Elliotte Rusty Harold wrote:
> I think there's a way to limit the damage this does to the existing
> infrastructure. Whatever the eventual identifier is chosen for
> Blueberry (version="1.1", unicode="3.1", etc.) I think it should be
> a *fatal error* to use this identifier in a document that does not
> actually use any of the newly introduced characters in an XML name
> somewhere.
John Cowna:
>The trouble with this scheme is that it makes generation on the fly,
>particularly by separate modules, difficult.
I'm particularly having trouble with what it does to manual editing: Someday there will be doctype in which the only Blueberry names are in optional elements. So then, somebody opens a document in Notepad, deletes a single optional element, saves it, and suddenly it's illegal because it no longer includes the bluberry characters.
I grant that the characters are still in the DTD/Schema, but if I understand, a simple WF check should be detecting the error.
Needless to say, this will happen when someone is trying to get an important document out the door and go on vacation.
What am I missing?
Frank