Why the fast_* slots?

[Jeremy @ Cowgar]

Hello Lisaac,

  I  am curious, when looking at the API (right now ABSTRACT_STRING) I
  see  all sorts of fast_* slots then non fast_ slots that simply call
  the equivelent fast_* slot. An example is:

  - fast_last_index_of ch:CHARACTER :INTEGER <-
  // Gives the index of the last occurrence `ch' or
  // 0 if none.
  (+ result:INTEGER;
    result := 1 + storage.fast_reverse_index_of ch from (upper-1);    
    
    ? {(result != 0) ->> {item result = ch}};    
    result
  );
    
  - last_index_of c:CHARACTER :INTEGER <-
  // Index of last occurrence of `c' at index upper or before index upper.
  (
    fast_last_index_of c
  );


  Why is this? It seems to be duplicating a lot of work and making the
  API  confusing. We have two slots that do the same thing, we have to
  duplicate  the  slot  definition  (with a different name of course),
  then  the  parameters,  the return type, and the documentation (which is
  different in the above example).

  Further,  why  would one use last_index_of instead of fast_last_index_of?
  The  existence  of fast_last_index_of would infer that last_index_of
  is a "slow" last_index_of.

-- 
Best regards,
 Jeremy                          mailto:jeremy at cowgar.com