Great question, janetfan. You raise two points I want to address.
For me, the "harshly logical (but not complete)" is the way the sport is heading, theoretically. How do I mean? I mean edge calls get penalized. If you UR you get hammered. If you achieve a certain difficulty level, you get rewarded for it, no matter how good/bad it is.
I think Rachel Flatt vs Mirai Nagasu at Nationals will go down as the classic example of the dichotomy I speak of. If one read the boards and gauged the enthusiasm, they'd come away thinking that Nagasu won and won easily. But that's not the case. Intuitively (how most of an audience operates, no matter how knowledgeable) - Nagasu won. She's more aesthetically pleasing. She's bendier. She's got the right body shape. She didn't fall or make "intuitive" errors (stumbles, step outs, etc). Her errors were logical. She didn't rotate fully. Flatt did. She got points for difficulty if not quality (the anti-Lepisto). Flatt won.
Hell, even the way falls are judged fits this schism. What is a triple flip? It's a jump with three rotations "in the air" that takes off a backward inside edge with a toe pick assist, and lands on the backward outside edge of the opposite foot. (stealing from wikipedia). If you UR, you don't complete the first part of the requirement. If you take off from a poor edge, you don't do the latter. The fall? Doesn't affect the fact that it's a triple flip. So the jump gets penalized for quality, the fall gets penalized again (the deduction), but the triple flip stays a triple flip.
The not complete factor? Well, ISU doesn't want to go the whole hog. It doesn't want to say well, that lutz had an edge call, therefore it's not a lutz, therefore not an element and doesn't contribute to the TES. And it's actually more logical to give more credit for a UR'ed triple than a solid double (base value is assigned by difficulty, and a UR triple is more difficult than a double).
janetfan, your other point is a little more troubling, and COP simply will not fix it. And I don't mean they'll ignore it, I mean it's simply impossible to fix.
Calling from one event to another (both from judges and from the tech panel) is not gonna remain the same. We saw it this year, we see it all the time. Why? Well people see things differently. GOE has objective parts and subjective parts, and we see that when we see a string of grades that range by three levels. Hell, even when there's a fall, we see outliers (Patrick Chan's free program at Worlds, for example. All but one of the judges gave him -3, as we all think they should. One judge gave a -2, presumably because he/she felt that the insane transition leading into it should get a mark). Is it deliberate? I can definitely understand suspicions; I agree with them (see the GP threads). But even it weren't, there isn't much the ISU can do, I think. Two people view things differently - one person's slow is another's Speedy. In extreme cases, you'll see agreement (for example, the Olympic FD judges giving V/M straight threes for their lifts). But check out the top ten at worlds in men's FD. How many times do the judges agree when no fall is involved? Out of 130 elements, they agree.... once. People aren't robots, but COP requires that for judging.