New control characters! (was: Re: Unicode "no-op" Character?)

Sławomir Osipiuk via Unicode unicode at unicode.org
Tue Jun 25 19:27:47 CDT 2019


All right. Thanks to everyone who offered suggestions. I think the final
choice will depend on the specific application, if I ever face this puzzle
again.

 

If nothing else, this discussion has helped me formulate what exactly it is
I'm imagining, which is actually a bit different that was I started with.
So, just to put it out there for the internet to archive (with the likes of
the various proposed "unofficial" UTFs I've been reading about), here are my
two proposed control characters (why just one when you can have two at twice
the price?)

 

Implementors, feel free to jump right on this. :-) I chose to assign them to
0xE and 0xF because the use of ISO2022-style stateful shifts is expressly
not permitted by ISO 10646, so by my reading the existence of those code
points inside a UCS stream is a roundabout error. Therefore I'm reclaiming
them for something useful.

 

 

EP1 - EPHEMERAL PRIVATE SENTINEL 1 (0x0E)

 

EP1 is executed as a null operation at the presentation layer. The formation
of ligatures, the behavior of combining characters, and similar presentation
mechanisms, must proceed as if EP1 were not present even when it occurs
within sequences that effect such mechanisms.

EP1 is intended to be used as a private process-internal sentinel or flag
character. EP1 may be added at any positions in the character stream. EP1
may be removed from the stream by any receiving process that has not
established an agreement for special handling of EP1.

EP1 should be removed from the stream prior to any security validation. It
must not interfere with the recognition of security-sensitive keywords,
sequences, or credentials.

 

 

EP2 - EPHEMERAL PRIVATE SENTINEL 2 (0x0F)

 

EP2 is executed as a null operation at the presentation layer. The formation
of ligatures, the behavior of combining characters, and similar presentation
mechanisms, must proceed as if EP2 were not present even when it occurs
within sequences that effect such mechanisms.

EP2 is intended to be used as a private process-internal sentinel or flag
character. EP2 may be added at any positions in the character stream. EP2
may be removed from the stream by any receiving process that has not
established an agreement for special handling of EP2.

EP2 should be removed from the stream prior to any security validation. It
must not interfere with the recognition of security-sensitive keywords,
sequences, or credentials.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://unicode.org/pipermail/unicode/attachments/20190625/c5716037/attachment.html>


More information about the Unicode mailing list