> Stefano Mazzocchi <[EMAIL PROTECTED]> wrote: > > > <snip on procedurally oriented XSLT/> > > > > I don't have time to spend on this, but I think what you > really want > > is something more like > > See? me neither. I don't have time to spend on an *elegant* > solution for > my view.
Fair enough, but the solution I wrote took me about 5 minutes and as far as I can tell it works perfectly except for the empty cell handling. (I didn't have time to add the empty cell handling last night.) The point is: you wrote more code than I, so it probably took you longer to do it. The code you wrote was a procedural approach inside of XSLT. In other words, you used the tool incorrectly, so it was hard to use. Now, yes switching mind sets to make the flip from procedural to plucking the nodes you need into sets is not easy if you spend most of your time in Java. However, switching the XSLT semantics from XML to something more compact won't help solve this problem; you'll have to put fewer characters down on paper, but the code design still needs to be done properly. > Views should be dead simple for me to write, they should not > require me to sit in front of a whiteboard for crafting the perfect > recursively declarative pattern. Yes, and computers should magically do what I want just by listening to me tell them about the problem. Views can be that simple (we're doing it here), but the work to make the views that simple is non-trivial and computationally expensive. You know better than to ask for "dead simple": you can't get good code for free! > And anyway, when somebody points you to the stars, make sure > you don't stop looking at the finger. Sorry, I don't see any stars the direction the finger is pointed in this particular case....
