Re: System Developement: A Gather of To Do's
Playtest, playtest, playtest, and don't even bother waiting for a complete draft of the rules or a full party of players to start playtesting.
I am making my own system as well, and even though none of my test games actually started gameplay (yet), the questions and feedback (mostly questions though) brought to light so much about things I had forgotten, or things I thought were clear but weren't, or even things that turn off players or confuse them.
I.E. I had originally made humans strong (a bonus to strength) because humans are, in real life, strong for their weight, but during playtesting it became clear that it is better for humans to be considered the norm (mechanics wise), even when they are not.
Also, consider for a moment what you want to support in your players. Classes make creation easier or faster, and give inspiration to those that need it, but they also limit creative options. I.E. I want a fighter who attacks with swords and bows, but uses magic in a support capacity, invisibility, speed boost, walk on walls, etc, but DnD and PF make it very hard to make the concept work.
So classes don't support interesting creative characters, but they do help if you want to support a certain combat or dungeon delving style requiring certain abilities to be always covered by the group (aka the style of play that requires those people asking about what role they need to play)
Some good advice is to decide on what you want the players to do in the game, then to make mechanics that enable and reward them for doing that.
For example, I want to support creative thinking and creative characters, so I have no classes in my system, and abilities tend to be open-ended rather then specific, all to enable and support creative use, but it comes at the cost of being easily gamed during combat if the GM isn't careful. Thus the game will hopefully do well in balanced gamelplay, but not so much for "kick in the door and hulk smash kill everything!" style of play.