Hello,
XML technology excellently supports two fundamental ways how to think about document transformation - push style (XSLT, template based), pull style (XQuery, node constructors with embedded expressions).
Much less important, but perhaps not completely without interest in certain scenarios, may be yet another way - I call it "op style", for want of a better word: transformation as a pipeline of simple operations (e.g. grouping, flattening, deepening, moving, merging, separating, ...). Each operation might be defined in a purely declarative way - operation name plus parameters. Instead of writing code, one might implement simple transformations (of a certain type) by just writing an "opsheet", a flat sequence of operation descriptors. It could be done by people without technical skills, and it might also be useful for technical folks because removing the need to write code, thus saving time. And nice support by IDEs can be imagined. Another aspect - when controlling complex systems, the explicitness of opsheets might be helpful.
My question - does this make sense to you? Do you perhaps know projects or products which offer something similar?
Kind regards,
Hans-Jürgen
