Lists Home |
Date Index |
I'd certainly agree that PEs were/are one of the least well
("precisely"?) specified portions of the XML REC. It's
gotten better, but anyone expecting consistent behavior
in the corner cases there is asking for trouble.
So I agree with Michael Kay that the current state of PEs
seems to be, to a worrisome degree, fragile. I've not
seen such issues with GEs though; they're much simpler.
----- Original Message -----
From: "Richard Tobin" <firstname.lastname@example.org>
Sent: Thursday, January 17, 2002 9:28 AM
Subject: Re: [xml-dev] Parameter entity references
> >> I am very wary of proposing errata to fix this kind of thing,
> >> because it is so easy to produce unintended consequences.
> >This is an admission that the whole business of parameter entity references
> >is so fragile that no-one dares touch it in case the whole edifice comes
> >tumbling down.
> It was certainly not intended to be.
> For one thing, they're a subset of what you can do in SGML. I think
> it was hoped that XML could get by without the SGML mechanisms of Ee
> (entity and) and so on to describe it, but the result has turned out
> to have a number of flaws.
> I don't see any problem with parameter entities themselves, only with
> the way they are described in the spec.
> -- Richard