It is somewhat amazing that it works at all.
Good: The front panel is very nicely done, although documentation of controls is minimal. At least the part on the tab control. The controls/indicators way off in far corners of the panel need to be brought "home" and either used appropriately or removed.
It appears that there are about 40 controls on the front panel, mostly booleans, to which the program should repsond promptly when the user changes a value. This immediately suggests a Producer/Consumer (Events) architecture.
It already has separate loops for the Modbus communication and DAQ processes. Those should also be incorporated into a carefully planned Producer/Consumer (Data) architecture.
The data transfers between loops should be handled with queues or User Events, not local variables or property nodes.
****
The way I would start is to go through the front panel. Write down every control. Add a description of what the program does (or should do) with the data from that control. Write down every indicator. Describe what the data represents and what part or parts of the code generate that data.
Next write down every connection to the outside world: Files, Modbus, DAQ, mental telepathy, and whatever else is in there. Describe whether the data is input or output, what it represents, where it is generated or used in the program, and any specific timing or format requirements.
Now write a description of what the program should do. This description might be only one paragraph but not longer than one page. Omit details but be as specific as possible.
Get the boss or the customer to agree that what is in those written documents is what they really want to program to do. If they do not agree, then work to make the changes needed to get an agreement. From that approved set of documents you can now begin to design the program.
***
Now that I have seen the program, I think a rewrite is still the only realistic option. But it may not be as horrible a job as I first thought.
If it does not work or the user/customer/boss wants big changes, then the estimate is of no value.
Lynn
