Yes, I was intentionally ignoring some elephants.
That said… if I understand your “vector” theory… if you take 4 characters, one of whom roofies girls and promotes good causes, another improves the world while pursuing their genocidal quest, a third saves worlds while animating the dead, and the last commits war crimes while, uh, fighting for what they believe in, I guess… there exists some balance of their actions where they reach the same point, and, depending on the specifics of that balance / where that point lies, that point could lie within the range defined as good, neutral, or evil; thus, they could all be described as generally good, neutral, or evil, depending on the balance of their actions… and they could also all be described with something pithy, like “light and dark war within them” or something.
Regardless of whether I’ve read/Interpreted that right or wrong, the point remains, I still hold that developing the heuristics by which you a) assign numbers to those vectors, and b) convert those vectors into a descriptor of the individual’s “goodness” (which, on a second read, I suspect your implementation / conceptualization of the spell isn’t actually doing; that’s more for our benefit) is much harder (and much more likely to promote philosophical disagreement) than simply recording “gave cats thumbs: +1,000 evil”, “funded research into modeling falling damage by dropping puppies from various heights: +2 evil”.
Or maybe the implementation of the “know Alignment” Spell is different. Maybe the spell just (almost literally) paints you a picture, simply showing you the vectors, and leaving the caster to interpret them. A whole forest of little evils and one big good? A whole forest of little good deeds and one huge evil? What that means is open to the caster’s interpretation. And, over time, society learns that different patterns are more suitable to certain occupations. Employers reject candidates based on their “art”(“I’m sorry, but this job requires you to interact with all sorts, and you’ve shown a disturbing predilection towards strong good acts.”), and information privacy becomes a hot topic (“Steve is the only one to gain a new, large evil vector around the time of the murder” (or, if lying is evil, testing Alignment doubles as a “defect lies” spell, making trials disturbingly efficient)).
Regardless of the implementation, I still feel that writing down “+2 good” seems the easy part of the process.
Like… choosing that good and evil cancel out when producing what you dub a “high level descriptive alignment summary” is a choice. One which (for example) 3e determination of which plane a soul goes to does not follow, iirc. That uses a base logic that includes “7 evil acts guarantee an evil plane” (iirc).
Another choice for the “high level descriptive alignment summary” could be that it simply returns the direction of the largest vector (or “neutral” of no vector is larger than X). This would produce different sets of “high level descriptive alignment summaries” than one that worked based off additive logic (and would probably never describe any of my 4 sample characters as neutral, regardless of the balance of their actions).
But, assuming a simple sum was used, and you were simply transitioning from a trinary return value of good / neutral / evil to a something more informative, up to returning the actual current sum… it’s not any harder to implement the code for the new “Know Alignment” Spell. It’s still pretty trivial code. If that’s all you’re saying, I agree… but feel that the *choice* of additive vectors is itself nontrivial, and not guaranteed to be a de facto implementation at all tables.