I have inherited C++ "class spaghetti" code, which was - in fact -- easier to read the disassembled compiled code than the source code. Because the compiler was

I have inherited C++ "class spaghetti" code, which was - in fact -- easier to read the disassembled compiled code than the source code. Because the compiler was often able to prove and inline what actually gets called -- making the code logical (and relatively easy to follow) whereas the source code was abstract nonsense. "spaghetti hierarchy", common in C++ and even more so in Java, is in my opinion and experience, much worse than "spaghetti code" of the old Basic/Fortran and very-early-C days -- the old "goto" spaghetti was hard to follow, but at least every goto named a concrete target. In a spaghetti hierarchy, execution jumps every 2 lines (with actual actions sparsely sprinkled among those lines), but to determine where it goes - you have to keep track of which class/subclass every object was actually instantiated in, and what methods that class/subclass overrides.