|Version 1 (modified by 11 years ago) (diff),|
It is widely regarded as a mistake that H'98 treats I/O as character-based. The proposal is that I/O should operate over octets (i.e. 8-bit binary values) instead. This would permit the layering of character-encoding conversions (see Unicode) over the top of the basic I/O mechanism.
- cleans up an area of confusion
- backwards compatible with all implementations