[R6RS] The unspecified values, multiple-value semantics and all that

Anton van Straaten anton at appsolutions.com
Tue Nov 7 04:06:03 EST 2006

William D Clinger wrote:
>  *  Do you really want to try to explain all this over in
>     the R6RS discussion group?  

I think we should make a decision, and (at some point) announce it.

For me, the main obvious choices are (a) to rename the unspecified value 
to something less amusing; or (b) revert to R5RS semantics.

The concerns about the ramifications of a switch to zero-return values 
haven't changed for me, no matter which variation on those semantics are 
used.  Will has laid them out very clearly.

Regarding (a), Will observed that the proposed unspecified value "did 
not actually increase portability in any meaningful way."  To me, the 
main reason to standardize this at all would have been to achieve a 
minor cosmetic standardization: the ability to write down (e.g. in 
email, a book, or lecture notes) the external representation of the 
value of expressions such as (list (if #f #f)), without having to 
mention the umpteen pointlessly different ways that different Schemes 
represent it.

Since we haven't gone that far anyway, and if there isn't agreement for 
that, then I don't see much reason not to revert to R5RS semantics.


More information about the R6RS mailing list