Now that I've covered all of the major subsystems, let's get into some of the smaller details.
The basic mechanic: Why have 2 6-siders counted separately? Once I had divined the "size matters" principle for the combat system, the 1-5 size scale suggested itself almost immediately, along with the 1d6-low mechanic. The second die came in as I decided that I wanted to make success easier for younger players with fragile, developing egos. Patronizing of me, I know. It wound up correcting for itself, anyway. Now, if I want to give a character a roughly 30% chance of success, I set the skill level to 1 instead of 2. a 50% chance comes at skill level 2, not 3.
The contest and simple contest mechanics were in place before the 2 die system, but they converted over fairly easily. I originally used a "lower is better" style for contested actions, but it was pointed out to me that a "blackjack" (as high as you can without going over) method would require less math, which made it immediately more desirable.
It also created a viable "crit" mechanic in the form of double successes. I am rather fond of those "woohoo!" moments in a game where the dice let you "bat it out of the park." I had toyed with a few ideas for such a system with the one die, including something akin to the "confirm the crit" technique from D&D. The two die system allows the crit to be established with a single roll, speeding up play.
Templates: One thing that I wrestled with for some time was how to construct characters.
Not just a trait scheme, but how to populate that scheme. Combining attributes and skills was simply not viable with the scale of game that I had. So I looked to a couple of games already on my shelf. Notably "Dream Park" and "Pokethulhu", both of which featured traits that were not quite attributes, but were a bit more than skills. I came up with the final list by averaging out the lists of the two games and removing the fighting abilities (Remember a few posts back? "Not everyone is a fighter." I knew you could).
I decided rather quickly that I wanted a fairly closed character creation system. If I went too open, players were just as likely to recreate characters from their D&D game as they were to try something Ozzy. As I read the stories, I realized that the protagonists of the stories fell into a few major categories. These became the concepts that I would build templates around. Another upshot is that Dorothy becomes much more viable, if only to create party variety. Also, since the Dorothy template (Child in Oz) is based on the same number of skill points as any of the other templates, she should have just about as much to contribute. Especially with the high Wits rating that the template provides (Dorothy had quite a lot of perseverance and Trot was nearly fearless), a Child can stand up to just about anybody.
The Scholar template was one of the last additions. A commentator on RPG.net noted that a number of my templates were "skill paragons", such as the Child in Oz template
being the "Wits paragon." But I didn't have a "Brains paragon." Reviewing my source material, I realized that a Brains paragon was highly appropriate. There was the Woggle-bug, who participated in the adventures in "The Land of Oz" and the Frogman who was featured in "The Lost Princess of Oz." Although both of these characters were very unusual, there was nothing so strange that a standard template wouldn't suit for either of them.
Character Advancement: There really isn't any. Dorothy is much the same character she was when she first arrived in Oz. While she doesn't accumulate powers and skills, she does accumulate friends. Thus the Friends List becomes an implicit means of character improvement. There are some exceptions, though. The Wizard of Oz began as a humbug wizard, but later began to study under Glinda and learned real magic. Since he doesn't use the stage magic after this, it could be argued that he "traded in" his Humbug Magic skill for true Magic skill. A more final version of the rules will mention this option a bit more explicitly.