XOM API for Processing XML with Java

[Elliotte Rusty Harold <elharo AT metalab.unc.edu>]

At 4:46 PM -0500 3/9/04, jcowan AT reutershealth.com wrote:


> Well, within a single package of XOM subclasses, it's redundant
> to have checkAttach and checkDetach, because you can always use
> the double-dispatch pattern: the parent can ask the child via a
> package-private interface "Do you want to be attached to/detached
> from me?"  Where it gets important is when multiple packages are trying
> to interact, since the parent has no way to get the child's input on
> the subject.

That's ugly. With or without double dispatch my question is still why 
does the child care what its parent is? Especially if the parent 
doesn't care what its child is? Is it that you'd like an HTMLElement 
that could only be a child of an HTMLElement or an HTMLDocument, but 
could not be a child of a generic Document or Element?

I guess I'm stuck on the fact that with the notable exception of 
Schematron, most validity rules flow from the parent top the child, 
not the other way around. When we write a DTD or a schema we list the 
allowed children, not the allowed parents.
--

  Elliotte Rusty Harold
  elharo AT metalab.unc.edu
  Effective XML (Addison-Wesley, 2003)
  http://www.cafeconleche.org/books/effectivexml
  http://www.amazon.com/exec/obidos/ISBN%3D0321150406/ref%3Dnosim/cafeaulaitA