Re: Sets of constants

Frode Vatvedt Fjeld <[email protected]> writes:
> My own idea so far is to define for each set of constants a pair of
> functions, where the first function maps names (symbols, or keywords
> if the definition is supposed to be used from outside the package) to
> values, and the second maps values to names. Compile-time expansion of
> constant names will be provided.
[...]
> Any other ideas, concerns, or experiences?

FWIW, we have a similar need in CFFI for enums and bitfields and we use
two operators (integer to symbol and and vice-versa) that take the type
as their first argument.  In the case of bitfields an integer usually
maps to more than one symbol, of course.

Why not plain old +constants+ with defconstant?  I'm not suggesting
that's a better idea, just curious.

-- 
Luís Oliveira
http://student.dei.uc.pt/~lmoliv/