[r6rs-discuss] Re: [Formal] formal comment (ports, characters, strings, Unicode)

[Per Bothner]

MichaelL at frogware.com wrote:
> There are at least four schools of thought on all of this. First, I 
> believe that some people think a sufficiently smart compiler could hide 
> some/many/most of these issues by, for example, caching information or 
> switching to another encoding on the fly. 

You could have a 16+16-bit cache that maps scalar-value-index to
byte or word offset.  You'd want it an atomic value so you don't
need synchronization.  This cache would be updated on each string-ref.
Since normally strings are normally accessed sequentially, each access
is effectively O(1).

Of course it's kind of ugly having a writable cache word in a read-only 
string.  This might hurt the memory cache in some cases.  And it means
allocating a read-only string in read-only memory is difficult.

Of course string-set! is fundamentally broken and useless as I've
said numerous times.  It could be replaced by:

(string-append-char! str1 ch)
(string-append-string! str1 str2)
   Append ch/str2 to the end to str1.  str1 is modified in-place.
-- 
	--Per Bothner
per at bothner.com   http://per.bothner.com/