When examining an existing design I try to get "into the head" of the original designers, putting myself in their shoes facing the same design questions. This often helps me understand why they made certain choices.
For example, the top side of the original PCB in the Canon P170-DH calculator uses what I think are carbon ink traces rather than copper. Likely this is because carbon traces can be used as contacts for the conductive pill keypad and left exposed to air without corroding. However, carbon ink traces have significantly higher resistance, which explains why the traces were made so wide.
One of the oddities of this board is that the power supply and printer driver components are mounted on the top of the board, while the calculator electronics are mounted on the bottom. Initially I thought this had to do with fitting into the three-dimensional space available in the shell, and so I did the same with the design of my replacement PCB. In reality there is at least 0.65 inches of space between the PCB and the shell, and the tallest component is barely 0.55 inches tall.
One morning this week, as I sat in traffic, a more likely explanation occurred to me. All of the components on the top of the board are through-hole parts, while all of the components on the bottom of the board are surface-mount parts. All the soldering is done on one side of the board, which makes for easier assembly.
All of the components I'm using are surface-mount, with the exception of a big filter capacitor and the connectors. If I was starting this design from scratch I'd give serious thought to putting all the components on the top, and the conductive-pad switch contacts on the bottom. Although KiCad has the ability to visually "flip" a board, it doesn't reorder the layers (yet!) which makes routing a bit more challenging.