In the early 70's (of the last century), I was Plant Chief Engineer for a company that manufactured pneumatic tools. The term, "reverse engineering" was not that common. It was mainly thought of as stealing someone else's product. It was sometimes called, "back engineering". Today, reverse engineering is used to copy or duplicate a product or object and often used to improve on a product.
Companies that supply spare parts for products that are no longer available often reverse engineer them. Antique car parts are a prime example of this. Often a part is broken and a replacement part cannot be found and must be made. An engineer will take that broken part and fit it together so that a drawing can be made and the part machined or fabricated. But, that's not all there is to it. Subtle mistakes can be made when reverse engineering a part or assembly that are easily overlooked.
Back at the pneumatic tool company as Plant Chief Engineer, I reverse engineered many tools. These had no intellectual property attached - their patents had run out and were public domain for anyone to copy. One such product was an air motor for a pneumatic nut runner. In it were two end plates that the rotor and vanes were sandwiched between. Every tool that we obtained from the field had a very small indentation that the vanes and rotor ran against. We went to great pains to replicate these parts in exact detail, including these end plates. This small indentation was difficult to reproduce since it was only a couple of thousandths of an inch deep and was exactly the diameter of the vanes in which they were spinning. For years the parts were made per the drawing that showed the precise depth and diameter of these indents. One day, we got a hold of a brand new air motor and eagerly took it apart. What? The indentations that we meticulously added were missing! What we had done by reverse engineering the parts exactly as we had received them was to manufacture years of wear into brand new parts!
We laughed at our ignorance and it became a running joke that we were designing brand new pre-worn tools.
I was about 26 years old at the time and realized that what we reverse engineered was much more than measuring parts, taking readings on a Rockwell hardness tester or examining surfaces. I got so good at reverse engineering that I could make an educated guess whether the designer had half-eaten hamburgers stuffed under the front seat of his car or whether he was Mr. Tidy Bowl that arranged every pencil in his desk pointing in the same direction. It was easy to tell if I was working with an Imperial Inch guy or a metric designer. Mr. "Burger under the seat" required having to look more closely at what purchased parts he used or the type of threads he used in his design. Mr. Burger didn't care if he designed on increments while Mr. Tidy Bowl laid everything out in even millimeters or perfect decimal equivalents of fractions of an inch. Psyching out designers became a quirky habit of mine.
One caveat with interpreting and reverse engineering a design is knowing what tolerances to use for base dimensions. This is where common sense and understanding the intent of the design comes into play. Did the designer intend for a particular part to be press fit, push fit, slip fit, running fit and so on? Were the parts originally made on the high side or the low side of the intended tolerance? Did the inspector pass a rejected part that you shouldn't even have in your hand? This is where engineering experience has to take over and measurements taken more lightly. For example, a press fit for a 1/4" nominal hardened steel pin is typically line to line to .0002" or sometimes as much as .0004", depending on the material. After a pressed pin is removed, or a product has been in the field for some time, this tight tolerance can be measured where the hole has been wallowed or galled to a larger size. Common sense and engineering experience is necessary for a reverse engineered product to work as intended. Things like bearing bores can be measured a bit larger than they were intended and plastic parts can warp over time. In essence, just measuring surfaces isn't enough. One must understand the function and intent of a design.
There is a common misconception that reverse engineering using a scanner can replicate parts exactly like the original. I wish I had a dollar from everyone that believes this. There are buzzwords like NURBS (Non-uniform rational B-spline), Laser Scanning, SLA (Stereolighography) and SLA Files, desktop printing, point clouds and so on that are interwoven into miss-perceptions about reverse engineering. Some believe that simply because they have a part in hand that it can be scanned and a machine will exactly replicate it. This technique may be acceptable if the part is a Ninja Turtle that's going to set on someone's desk. It's also acceptable if the part doesn't require a precision fit or, in most cases, it doesn't require high strength or a certain machine finish.
We've been reverse engineering parts and products for decades. Scanning parts can bring an object into a computer with relatively close replication. If that part requires a precision fit to other parts, these surfaces must have an associated detail dimensioned engineering drawing indicating these high tolerance surfaces. Scanned models simply cannot produce a crisp parting line required to produce the core and cavity of an injection mold. These sharp corners are required in order to eliminate flash, the oozing out of plastic from between the two mold halves. I hate to pop the bubble of those that believe scanning a part will make another one just like it. Don't get into a tizzy about what I just said without reading further.
On the other side of the coin, scanning parts, and in particular, injection molded parts can be used as a guide for post clean-up. This clean-up entails accurate measurements of critical surfaces needed and possible re-building of areas in order to construct a CAD solid.
Without engineering experience and not knowing the intent of a part, reverse engineering by merely scanning may only get you close. In the realm of engineering and manufacturing, that's not going to work. Proper experience with reverse engineering techniques will make or break the reverse engineering effort. A whole lot of time and money can be flushed down the drain by thinking that all one has to do is scan an item and viola, a desktop printer will make another part just like it. That is, unless all you need are Ninja Turtles and artsy sculptures.
Before you use a scanning service or hire a CAD draftsperson, or use a good CAD designer to reverse engineer your part, product or assembly, make sure they have a track record of successful reverse engineering that made it to production.
Without this experience, the tool maker or fabrication shop personnel and the "reverse engineerer" may go back and forth many times wasting a lot of time and money.
Reverse engineering is more an art coupled with experience than simple methodology.