XOM API for Processing XML with Java

["Bradley S. Huffman" <hip AT cs.okstate.edu>]

John Cowan writes:

> Maybe so, and maybe not.  If I write a piece about Lenny, and you quote
> me in your document, your readers will not assume that the Lenny in the
> quote is necessarily your Lenny; since they see that the piece is quoted
> from me, they're more likely to assume that it's my Lenny that's meant.
> But perhaps they'll think that because you are quoting me and have not
> absolutized my reference to Lenny, that my Lenny and your Lenny are the
> same after all.

But a conclusion like that relies on the user knowning the quote came from
you.  If they don't have the infomation (that your the original source of
the quote) then the conclusions they can draw are limited.

That's why I advocate carrying the URI of the original source (or null if
one doesn't exists). This gives a application access thru the XOM api to
the following info.

    1) the ancestors of a node
    2) any base attributes on a node, which obviously when used with
       item #1 base attributes on any ancestors
    3) the URI of were the node original came from, not the *calculated*
       base URI, but it's actual source
    4) a general algorithm, thru getBaseURI, for calculating the base
       URI based on the current context that works in most cases.

Now in those cases where the algorithm in getBaseURI breaks down or doesn't
cover a given context, the application still has access to the information
for determining the "right" conclusion in they particular context.

With the raw data the user's application can adapt, without it's SOL.

Brad