title: case conversion

specification-1:

Reserve the following names for the following functionality:

 CHAR>LOWER ( c1 -- c2 ) "char-to-lower"
    Convert character to lowercase.

 CHAR>UPPER ( c1 -- c2 )  "char-to-upper"
    Convert character to uppercase.

 All these words may be locale-dependent.

discussion-1:

 The name CHAR>UPPER is also known under the names UPC and toupper
 (the latter beig the name of the corresponding C function).

description-2:

 The comp.lang.froth community completely failed to agree on the
 following proposal:

rejected-proposal:

 Reserve the following names for the following functionalities, allowing
 the people to choose the version they prefer:

 STRING->LOWER ( c-addr len -- ) "string-into-lower" "string-dash-to-lower"
    Convert the string at c-addr of len characters to lowercase.

 STRING->UPPER ( c-addr len -- ) "string-into-upper" "string-dash-to-upper"
    Convert the string at c-addr of len characters to uppercase.

 S>LOWER ( c-adr len -- c-addr len )
    Convert the string at c-addr of len characters to lowercase.

 S>UPPER ( c-adr len -- c-addr len )
    Convert the string at c-addr of len characters to uppercase.

 All these words may be locale-dependent.

discussion-2:

 There have been the following arguments:

 1. If we convert a string to uppercase, in 90% of cases we need
    its address and length after that

 2. It is important that conversion words have a same stack effect
    compatible with NOOP, that is, that conversion words may be
    painlessly and straightforwardly commented out or replaced by NOOP.

 3. The stack effect ( addr len -- adr len ) violates an important
    principle that words consume their arguments.

 4. We need only one pair of string case conversion words.


Page-written-by:
   mlg (who prefers the ( addr len -- addr len ) stack effect)