I am wondering how much effort it would take to add a feature to Xalan.
Here's what I'm thinking:
I have some sales data in a DOM source, which is transformed into SVG, and
then rendered to the screen to show a sales chart. The user types in some
numbers and clicks on a button, which adds more data to the source. The
transformer is invoked again, but instead of regenerating the entire result,
it determines which parts of the result are dependent on the changed source
elements, prunes them, and regenerates them (and only them).
The motivation here is performance in an interactive application.
I don't (yet) understand the Xalan architecture, and before I sink my head
into the source code, could someone advise me on how feasible/difficult this
might be? Can you foresee anything that would make it difficult to map the
dependencies?
Thanks,
Paul
