Thanks for the comments!
I think one of the most important things is making sure the puzzles are reasonably logical whilst not being a pushover to solve, so there is a balance to find and I think this is the hardest bit - making sure the levels as a whole are balanced correctly so it will doubtless require lots of fine-tuning and playtesting. The majority of the actual engine code is now in place and functioning as it should be (one or two minor issues to tidy up aside; mostly cosmetic though) so apart from maybe a couple of extra small bits to go in it's getting to the point where we can concentrate on making the levels as good as they can be. Although our demo level will probably be ever-so-slightly easier as a 'real' demonstration of what the game will be about, but that will no doubt end up being tweaked and changed a million times before it goes out!
I was quite keen to get the character moving fairly quickly as one of the things I didn't like about Dizzy was how slow and plodding it could be when you need to move back and forth between a number of screens, especially when you know exactly what you have to do. And one of the things I am keen to replicate is that buzz of exploration which you mention - I used to love just wandering around seeing what the screens look like and any bits that were tantalisingly out of reach just made me more determined to find a way through - who knows what's over that big gap, or through that locked door...? You can use your imagination when playing... and when designing too