With great magic power comes great responsbility... for Toady to makes it balanced. I believe that it will be quite a challenge for him - if few characters have too much power then they will most likely dominate the world too easily, and if it is too common than death toll will be huge as a powerful wizard will be able to burn whole towns easily if he had the power to command fire.
Fire magic is what I am most afraid of. In casual RPGs like Skyrim for that matter you can cast fireballs without much thought, but in DF one fireball can set grass on fire, and then burn quite a large territory - last time I had fire it burned the grass on the whole map and its spread is unrealistic too - it takes roughly the same time to spread in every direction forming a huge circle. In adv mode I believe it will have to spread more realistically and over a finite distance, but still, thinking realistically, one fireball has the potential to burn the whole forest with all civs that call it home.
Why is that a problem? DF is intended to be able to generate and simulate a huge range of types of fantasy worlds (including some that aren't very fantastic). In some of them, it might well be the case that fire magic is easy and/or powerful. Important structures and items will be made of stone or metal. Living underground will be popular. Shields will be important. Managing line of sight will be tactically relevant. Ways to survive without extensive surface farms will be logistically crucial. Various secondary technologies based on ready availability of high heat levels will be practical. Sounds... pretty dwarfy, really; in fact, one could argue that a significant part of what we think of as the "default" mythos of DF is actually pretty compatible to a world with too much fire in it.
Worlds don't have to be "balanced". In fact, overly balanced worlds seem fake and artificial, because reality isn't balanced. Frankly, I'd be more worried about death magic or meta-magic than fire magic; either has the possibility to destroy a world far more firmly than just setting the surface greenery on fire. (Any magic that affects the effectiveness of other magic has the potential to spiral out of control in a feedback or feed-forward loop, and that's the sort of thing that a computer will have difficulty spotting until it happens.)
(At this point, I've stopped really addressing the poster I was replying to, and am responding to the thread as a whole.)
It regularly baffles me why people seem to have extremely narrow expectations or desires for something as powerful as DF. It's like getting three wishes from an immensely powerful genie, and asking for $1,000, a Honda Civic, and a ham sandwich. Sure, there's nothing wrong with any of those things, but they are all things you could probably get in other ways, and don't even begin to explore the vast space of possibilities available to you.
For a more mundane comparison, it's like going to a huge, high-quality, all-you-can-eat Chinese buffet and just getting a slice of pizza and some french fries from that bit on the end of one row put there for picky kids. It's arguably not the best pizza, a specialized pizza place would probably do better, but you know you like pizza so why try any of the other hundreds of items? Why, there's a risk you might not like something! Ignoring, of course, that if you don't like something, it didn't cost you anything directly, and very little in opportunity cost, and can just move on to try something else.
The likelihood that my "favorite" type of magic system is one that hasn't *ever been written before* is statistically quite high. How will I know until I try a whole bunch of them? Humanity has probably only generated a few tens, perhaps hundreds, of thousands of well-fleshed-out mythos and magic systems (*); DF players may end up generating that many in an afternoon. Sure,
Sturgeon (and Kipling before him) was an optimist; but there's little cost and no reason to linger amongst the dross of bad rolls; pick out the glittering, never-before-seen gems and enjoy the wonder.
(*) I'm making the simplifying assumption that many crappy fantasy game and novel settings, particularly the sort sometimes referred to as "sword and sorcery", reduce to a much smaller set of functionally similar world settings, just with different names.