Mark Waschkowski wrote:
My understanding of this is that column specifiers can only be used for properties of a node that have been declared using custom node types? Is that correct?
your understanding of the specification is correct, but jackrabbit allows you to use residual properties in a query.
If so, I'm curious about why of this choice? Is it somehow performance/indexing related ?
kind of. there might be implementations that store residual properties different from regular properties and are therefore unable to (efficiently) execute such queries.
regards marcel
