I think Gen7 is good and Traumflug is awesome guy and top contributor. My opinion is that if you build and debug your own made electronics then you will understand much more than the usual user.
Ofc ppls will help if you want to learn more, thats why the forum is here. I think also what is important about this, is to have your opinions sort of flexible, and be prepared to view other ppls standing point, and be ready to change yours when you meet something better. Development depends on ability to change one's self, sort of speaking will not evolve if not let yourself change.
I think you will need a programmer, i would recommend AVRISP mkII There would ofc be other alternatives, perhaps like 10-15$ cheaper, but might require things like own drivers, perhaps other changes, and may have some limitations like <128K. But that programmer is by atmel, can program full 256K if needed (mega2560), perhaps bigger range of chips, perhaps bit better support in many programs.
One thing which i would consider pitfall as from my own experience, falling to the temptation to try test a board at an incomplete / intermediary stage. I had few of those and while thinking that i had in mind a clear part of schematic that should of work in my head, the board would be at a stage actually something different than what i had in mind, but differences are hard to realize between a half schematic and a half populated board. So now i would not test something unless it is 100% fully populated pcb.
Ofc ppls will help if you want to learn more, thats why the forum is here. I think also what is important about this, is to have your opinions sort of flexible, and be prepared to view other ppls standing point, and be ready to change yours when you meet something better. Development depends on ability to change one's self, sort of speaking will not evolve if not let yourself change.
I think you will need a programmer, i would recommend AVRISP mkII There would ofc be other alternatives, perhaps like 10-15$ cheaper, but might require things like own drivers, perhaps other changes, and may have some limitations like <128K. But that programmer is by atmel, can program full 256K if needed (mega2560), perhaps bigger range of chips, perhaps bit better support in many programs.
One thing which i would consider pitfall as from my own experience, falling to the temptation to try test a board at an incomplete / intermediary stage. I had few of those and while thinking that i had in mind a clear part of schematic that should of work in my head, the board would be at a stage actually something different than what i had in mind, but differences are hard to realize between a half schematic and a half populated board. So now i would not test something unless it is 100% fully populated pcb.