I heard from a colleague that typical PLC serve as an interface between a the signals and, for want of a better word, a PC. The connection, being typically done with ethernet, and the PC runs some Codesys, or Labview applications.
This was the way the application could be changed while running, which is one of main advantage of PLC for debugging and maintenance.
However, looking at one example, like this one, I can see the Ethernet ports, but I can't find any precise information about that topic.
Looking around, I found that a remote PC can be used as a server and thus the PLC would communicate with it, typically via Ethernet. But it is not a must.
Can someone clarify that trivial point for me? Must or must not a PLC be connected with an external PC to run the application. And if it's not, can we still change the code on-line?
For the background, I have to decide on using a PLC or a µC-based solution, and up to a few days ago I ignored the existence of PLCs.
I should add that I have read the question: Why use PLC instead of microcontroller?, but unless I missed it, it does not answer that specific question.