> Although there are still a few things Ron and John disagree on,
> I have written a unified draft, basically by taking Ron's draft
> and hacking on it a bit. Ron hasn't seen it yet, though, so
> it may change further in the process of converging.
>
> The URL: http://www.ccil.org/~cowan/XSchema-draft-DTD-19980605.txt .
> You may also need: http://www.ccil.org/~cowan/ibtwsh.dtd .
> In the interests of speed, I have not published new prose or a new
> meta-XSchema yet.
>
> New and changed stuff from Ron's last draft (*s mark what Ron has
> NOT approved-in-principle yet):
>
[change list snipped]
I've read the changes and agree. I'm still a bit uneasy with the following two
items, but I'm willing to agree to whatever the list decides, including leaving
them as they are:
1) If an attribute type is not declared, it is assumed to be CDATA. I lean
toward explicitness.
2) A Choice or Seq must have two or more elements. This makes logical sense,
but I'm still a bit afraid that people will commonly include only a single
element and that we ought to allow it. Perhaps it's due to one of my mother's
favorite sayings: "We got vanilla [ice cream]. Waddya want?"
My Web site now points to John's.
-- Ron Bourret