I recently encountered an interesting group of people doing what they call "sketching in hardware". Look at last year's conference to get an idea of the diversity of this concept. Underlying all this diversity is a breadboarding approach that feels very much like Max: connecting modules as a form of experimentation, trying to simplify the transition from idea to hardware.
In this kind of environment, it's tempting to start tinkering right away. Sometimes that works, but it's easy to get sidetracked. So I resist the temptation and instead try to spend a lot of time exploring ideas beforehand. I go to art museums and performances, explore the Cycling '74 projects page, and talk with friends.
Once I have a clear idea of the project I want to try, I create a new patcher
and then leave the computer at home and go for a long walk. There are so many ways to realize an idea, and implementing an idea awkwardly takes just as much time as doing it right, so I take time to sort it out and focus on the essentials.
Then it's time for pencil and paper. Start by writing down the key organization and/or esthetic requirements for the patch. What would Version 1 look like? How will I know when I'm finished? Staying with pencil and paper, I sketch out the overall plan for the project, decide which modules I need and what they need to do, and imagine what it will be like to play with the system. Finally, I put together a little to-do list, and head back to the computer.
As I build the patch, I try to keep it in a functioning state: "always up and running". That reminds me of one of my favorite orchestral conductors, who would begin each rehearsal with a run-through. If the patch is always ready to go, then making the final version reliable is easy!
As the functional aspects of the patch come together, it might be a good point to look at ways of refining the overall idea. This would be a great time to step away from the computer once again and pick up a good book. Gregory Taylor turned me on to this gem, Universal Principles of Design:
Looking through this book and thinking about my project, I sometimes get insights into how to refine my project to achieve a kind of elegance. Just because my patch works, doesn't mean its function is optimized. It's satisfying to finish with a Presentation Mode full of grace and functionality.
Keeping the Big Picture clear from the beginning through the end results in a Max project that is fun to use, works well, and opens up options for the future.