I'd given the mining thing a bit of thought over the years and here's what I'd prefer to see (and others have already touched on). Mining is currently imbalanced as the only penalty for poor mining skills is speed and lack of drops - and sometimes you *want* a lack of drops, so that seems odd. And there's a conservation of mass problem that everyone notes. I don't really like the fewer drops because of this.
I'd prefer to see rock replaced with three items: boulders, rock, gravel. A legendary miner would produce only boulders. Envision them as quarriers that can remove an entire block of stone at once. A boulder could be broken into (for example) 4 rocks (4 rocks would stack in a stockpile tile), each of which is ¼ the volume and ¼ the mass of the boulder. And a rock could be broken into 4 gravel (16 gravel would stack in a stockpile tile) which is similarly ¼ the mass and volume of the rock. Masons could break boulders into rock and rock into gravel and would do it on-site. The stone stockpile would hold all three items, and you could set it up so that when the mason breaks the boulder, the 4 stone remain in the tile and are accepted into the stockpile - no extraneous hauling needed, unless you want it.
Miners at lower skills would produce primarily gravel (and a lot of it). You could clear rooms, but would not get much usable rock in the process. This seems appropriate. As they skill up, they'd produce increasingly more rock (1 rock, 12 gravel; 2 rock; 8 gravel, etc), and then increasingly more boulders, but always the same ultimate volume. In the case of metal ores, they only ever produce 'ore', which is the size of gravel (16 drops per tile) and gets processed as it currently is, with each ore producing a fraction of a bar of metal. This would open the possibility of furnace operators gaining skill as well - by extracting more metal from the ore as they skill up - but 16 ore would at the lowest level produce one bar, as it is now, but a legendary operator could get 4 bars, maybe. Gems would work similarly, but they would drop some variable number of gems (1-4?) and the rest would be gravel - but again 16 drops per tile. The number of gem drops could be based on miner skill. Coals would drop 16 ore as well, and each ore would do the job of a piece of coal now. Flux stone would need to be converted to gravel to work in reactions - again one gravel per bar in the reaction. So in the case of coal and flux, you get 16x the output as you do now, which seems reasonable - coal mining should be much more desirable than it is now as it was a critical industry for millennia, but doesn't produce enough output now to be sought after or efficient. The hard thing should be the metal, and there you're limited to one(ish) bar per tile.
A more interesting approach to ore would be for any gravel to produce a small amount of a variety of metals - maybe based on the ratio of ore squares on the map. If you have a lot of gold, the generic gravel would produce more gold. If you have a lot of hematite, it'd produce more iron. But you'd only get 1% of a bar max or so per gravel. That way, if you didn't want to exploratory mine, but were digging out a vast area, you could turn your output into gravel and slowly convert it into metal. That'd be a fairly big departure for the game though.
As to mining speed, I would argue that legendary miners can't work any faster than novice - but they produce more desirable results. If you just want to clear a room and don't care about having material to use, then draft the fortress and mine it out and then deal with the gravel. If you want material, then use legendary miners - and if you want it done quickly, well, you better have a lot of them. Mining should be hard. It should take a lot of time.
In general, blocks and furniture would require boulders to produce while crafts would require rock. You should get more crafts than floodgates from a given amount of stone, and perhaps two blocks per boulder. Gravel would in general be useless (except as noted above) but could be used for road building and construction. I'd like to see road construction require two gravel for each block (with wheelbarrows, it wouldn't make the task any longer.) I think it's fine if it's just a pure waste product, though. Once you get to legendary miners, you won't produce any that you don't explicitly ask for.
I would also argue that the mason job be split in two - into stoneworker that does all of the workshop work, and the mason who breaks stones and does construction, and that construction done with rough stone take notably longer than construction done with blocks. There should be a benefit to setting up a block industry. Having masons run between workshop tasks and construction tasks with no way to designate one or the other is frustrating.
Now, this would all have been impossible before the hauling fixes. With the changes to hauling, the amount of hauling from this goes up only slightly. Legendary miners create almost no more work (one boulder per tile, or possibly two jobs for gems. I assume one hauler would shovel 16 ore into a cart and treat it as one job). Unskilled miners create as much hauling work as legendary miners (where now they create almost no work). It requires many more miners than currently, however, and would create some additional unskilled mason work if you need to break boulders into rock for crafts. But hauling was 50% (at least) of our labor before. With these changes, it should go down to 20% or less, and I would argue that what we gain in hauling should go back to balancing out other parts of the game. It's just odd that in the amount of time it takes one hauler to put a seed away, my legendary miner can clear out a dining room - and yet that dining room would take 5x longer to smooth. Gravel and rock could be carried by hand so you can clean up without an infrastructure, whereas boulders would require a wheelbarrow or cart.
What we get:
More realistic mining results
Better gem/ore/coal/flux output and balance.
Greater penalty to using unskilled miners.
Slightly more use for masons.
Greater challenge in the early game getting furniture/block production going and clearing space. A big fortress will require a big workforce.
Greater challenge to making an efficient hauling setup.
I'd propose something similar for the wood industry as well. With a tree producing one log. One log is needed for furniture similar to the boulder, but could be broken down into 4 boards for use with crafts, etc. One log would produce 4 charcoal, ash, etc. to keep coal mining more productive once you hit a seam and to reflect the greater mass/volume to stone over trees. Because woodcutters gain no quality benefit for skills, they would get faster with skills. For construction, because you have a 100% success rate even at novice with woodcutting over mining, and because woodcutting can be done more quickly with practice, your best bet to getting an early fortress going is wood rather than stone. Once you've gotten minimally situated, then you'd probably move over to proper stone working.
Finally, if we gain volume calculations, I like the idea of making movement through full tiles (full of stone, gravel, bins, whatever) proportionately slower based on volume than over clear tiles. Again, I don't think this would have been viable before the hauling changes.
Not sure what Toady has in mind overall, but this is what I'd like to see as a general direction. And a number of other people have touched on these as well. Regarding the overall game balance, players that want more military involvement still need to balance that against hauling and other things, and it's not that hard to balance them. Rather than build large constructions and megaprojects, make a simple fort that produces a ton of roasts and processes goblinite like nobody's business - and it doesn't matter what the mining changes are, you can still easily run (and support) a military-heavy fortress. If you want to build that tower to the clouds, you're going to balance away your military for masons. If you want to do everything, then the challenge comes in how to make everything work within your resources. Typically the limitations within the game aren't the number of dwarves or even resources in most cases - it's management. It's keeping production chains running, designating building and mining, and dealing with the various issues that occur - like moods and when invaders screw up your production or kill your soaper. I have some legendary experience with that, I can tell you.
Changing things like the frequency of mining drops isn't going to radically alter the game. It shifts the resource balance a bit, but doesn't do anything (good or bad) about the management issue. So don't sweat that too much as a resource shift doesn't change the game, just the pace at which certain things happen. The hauling changes massively shift the resource balance (enough that I'd argue other things need to get harder) and probably has a neutral effect on management - adding management around the carts/wheelbarrow while removing it from the sheer logistics of the job and some of the stockpile changes (most of us design fortresses around hauling more than any other thing which is another management impact, which would be nice to change). But adding depth to mining without adding management should be a net plus if we can find a way there (which is what I was aiming for above - particularly with all gravel being able to be turned into metal, and some of the other tweaks around coal, flux, etc).
And part of the game is invoking your own sets of guidelines to shape the game you want. If you want a military challenge, then don't turtle up underground with a massive trap gauntlet to hide behind. Instead, build a functional town where there are no remote-control levers to lower the gate are at the gatehouse and where everything is building destroyable and where fliers/swimmers can path inside. Even with 200 dwarves that'll be a challenge in the later game. Or if you want that challenge along with the economic stuff, turn off immigration. Getting 7-10 dwarves to do all of your production and prepared to fend off a couple of gobbo squads once they start dropping kids is pretty damn challenging. But even in that case, outfitting your fortress with steel isn't too hard if you make good use of the trade caravan.
Oh, and hey all.