Re: Internal subset equivalent in new schema proposals?

Michael Kay (M.H.Kay@eng.icl.co.uk)
Wed, 2 Dec 1998 10:18:37 -0000


>How can you represent a strictly acyclic graph using ID/IDREF?

You make the IDREF attribute optional, and use the absence of a value (the
#IMPLIED default) to indicate a node that has no "parent".

There are many other solutions, e.g. you could have a dummy element with the
special ID value of "NULL", but I can't see why one would want to.

Mike Kay