Issue DIGIT-CHAR-P-NOT-AN-ACTUAL-PREDICATE
Issue: DIGIT-CHAR-P-NOT-AN-ACTUAL-PREDICATE

Forum: Editorial

References: DIGIT-CHAR-P

Category: CLARIFICATION/CHANGE

Edit history: 2017-01-28, Version 1 by Michal Herda

Status: For CLiki consideration

Problem Description:
The output of DIGIT-CHAR-P might be confusing, as this function's return value also serves a different purpose. DIGIT-CHAR-P, aside from being a predicate, is also a converting function: it converts a digit in a given radix to its corresponding integer value.

Proposal (DIGIT-CHAR-P-NOT-AN-ACTUAL-PREDICATE:SPLIT):
Rename DIGIT-CHAR-P to CHAR-DIGIT and possibly make it signal an error instead of returning NIL. Create a DIGIT-CHAR-P which returns either true or false.

Test case:
(digit-char-p #\E 16) ;=> 14

Rationale:
A predicate function should be a predicate and not additionally serve as a converting function.

Current practice:
DIGIT-CHAR-P is used as a converting function

Cost to Implementors:
Tiny.

Cost to Users:
Incompatibility with previous uses of DIGIT-CHAR-P.

Cost of Non-Adoption:
Confusion coming from the fact that DIGIT-CHAR-P may be used in non-boolean contexts.

Benefits:
Abiding to the single responsibility principle.

Aesthetics:

Discussion:
  • Michał Herda supports DIGIT-CHAR-P-NOT-AN-ACTUAL-PREDICATE:SPLIT.