[R6RS] Renaming write-...

[Anton van Straaten]

I think Kent wanted peek-char in the proposed I/O convenience library, 
but it's also in the Port I/O library.  Any suggestions about 
disambiguating its name?

Alternatively, peek-char could be exported from both libraries, except 
that the port argument would be optional in the convenience version. 
I'm not sure that's a problem - the semantics of the two procedures are 
otherwise the same.

(BTW, it occurred to me that if it weren't for write-char, whose 
signature is (write-char char port), we could simply export all three 
conflicting procedures from both libraries, thus avoiding having to 
rename all the read-... and write-... procedures in the Port I/O 
library.  Not sure what would be done about write-char, though.)


Michael Sperber wrote:
> If we're renaming write-..., we should also consider renaming
> read-... to get a matching pair (or at least avoid having a
> non-matching one).  Choices:
> 
> 1. display / read
> 2. put / get
> 3. output / input
> 4. out / in
> 5. push / pull
> 
> (I'm in favor of #2.)
> 
> Please also consider the names of `read' and `write', relating to
> this.  (And the sound of pronouncing "xxx/xxx invariance.")  
> 
> 1. read / write  (awkward with #1)
> 2. read-datum / write-datum (ditto)
> 3. <write-prefix>-datum / <read-prefix>-datum (for choices #2-#5)
> 4. <write-prefix> / <read-prefix>
> 
> (I'm in favor of #3.)
> 
> I'm also thinking these two should be in a separate library.
>