This is exactly an XPath ( http://www.w3.org/TR/xpath.html ) query, it will select all <density> elements children of <indumentum> elements children of ... etc.

Other example:
lamina[ancestor(species)//petal[contains(./text(),'white')]]
will select <lamina> elements descendant of <species> elements, whose <petal>
element's text content has 'white' string in it.

The whole thing is to set a botanical vocabulary, allowing indumentum and disallowing (e.g.) pilosity. This is a secondary issue wether this vocabulary will be a DTD, an XML Schema, or a RDF Schema. Anyway, there are tools like XML Authority of Extensibility to translate beetwen these formats (not yet RDF Schema, alas), so we can provide all Schema dialects. This is what most business Consortia do.

For the relational database fans, it is easy to export a many-columns table using a XSLT (http://www.w3.org/TR/xslt.html) transform, with columns names exactly like this:
leaf/lamina/abaxial_surface/vein_islands/indumentum/density

For the DELTA fans, the principle is to associate a character number to each leaf, like
leaf/lamina/abaxial_surface/vein_islands/indumentum/density
of the document Schema. It can be done even without DTD/Schema, see http://jmvanel.free.fr/Samples/DisplayDescriptions/speciesGUI.htm, where I list all the unique nested tags names in the document; this is what I call the implicit Schema. I do Javascript with the Document Object Model.

I have proposals to extend the DELTA system; this will be another thread.

see for more details:
http://jmvanel.free.fr/

Regards
Jean-Marc

Kevin Thiele a écrit :

 Is the entity/property/value schema restricted to 3
levels?

What to do about

leaf/lamina/abaxial_surface/vein_islands/indumentum/density.
I agree that a hierarchically structured schema may be wonderful, but in a
biological context I wouldn't like to see it constrained at 3 levels just
because that's what existing computer formats do.

Cheers - k