On 1/27/24 23:17, John Cowan wrote:
For implementations that use unicode grapheme clusters as characters,
Such implementations are neither R6RS nor R7RS, and dno not in
practice exist AFAIK.
True, it was never released. I hadn't remembered one of the reasons why
until you just now reminded me. R6RS made an implicit decision to
optimize for very short text values based on contiguous non-shared
memory representation and an FFI that required repeating representation
mistakes inherited from ASCII. That was one of the reasons I went quiet
on the standardization efforts, I guess; I had been writing a scheme
implementation optimized for corpus-linguistics work, but after R6RS
declared it to be non-scheme, I decided R6RS was wrong. Afterward my own
lisp diverged pretty hard into private experimentation.
On the bright side, I learned a lot from the experiments.
Bear
