Lists Home |
Date Index |
John Cowan <email@example.com> wrote:
| Am I right in thinking that a SHUNCHAR should not appear as-is in an
ISO 8879 says
4.297 shunned character (number): A character number
...that should be avoided by documents...because some systems
might erroneously treat it as a control character.
Note the "should" there, and that "there is no statement of what an SGML system
should do if it does find such characters" "Shunned does not mean prohibited"
but rather "ignored"
So it is not an error for an (external) entity to contain a shunned character (number). Whether a processor strips out a shunned character, or labels it in the infoset, or
treats the shunning as a matter of human documentation is not mandated by
ISO 8879 AFAICS.
If the file is coming in from the outside world, it is clearly a mistake if
there is a control character in the data, and an entity manager which uses
the "shunned characters" to strip out the control char is not doing the wrong thing.
(I would say that for internal entities, it is a different matter.) The right
thing to do does not come from ISO 8879": it just provides a way
to label characters with their class.
Ultimately, boils down to whatever SP does.