Maybe consider the relationship between consequentialism (theory) and deontology (practice): the rules of the latter can be considered pre-calculated shortcuts to the former. For example, "do not kill" and other commandments are widely applicable shortcuts for most real-world consequentialst calculations, though they obviously fail in some cases. An example from religious ethics: you ought to donate some of your income to charity (through church), but how much? A tithe (1/10) of your material and/or financial revenue is a rule that makes it workable in practice in many cases, without an undue burden.
Of course, with time the rules of deontological ethics tend to become "imperatives" due to lost purposes, and "practice" becomes "theory".
One of the criteria moral philosophers use to asses the credibility and power of a moral theory is "applicability". That is, how easy is it for humans to implement a moral rule? For example, a rule exists like "donate 23 hours a day to charity" it would be impossible for humans to fulfill the goal.
This lead me to start thinking about whether we want to be able to to pursue "the moral theoretical truth" should such a truth exist, or if we want to find the most applicable and practical set of rules, such that reasonably intramentaly rational (human) agents could figure out what is best in any given situation.
I feel like this is sort of like a map-territory distinction in a loose way. For example, the best thing to do in situation X might be A. A may be so difficult or require so much sacrifice, that B might be preferable, even if the overall outcome is not as good. This reminds me of how Eliezer says that the map is not the territory, but you can't fold the territory and put it in your pocket.
I'd love to be able to understand this issue a little better. If anyone has any thoughts, ideas or evidence, I'd appreciate hearing them.
Thanks,
Jeremy