[
Lists Home |
Date Index |
Thread Index
]
[Bullard, Claude L (Len]
> Ok. The serialization issues don't bother me. Should they?
>
> The comment that use of RDF rules to interpret the XML content
> results in loss of information and misinterpretation does.
> Please discuss that further. What is the most I can expect
> from an RDF processor reading the XML? I saw John's reply
> on what it can say (this is a property of this... etc.).
>
> So it seems there has to then be an interpretant that can
> take those "facts" and reason on them. Yes? And then
> for the right things to happen, the results of the reasoning
> engine should not contradict the intending meanings of the
> original XML producer? Yes? Does the loss of information
> you talk about make it difficult to write such an interpretant?
>
The problem, if it is a problem, is that the relationship in XML between an
element and its children is no more or less than "containment" or
"childhood" or some such notion. Any other relationship or meaning that you
want to assign cannot come from the markup structure itself. It is like
having a labeled graph where some of the arcs are missing labels. You can
still construct the graph, but you have to guess at or infer the labels.
With good practices, good naming conventions, etc., you could have a
processor infer those things reasonably well, thus extracting metadata as
you say. But it is not totally cut and dried.
Cheers,
Tom P
|
