the set of characters that can be used in identifiers.
While it may seem more descriptive to use a term like letter-or-underscore to describe this set of characters, it is important to note that this could potentially lead to confusion or misinterpretation. For example, it is possible for a programming language to include other special characters in its set of valid identifier characters, such as accented letters or currency symbols. Using a more specific term like nondigit makes it clear that only the characters listed in the C specification are valid for use in identifiers.
Ultimately, the use of the term nondigit in C is a matter of convention and historical precedent. While it may not be the most descriptive term, it has been used in the language for many years and is well-established in the C specification.
