I've been following this thread with interest, albeit from within the air raid bunker. Martijn seems to have touched a nerve. :^) Here's what I'd like to see. Could someone post a recap of the discussion? I did this recently for the "Zope Front Door" thread and found it to be an invaluable exercise. Zope's usability needs to improve a LOT. I got a good dose of this when I taught a three day Zope course several months back. When you have to explain it to people, you really find out how baroque Zope can be. Though I agree with Andrew that the problem isn't just DTML, that seems to be a locus. DTML *is* being used in a way that violates its design (separation of presentation and logic). But at the same time we at DC lead the way on this violation, and aren't actively promoting viable alternatives. I'll say this about an ngDTML effort: good luck. The current quandry exists for a reason: there are a lot of tough design choices. That which is clear to the Python hacker is obtuse to the HTML newbie. That which is clear to the DTML developer is inefficient to the parser/renderer. That which is... I *really* hope we get some traction on what to do here. --Paul