[R6RS] Case-sensitivity response: proposed addition

[Anton van Straaten]

Michael Sperber wrote:
> Anton van Straaten <anton at appsolutions.com> writes:
> 
> 
>>"However, to accomodate those who rely on case-insensitivity, the
>>report will allow the case sensitivity of expressions to be specified
>>in the read syntax.  The lexemes #!cs and #!ci will be used as
>>prefixes to mark expressions containing symbols which should be
>>treated as case-sensitive or case-insensitive, respectively."
> 
> 
> I think that would be a horrible kludge, much worse than either of the
> individual choices.  

It's more flexible than either of the individual choices.  It addresses 
the fact that case-sensitivity is a kind of cop-out, that is primarily a 
concession to integration with non-Scheme systems, and that abandons a 
feature of Scheme which many (including you, afaik) consider a benefit.

> It piles feature upon feature in the worst possible way.  

Do you have criteria for reaching this conclusion?  To me, the "worst" 
possible way would involve changes to the semantics, not just a feature 
of the external representation which is not required for writing 
ordinary code.

> I could see us talking about in the appendix, allowing
> us to have this kind of thing for a transitory period, but to have
> this as a permanent fixture would create all kinds of problems, and
> confusing programs.

Putting it in the appendix could work.  But what kind of problems are 
you thinking of?  These tokens are somewhat analogous to the #!r6rs 
token.  The #!r6rs token potentially raises a much broader class of 
similar issues, related to mixing of #!r6rs libraries with other kinds 
of code.

Perhaps Matthew could say something about the experience with #cs and 
#ci in PLT.

Anton