As a process engineer I got to see design engineers design impossible stuff.
The only way to create one weldment was to have a dwarf welder permanently weld himself inside.
Everyone who designs anything has done something like this on occasion, whether they have an engineering degree or otherwise. Can't always see on a piece of paper or flat computer screen how it's going to be in reality. Solid modeling that will run on a computer that anyone can afford helps ... didn't have that back when I was designing stuff. Still doesn't help if the operator doesn't have a clue.
My theory is that everyone has a limit to the number of pieces that they can visualize going together in their head. We've hired (and subsequently fired) mechanical designers whose limit was one: they couldn't get two adjacent pieces to have the same bolt pattern on them. When this was pointed out to them, they'd just change it on that drawing, without addressing the repercussions downstream. Arrrgh ...
Been there myself ... Management was worried about how a certain workpiece was going to be clamped in a machine that we were building, which was my design. I wasn't worried about them, but that's another story. Management insisted on having extra clamps designed in (everything around this area had already been designed and most of the parts were already built). I drew them up but forgot about a certain major un-movable part of the machine that would be sharing the same space, and this wasn't discovered until the shop tried putting it together. Phone call, "get down here and have a look at this, what do you want to do?" Answer ... "Leave 'em out!" Ended up working fine without (as per the original design ...)
Sometimes it's an organizational deficiency ... We once subcontracted out the mechanical design of a heavy-duty conveyor system. Biggest mistake ever. I (project manager) had a certain date that I had to commit the locations of all the footings in the concrete. A few weeks before that date, the inquiries went out: "We need to know where you want your footings". No response. No drawings. No reaction. A few days before, the request was sent again. No response. The day before, "Tell me where you want your footings by date XXXXX, otherwise I'm going to tell you where they're going to be." No response. I spent that day drawing up where the footings were going to be, with no knowledge of the detailed design of that conveyor frame, and on date XXXXX, that drawing went out to the concrete contractor and to the conveyor designer.
Only after that, did I get a phone call complaining that they had to redesign the conveyor frame to fit the footings. "Too bad. Too late now. Deal with it."
That part of the project worked out in the end, but did it ever look messed up. Not much else worked out ... this was symbolic of the whole mess on that project. Ill-defined specifications, overly ambitious goals, committing to a project knowing that there were things we didn't know how to do, not enough money in the budget, not enough real estate, you name it. Company subsequently went bankrupt, but I was outta there before that happened.
I still see designs that are half-baked and not thought through properly, or in which the design that was implemented has been done for political reasons (to keep someone happy) even though it's not the right solution, etc. Most of the time the designers know full well that it's not right, but it's being thrust upon them by others, "thou shalt do it this way".
