Gorobay's questions first. For the questions that I didn't address in this post or the next one, I agree with Lancensis.
What about differences in biome?
Consider making a more generalised creature, and giving it extra biomes. Some biomes, like mountains, deserts and tundra need all the variety they can get, however.
If the only significant difference is biomes, then yeah, I'd just do a generalized creature. But if they have other differences, different biomes would definitely factor into the decision.
Terrestrial versus arboreal?
? I'm not sure where this would be an important distinction.
It doesn't matter in DF yet, but it's a big enough difference IRL that I'd take it into account.
Should extinct creatures be considered [MUNDANE]?
Hmmmm. Up for discussion. I would say no, though
I'd say yes. The only creatures that don't have it (in vanilla) are fantasy creatures like gnomes and beak dogs. Extinct animals aren't particularly fantastic.
Should parrots have [UTTERANCES]?
Heh. Well, wild parrots wouldn't know human speech. They could be given [SLOW_LEARNER] however, and they would slowly devolop conversationalist skills
Yeah, it would be fun to give them SLOW_LEARNER and CAN_SPEAK. I think the current version supports conversations between dwarves and talking pets.
Can a creature variation delete certain body parts? I know it can swap existing parts (like QUADRUPED with HUMANOID in the example), but can it swap a part with nothing? This would be useful with legless lizards, so they can [APPLY_CREATURE_VARIATION:LIZARD] and then delete the legs.
I think they can, yes
I'm doubtful. If they can't, though, it's not a really big deal -- you can just remove the BODY tag and add it back from scratch. It isn't ideal from a templating standpoint, but hopefully we won't have to do it often.
Where will we keep the new [BODY] parts that there surely will be?
Just in a new body.txt file
Yeah, although we'll need a little infrastructure in the wiki to facilitate that -- taxonomic groups and creatures will need the ability to define new bodies alongside their template variation and raws, respectively.
This also touches on the broader question of how we're going to approach anatomy. Broadly speaking, we should probably aim for anatomical accuracy on par with the creatures in vanilla DF, but the specifics will be tricky.
How will we assign symbols for so many creatures?
Convention is just the first letter of the animal's name. Vermin tend to have their own symbols.
There's another convention of using vaguely taxonomic symbols, e.g. F for feline. I think that'll work out much better for us, both for ease of identification and ease of assigning the symbols (since we can do it in a template), but there'll be problems either way. ASCII wasn't really meant for this.
How do genetics work? Are there dominance, codominance, and carriers?
This is the most I could find:
"Right now it'll do dominant-recessive stuff with the color variables, and it can also do that with the appearance variables or do averaging. It's pretty simple at this point, but what's there now could be used to, say, breed your dogs toward certain colors and body dimensions, as well as whatever facial features there end up being (though the colors and patterns would be limited to whatever is in the raws)."