[
Lists Home |
Date Index |
Thread Index
]
From: "Paul Prescod" <paul@prescod.net>
> Here's a simpler proposal. Take this:
>
> [1] document ::= prolog element Misc*
>
> And make it:
>
> [1] document ::= prolog (element Misc*)*
>
> This is not straightforwardly SGML compatible, but I think it is
> significantly easier these days to change SGML than it is to change XML.
Nice. But I don't think that it does the same thing: my proposal provides
an alternate syntax for internal entities which also allows bundled
other documents. (Thinking about it more, it suggests that it might fill
some nice needs if the entity mechanism was rejigged to allow appendable
entities rather than having them in a prolog.)
I think you would need to have some top-level attribute available to
say "this branch does not form part of the infoset" so that it can
be referenced from an XInclude without appearing in the document
twice (assuming we will perform some kind of XInclude linking early).
Cheers
Rick Jelliffe
|
