Hmm... Sorry if this is a little off-track, but I'm going to sort of dump a jumbled collection of thoughts I have on the subject onto the thread. I'm actually actively thinking about this very subject right now, so I have several different things to say about it.
I actually remember getting a comment from Toady, when I was making a mod that involved having playable minotaurs, (back about 9 months ago) and I wanted to make them herbivores, only to find no tag for that. He basically responded that he hadn't really made up his mind what sorts of herbivores and whatever he wanted in his game. Obviously, grazers are something that went in, but we still don't have something that replicates "forced vegans" from a dietary inability to process meat sort of thing we can shove on a race that is capable of farming.
This would probably go well as part of a broader effort to work on rewriting nutrition as a whole, however.
To just quote myself on something I am talking about in regards to a different thread...
Thing is, the amount of food the animals eat is directly proportionate to their species' average size alone. Toady obviously just sat down dividing the number 60,000,000 by the animal's average size.
Meanwhile, the animal's milk production is oddly the same no matter the creature size, making goats the best milk-producers, since they are fairly small and eat little. Cows are only useful for the meat they produce.
Now, making the amount of food a creature eats derived from its size is not a bad idea by itself, but we can make the game calculate out that part on its own, so that larger or smaller versions of the same animals eat more or less food. What should go into the raws is something about metabolism so that you can throttle up or down how much food that animal needs to eat per mass of the creature.
Likewise, the size of the cow or goat should have a direct impact on how much milk it can produce, with a variable that makes cows much better milk producers for their size than a horse is.
This means setting some standard model, and multiplying it by the size of the animal, then possibly adjusting by certain traits of the animal.
Already being strong makes animals more beefy and more massive, so there is an evolutionary advantage for being smaller and hence, eating less, if there is a strain on the food supply, while being stronger is an evolutionary advantage when you are fighting against predators that hunt you in great numbers because there is a plentiful food supply. That's the sort of evolutionary equillibrium we need.
Other traits on the animals, such as speed should also affect metabolism - slower creatures should consume less energy and hence, eat less, basically, and this should be a multiplier on the food consumption rate, as well. Cold-blooded creatures should get a massive bonus to their ability to go on little food, at the expense of vulnerability to cold and a penalty to their speed when they are exposed to cold, should that ever be implemented.
This is all talking about animals, however, and not dwarves. Dwarves need to have food requirements based upon their own mass and their own stats, as well. Even things like being more intelligent should probably have some calorie-based penalty. Basically, larger dwarves with better stats eat more. This means starting from a point of either multiplying what the average dwarf eats by about 4 or 5 times, and modifying it from there based upon stats and mass. We can also just multiply the whole amount of meat that is provided by animals that are butchered and the amount of food that is produced by fields to match what is needed for the chosen nutrition model. (The amount of "4 to 5 times" was derived from the statement that a cow's meat should be able to feed a human for about a year, and a whole lot of reverse engineering of the numbers in the Volume and Mass thread.)
A problem is we don't want dwarves taking more food breaks, we want dwarves to be capable of grabbing all they need to eat all at once. This means at least some fundamental ability for dwarves to carry more than one kind of object if the food stack doesn't already have 5 or so food items left in it. This is, basically, more stacking and hauling requirements.
I'm basically trying to work some bits on nutrition into the
Improved Farming Rebooted thread, and there's also a
Prepared Food megathread, for the "delivery" end of the spectrum.
Obviously not to say that there isn't room for a nutritional model topic all its own (and I think there have been a few), but that there are some very big and complicated overlaps going on. I pretty much agree with everything you have said, although I wonder if we really need to talk about a need for distinguishing fish-eaters from meat-eaters in general.
I'm actually re-writing the section on nutrition right now for the Improved Farming Rebooted thread, in case you're interested in that, although it is meant to accomidate some of the more advanced reaches of the farming thread's needs. (The quote comes from what I'm rewriting for that...) The point of which is to talk about making players be more conscious in their balance of fruits, vegetables, fungi, and meats, plus also having some tie-in with a waste system, so that we can model the full cycle of soil nutrient to crop to meal to dwarf to dwarven waste to soil nutrient again.
Oh, and as part of the ecological simulation, one of the things I wanted to throw in is a food web system where vermin will target (and potentially devastate) certain plants and crops, supporting populations of predatory vermin (like spiders or other insects) that eat those vermin, leading up to large creatures (non-vermin creatures) in a food web that can actually make certain animals hunt specific types of prey based upon some general characteristics.
Speaking of which, you might want to add "EATS_EGGS", since there are some snakes and other creatures that specifically target eggs.
Of course, when you're talking about things like honey, I think that almost any creature can eat honey, it's pretty much pure sugar, but few creatures actually try. It may be better to work on trying to seperate out some sort of dietary tolerances for what sorts of things animals are capable of eating versus their specific tactics and evolutionary survival strategies that make certain types of food more viable. Almost any carnivore or omnivore will scavenge if forced to do so, after all, but vultures do almost nothing but scavenging. That may be something more complex than a ranked priority system will be able to handle.
Sorry, I started writing this when there were no responses, so I haven't read those yet...