Mehdi, when starting a project it is easy to get overwhelmed by the amount of design decisions you need to make. As often said - 'divide and conquer'. When I was younger I get stuck into the thought pattern of solving everything at once - you can't do it on a complex project. so - divide up your project into bite size chunks. Normally you start out with a specification or at best a wish-list. Then you might start to map out the hardware you might need. Then you might start looking at interfacing constraints - 3.3V vs 5v, number of port pins, etc. From that you might be able to start drawing a preliminary schematic. Similarly with the software - don't try to start cutting code (unless its a small program) until you've mapped out your major functions
you might have something like:
menu system
control logic
task switcher
comms protocol
etc....

Then you might start to look at data structures - how these items are going to talk to each other etc. Then you might start to break the major functions down to smaller functions etc. When you get to write the code, you can then concentrate on the function itself without having to worry about the data it uses or how it it passed - you've thought about that already. Eventually you'll have all the code written and you can test each module separately then start to merge everything. You'll find that your project should start to converge at a fast rate.
Also, if you've got a manager or customer breathing down your neck you can start to assign times to each of the items and track your progress. It also allows you to estimate the time so you can quickly figure out if you've underbid and by what margin.


Hopefully I'm not teaching you how to suck eggs! But it sounds like you've got this project and now the panic starts to set in. You're reading the absolute maximum ratings without looking at the normal specs! If the device is 3.3V you should have picked that one up first - it might even be 1.8V!