Summary: I want a byte-string API. I want string integrations. I want
global UTF8 strings.
Is Chicken to be a development tool for all kinds of software,
including i18n applications for general users (my major focus, all
indications to the contrary), or for in-house tools only.
Text is still the method for communication btwn people & machines.
People like to use their native "tongue". Programmers like to use
common APIs.
The current method of relying on (as opposed to using) integrations
is unacceptable. But a separate namespace string API (for text
strings, not binary strings) is also unacceptable. An arbitrary text
algorithm should work w/ byte-strings & utf8-strings. The programmer
using the algorithm makes the "call".
An application should be all byte-strings or all utf8-strings (Alex's
concern about character indices noted). Where an algorithm is
concerned solely with binary data in byte-strings then explicitly use
a byte-string API. (The same could apply to UTF8 but the "generic"
string API is fine for this case.)
I may be missing something but an algorithm that processes both text
& binary data is pretty rare. While common sources & sinks exist
there is usually a way to determine the "type" & direct the stream
accordingly.
(IMHO the situation w/ numbers is actually worse.)
Best Wishes,
Kon
_______________________________________________
Chicken-users mailing list
[email protected]
http://lists.nongnu.org/mailman/listinfo/chicken-users
