Reverse-engineering is basically from down to up (from fine details to abstract) engineering while "normal" engineering is so called forward-engineering and is from up to down (from abstract to finer details). Reverse-engineering is basically the first part of the re-engineering process.

If you have a ie. traffic light controller PLC with few hundreds lines of non-commented code and you make some efforts to get the logic behind the program because you want to improve it. Then you are basically making reverse-engineering to get the system re-engineered.

I'm basically interested how that reverge-engineering could be efficient. I know that there is plenty of how-tos and more theoretical studies from the subject in PC programming world, but i'm not sure if there is any usable one in automation field.

-LV

Hi LeoV.

1) First i would start from is IO signals. You can find out what sensor connected to what input - by schemes or by experiment (also ensure it's safe if your machine is working).

2) Then you can give names for IO areas.

3) Then find out which variables depend ONLY from already known variables - and name them

4) Repeat (3) till you get all variables names

5) Check code by comparing blocks of all projects you have - and define which version is the latest (best code from your point of view)