Lists Home |
Date Index |
From John Cowen:
> However, the control characters are *characters*, not really very
> different from other control characters in the Unicode space
> which are already allowed: not only the ISO C1 controls, but
> also such things as: the Mongolian variant controls (and the
> Unicode 3.2 generic variant controls); the bidi marks, overrides,
> etc; and the music symbol begins/ends.
The Unicode recommendations w.r.t. control characters are in
That makes it clear that control characters are unlike other characters,
for which Unicode provides "semantics". The only C0 or C1 characters for
which Unicode provides "semantics" are TAB, CR, LF and NEL.
Unicode completely defers the use and semantics of the other control
characters to whatever makes sense for the application in question.
There is no justification for saying "we need to support the C0 and C1
characters in order to support Unicode" because Unicode does not
require any such thing.
But what if we do decide to support these control characters: what does
it mean? It means that we recognize their semantics, according to which
it is inappropriate to embed most of them (e.g. EOF, BS, BELL, flow control,
etc) in a text file for transmission anyway.