[r6rs-discuss] Case-Sensitivity Optional, Case-Preservation Mandatory

[Sam TH]

On Mon, Feb 23, 2009 at 11:11 AM, John Cowan <cowan at ccil.org> wrote:
> Erich Rast scripsit:
>
>> I submit that all implementations MUST be case-preserving under
>> certain circumstances. Whatever is done to source code/data file input
>> MUST preserve case information if (a) the implementation provides a way
>> to retrieve the original symbols of the source data (i.e. the source
>> data is not compiled to something else entirely anyway), and (b) the
>> respective symbols are not mapped to another case explicitly.
>
> With respect, that seems to me utterly incoherent as stated.  The essence
> of case-insensitivity is that foo and FOO are the same symbol; that is,
> when you "retrieve the original symbols of the source data" for these
> two spellings, they are eq? as symbols.  How then can they be both eq?
> and distinguishable?
>
> Usually when we speak of case-preservation, we mean that the case of
> the *defining* instance is preserved.  On the case-insensitive NTFS file
> system, for example, the case used when a file is created is preserved.
> Is that what you mean?

NTFS is a case-sensitive file system.  However, the Win32 APIs provide
a case-insensitive (but case-preserving) layer on top of that.

 http://en.wikipedia.org/wiki/Comparison_of_file_systems#cite_note-note-36-56
 http://en.wikipedia.org/wiki/NTFS
-- 
sam th
samth at ccs.neu.edu