IamNotgod was referring to a different issue (equal weighting of all parts of the map, despite inconsistent difficulty). Last I heard, Tom was interested in having pp consider "per-Hitobject" data in order to solve that, but it's not currently possible (and I don't see how it would be possible to get accurate data on that from old scores).
Personally, I think HD is the biggest "issue" right now. HD does deserve a bonus, but it doesn't belong in aim and it doesn't belong in accuracy (it also doesn't belong in speed, but it doesn't give a bonus to speed right now). The accuracy bonus in particular seems silly to me - HD gives a 2% (or something like that) bonus to accuracy under the assumption that people will mess up at reading a map when playing with it. Apparently playing with HD means we're supposed to be playing maps we can't read well and playing nomod means we're supposed to be playing maps we can read perfectly. I don't get it. The 18% bonus to aim...well, that's just really misplaced and inflated and it's even more egregious on slider-heavy maps. If and when pp incorporates pattern/reading difficulty, all of that should be solvable by just throwing the HD bonus into that instead. That could also help with flashlight - flashlight is crazy hard, but the primary difficulty with it has nothing to do with aim, speed, or accuracy.
Speaking of flashlight, its bonus should probably scale with the player's max combo instead of being a flat 36%. Flashlighting a 50 combo map is easy. Flashlighting a 1000 combo map is very difficult.
In fact, maps with high max combos feel underweighted in general. I don't know by how much or what exactly should be done about that, but long maps that are difficult throughout could use some love imo.
Overall, I really like pp right now. I feel like it for the most part rewards good plays - exactly what a ranking system should do. It makes it feel worthwhile to sometimes play specifically for ranks, which always felt worthless with ppv1.
Something that happened earlier seemed very wrong to me. I had the following plays in order:
Seemed really strange to me, especially considering what
osu!tp has to say about the relative values of aim, speed, and accuracy in that map. Is combo weighted considerably lower in pp than in tp or is there some other factor pp takes into account that I'm not considering? Or did my pp for it not calculate immediately (as of 2 hours later, it still hadn't changed, so that makes this a bit less likely)? It would be convenient if there could be some sort of indicator on the score screen that tells you if your pp for the play has been computed and added to your total (or, better yet, as has been suggested, something that tells you exactly how much raw pp your play was worth).