My fortresses do remain fairly small (I don't enjoy managing more than a hundred dwarves, and I can barely employ that many no matter how much stuff I have them do), but I still don't see how you could cover the surface with just a fortress, unless you regularly clear out z levels of rock for no particular reason, or engage in extreme exploratory mining that reveals every square. Underground megaprojects might do it, but megaprojects
should involve vast expenditures of labor and careful planning, and modding could aleviate the problem anyways. Very large fortresses don't usually mine out more than a few z levels of total area (I think... Does anyone have actual numbers for total mined area in a large fortress?), and if rubble is made to pile nicely into pyramids with ramps and stuff, as per the sand suggestion, that would easily fit in probably not much more than a single embark tile (haven't bothered to do the math, anyone interested?), even with a 1 mined tile equals 1 tile of rubble heap system. Unless you're in a 2x2 embark, that shouldn't be a big deal for most people. If it is, find a different place to dump your rubble.
I think one of the easiest ways to do it would be to have 1 mine tile equals 1 ramp, and 2 ramps equals 1 wall of rubble, but it wouldn't stand up without ramps beside it and would collapse back into 2 adjacent ramps. Yeah, as you pointed out, people could micromanage their rubble dumping zones to an extreme degree, and turn a wide hallway into a slightly narrower hallway with rubble piled at the sides, but this would be so intensive on player labor (very much unlike just letting your dwarves automatically haul it to the tailings heap) and so damn ugly that I don't think many people would bother. People already order the stone cluttering their fortresses dumped, despite it having absolutely no consequence except for increasing construction time slightly and being ugly, and this would be even uglier. Toady could make the default color for rubble that ugly brown used by most sedimentary rocks. Few players experienced enough to pull off that sort of micromanaging of dump zones would tolerate that besmirching the walls of their fortress, blocking access for engravers and covering the lovely marble.
Despite being for rubble, I don't want mining to be slowed significantly, with some qualifications. I think mining should go at current or nearly current speeds, so long as you have adequate removal of rock. During the founding of your fortress, this shouldn't be an issue. Assuming the typical 2 miner embark crew, you have 5 other dwarves who were probably just hanging out by the wagon waiting for the miners to do their thing anyways, and the stone doesn't have to be hauled very far. Later, when you've disabled hauling on those dwarves, and they are productively employed anyways, you might have to set aside some of your new migrants as stone haulers, and remove all other duties. Otherwise, your mining might slow, or even stop, as the miners have to crawl over rubble. I would rather that mining is not totally stopped (realism has its merits, but this is a game and I don't think that would be fun), but just slowed. Also, as you delve deeper into the earth, it's not so easy to drag the stone all the way back to the surface, especially if the haulers are getting tangled up in narrow mineshafts, causing traffic problems. It might be time to do a bit of careful civic planning, and dig a real city, with wide high traffic routes that go around your main fortress areas and provide easy hauler access, or a nifty minecart system that goes from the mines to the fortress and the tailings pit. Ideally a minecart could carry more than 1 tile of rubble (ignoring all the physics involved with a minecart taking up 1 tile of space, this is a necessary concession to gameyness), and would be of significantly different weight when laden with rubble than when ladden with ore or quarried stone, otherwise you might have to dig two railways.
I like the mining versus quarrying idea. It has appeal from both a gamist and a simulationist standpoint. After all, if you want to dig a hole through rock with a pick, you're probably not going to be producing many boulders, you'll just get shards of rock. Whereas if you want to get a big block of rock out of a rockface, you will be doing a lot of slow work with chisels, and it will not be a very quick way to dig a tunnel.
EDIT: about the sluicing thing: found more, used for lots of things, and pretty old (the earlier sources seemed to claim it was invented in the california gold rush, which seemed a mite fishy to me). Especially great for tin and gold mining, but used for all sorts of situations where you have mixed ore and rock, and the ore happens to be significantly denser. This isn't used for just ordinary rock, and would not produce anything useful from it. Ore does not come in nice 1 tile units of pure cassiterite (at least in
Dartmoor), it comes in thin veins running through granite, which can be separated easily using what is basically a sluice, but in Dartmoor at least was a slightly sloped ditch dug in the earth, allowing water to wash away the granite leaving pure cassiterite. This doesn't mean that the gabbro you dug out of your dining hall can be sluiced to produce metals, it means that the vein of cassiterite you discovered must be processed to be smeltable at all. I think that would be an interesting change, especially if sluicing weren't abstracted into a building but required an actual channel of flowing water, but it would ultimately be extremely frustrating to newbies and a great many other people, and would not be a good addition. Perhaps ore veins could have fuzzier borders, with pure veins in the center, like it is now, but with a thick layer of mixed rock and ore that could be set aside for when you get around to setting up a proper sluice channel. This might make lower mineral abundance more bearable.