Re: Parameter entities vs. GI name gro

I'm stuck in the command line world right now,so I'm not going to quote
directly, but jus say my piece. I would hate to see parameter entities go
because they provide the only validation-safe way to do modular DTDs.
And I feel that this is, asPaul P. said part of the whole reason for
doing XML at all!

    That said, I agree that parameter entities are an unholy mess. Maybe this is
a place where we should pressure the revision to include a real macro-style
parameter entity substitution rule. We would say that a declaration like
<Entity % foo "bar | baz"> 
defined a string substitution. We would simply eliminate the complexity of SGML.
Then we could have an appendix documenting the complex rules for people who
have a hard requirement for SGMl comaptibility. Given the xml-dev traffic this
would be another great piece of fodder for the TC, if it could still be managed.

   I think losing PEs is losing a very great deal, but simplifying them might
be a practical necessity. I can't address the politics of this, but consider
that wihtout PEs some very important DTDs are going to be _very_ hard to wedge
into XML.

   -- David

Received on Thursday, 19 June 1997 20:50:17 UTC