One cannot assume that players will have access to clay, this is true. If this were to be implemented I would imagine *perhaps wrongly* that it would be relatively trivial to include clay as a tradeable substance, just like sand will be in the next release.
Since starting a farm is a fairly basic thing, you would probably need to embark with clay if it wouldn't be available where you were going. Home civs never have everything, though. It's possible you wouldn't have it at all, in fact.
Sand, and hence, glass, you can live without. Farms are pretty much required, especially if you're going to make butcherable animals require feed.
We were talking earlier about how requiring water would make farming in a desert too difficult, and the argument against that was "well, shouldn't it be?" When the difference between a fort with a mighty agricultural base and one without it is whether or not they have access to clay and other types of soil, as well, it's getting a little weird.
Logs are frequently available, inexpensively, from traders. This means trading would be that much more important, which I only see as a good thing. Most of my charcoal, potash, and wood production comes from purchase logs.
Apply these same principles to alchemy and purchased items.
As you already say, though, you're already using the logs that you purchase. Logs are fairly heavy (meaning they take up much more room than other products players may request). You're saying we should put a significantly larger strain on what is already a limited resource (on an annual basis).
Likewise, if we are punishing a player for embarking when their civ is at war with elves or humans, that creates the same problem as with clay - farms are such a basic necessity of DF that it really shouldn't make farming too difficult without relying upon an embark parameter.
Incidentally, at no point did I say that logs would be required for SUSTAINING underground farms, simply building them. Perfectly happy to stick with using potash to continue fertilizing fields, in conjunction with the other elements listed above.
That's the problem with "realism" as a design goal - things that one person considers realistic often leaves in things that are highly unrealistic in other aspects, and the "realism" often leaves in severe game imbalances that are, themselves, fairly unrealistic. Still, I'm really sounding like a broken record talking about how realism alone should never be a design goal.
If logs are the fuel for building a farm, why wouldn't they require replacement? If the basis of a farm is a rotting log, wouldn't that log eventually be eaten all the way through?
This also makes subterranean farming require above-ground resources (excepting, perhaps, tower caps), which may be realistic, but I think part of the concept of dwarves currently is that they can be entirely subterranean creatures, realism be darned. (Of course, I still like the chemosynthesis based upon volcanos as a possible source of a subterranean ecosystem's energy.)
I like animal chunks being able to be used as a feed for animals, but not so much as a fertilizer. The idea of using raw meat straight from the butcher as a fertilizer doth offend my sense of realism a bit much. (Many things I can let go, this is a bit far, simply because I've worked with organic gardening and processing meat into compost and know that it takes a goodly long time to become usable).
If animal chunks are unsuitable for you as direct fertilizer, that's fine, just shove in a middle step. Have a worm farm for a workshop that can decompose bodies into usable soil. Can work for "solid waste" as well if we ever get sewage as part of DF.
Collectors may bring chunks to the worm farm, and, when they have decomposed properly (changes to miasma would be nice, so that ventilation can be worked out), they can be used as fertilizer.
---
Anyway, you should be able to find what I proposed for a system for the past few pages. I agree with a step up in complexity in general, however, I have no particular fondness for realism per se, I simply want a more involving system. (I'm perfectly fine with just having three factors of soil quality, that can just be called "mineral A" "mineral B" and "mineral C" for all I care, plus whether they had been watered recently or not.)
I don't really think fertilizer should be a required aspect of the game (although underground farming does seem to call for it, doesn't it?), and would be happy with just plain crop rotation.
I'm also a little leary of irrigation, as either it will be a matter of simply digging a couple holes you fill with water, or you will actually have to micromanage exact water levels of ditches feeding specific crops, which seems to be what you want, which will be a micromanagement nightmare if you put in crop rotations, which is something I would really like to see.
The idea of going all the way with setting up specific drainage systems is going a little far in my book - they not only take quite a bit more manual labor to set up (and farming is a basic necessity that people will want to set up quickly), but would require destroying your farms to change drainage systems between crops. (Which again, is very bad for crop rotation.)
Also, I am no real expert on medieval agriculture, but I don't really think this was acutally a technique used by the sort of culture DF is modeled upon. I believe medieval farmers largely just cleared the boulders away and plowed the land. Rice farmers certainly worried about this sort of thing, but we're dealing with a much more wheat (and mushroom) based agriculture, here.