Re: Parameter entities vs. GI name groups

David Durand wrote:
> If every string-substitution effect were legal for a PE, then
> implementation would be so easy it's not an issue.
I agree.

> Only WG8 could enable
> this kind of processing. for valid documents, however.

Alternatively (although I'm not wild on this) it could be made a
well-formedness/validity constraint: a valid document would not use
PEs in a way that isn't valid SGML under the 1986 Rules.

A well-formed document migh have fewer constraints.

This in some ways might mirror the "obfuscatory entitiy references" note
in the SGML standard itself, I suppose.

Lee

Received on Monday, 23 June 1997 16:14:51 UTC