Fri Nov 16, 2012 10:57 am
I quite agree, that four out of five times a case gets in the way when developing hardware. But very often, after the hardware works, you have to finetune and debug software, and try out a device in the real world.
A completely hypothetical example: You want to create an alarm clock that wakes up not by sound but by scent. You want to hook up a scent disperser to a buffered output, maybe a light sensor. You get the hardware working, but now you want to try this thing out. Next to your bed. Because you want to know: Will the scent even be noticed? What is the right scent? How much of it? Will it affect your dreams? You can't test this on your workbench.
If you've got loose boards, cables sticking out in every direction, breadboards, etc. this will work for maybe the first day, but it won't last a week even for testing, because some day you will put your pillow on it, by accident. And then some wire will disconnect.
Purely hypothetical, but sometimes you *do* have to test devices in the field, even if a rough enclosure just for protection and bundling cables is all that is needed.