Cart
[Home] [By Thread] [By Date] [Recent Entries]
> > The critical thing here to understand is that the > relationship cannot typically be deduced from the data > itself. The relationship usually originates in the brain of a human. > Again, I think we're mixing concepts that belong on different levels. Data represents entities, attributes, and relationships. Whatever data model you use, those three things are intrinsic, and have equal weight. Saying that John is the father of Mary is exactly the same kind of assertion as saying John is six feet tall - they are both intrinsic parts of the data. Relationships are not different in nature from attributes. The hyperlinking community is trying to enable a different kind of relationship to be expressed, a link that is superimposed on the data by someone other than the originator of the data: for example "The John Smith who was born in London is the same as the John Smith who died in Paris". Now, that's entirely valid and it would be really nice to be able to do it. But why confine it to linking? It applies equally well to any other kind of annotation of the data by third parties, for example "Did you know, John Smith had a beard?". So in the annotation layer too, relationships are not different in nature from attributes. I think XLink isn't the right shape for the architectural hole it is trying to fill. For some people, it is trying to be XML's one true way of representing relationships in the data, whatever their role and purpose. For others, it is a high-level abstract description of user interface behaviour. For others, it is an annotation layer, describing relationships between data entities that are in some sense not part of the data. That last role seems to be closest to an area where there is a gap to be filled: but in my view, such an annotation layer needs to cover much more than linking. Regards, Michael Kay http://www.saxonica.com/ http://twitter.com/michaelhkay [Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] |
