I suspect that on the issue of derating, there really is not a problem. Looking at the DIN rail terminal strips, these are probably lighting and control circuits operated at well below their ampacity. There is almost certainly a violation of the code in terms of cross section fill and derating, but IMHO there is probably not a significant conductor heating issue here.
On the shear messiness and difficulty of debugging....arrrrrrgh.
This particular issue has been discussed several times, usually with respect to a trough feeding a bunch of emt and several panels. There might be hundereds of conductors in the trough. With any complex system, you either have to spend lots of time making sure that all of your breakers are in order relative to your incoming emt, or you have to have some sort of 'crossbar' where you can match the innies and the outies.
I've though that this presents a market for some sort of 'matrix trough'. Imagine a trough say 24" high by as long as necessary, with insulating pillars or bridle rings, say on 2" centers. Each row or column would define a 'wireway' that would be tested and _listed_ as having a capacity for (say) 20 current carrying conductors without derating. (No, I've not done the physics, and don't know what the capacity would be.) Wires would be separated throughout the entire 'trough', and so easy to get at, heat could be dissipated, etc. Complete circuits would be required to sit in the same 'wireway', to reduce loop area, and there would probably need to be some limits on the number of splices.
Another approach (this one suggested by the picture of the panel with the neatly bundled wires) is some sort of 'stacker' where individual conductors are held in neat _separated_ rows, where each conductor can be easily found and traced.
-Jon
