Kevin Messman writes:
> The editor doesn't like entity references.
> The ones I'm worried about are:
> “ ” ‘ ’ ‐ …
> They would be typed as:
> ["] ["] ['] ['] [--] [...]
> Can anyone see any problems that might arise
> from using the ascii characters in place of the
> entity references?
Yes. Your proposed alternatives do not reliably represent the intent.
You have to interpet 'hello' to work out that the ' characters have to
be replaced by someone else. Who will turn ... into an ellipsis?
But why not just enter stuff in a known encoding which supports the
characters? Latin 1 would do fine.
As an aside, are you sure you want SGML still? You dramatically
increase the range of tools available to you if you switch to XML. The
TEI is now fully XML, so the advantages of SGML are few. Unless you
really need something like CONCUR (and have the software to prove it),
its hard to argue for SGML.
And before someone says "XML doesn't support exclusions, etc", I would
reply "but neither does SGML support all sorts of important
constraints you want to impose on your data, so you are better off
looking at better constraint notations anyway (ie schema)"