[R6RS] R6RS Unicode SRFI controversial issues

Michael Sperber sperber
Sun Jun 19 10:05:00 EDT 2005


Here's the redux of my positions:

- I'm for:

> 2) adopt the here-string syntax and allow newlines in the normal  
> string syntax

  (I wouldn't fight for here-strings, though.)

- I'm against any form of \<newline> special, specifically if it has
  space-eating semantics.

- I'm for Kent's suggestion to require characters to be delimited.

- Let's either replace #\newline by #\linefeed or not introduce a
  redundant character name.  The presence of both would be confusing
  if they mean the same thing.

- I agree with Marc that "octal is dead."  But if we don't support it
  for characters, let's not support it for strings.

- I prefer Marc's fixed-length scalar-value notation in strings to
  Matthew's variable-length notation, but am worried about the
  proliferation of different gadgets for doing the same thing.

  Ditching octal escapes would be a first step.  As I argued before, I
  prefer Gambit-C's (early?) method of letting me write a
  Scheme-syntax exact-number literal to denote a Unicode scalar value;
  for strings, we'd merely need to terminate.

- I'm against adding scsh's #< syntax.

I'm for all the rest of Marc's original post, specifically, the case
of character names, #t, #f, and not decoding escapes before lexical
analysis.

-- 
Cheers =8-} Mike
Friede, V?lkerverst?ndigung und ?berhaupt blabla


More information about the R6RS mailing list