That is good to hear Adrian - A few questions:
How do you manage your Robot libraries, that was the big thing that always confused me, creating keywords in libraries and making them searchable/documented always felt like a missing piece of tooling requiring better text editor support to ever feel comfortable with on account typos merely frustrating. Do you have an Atom or a VSCode plugin that scans your libraries for keywords?
When I used robot (over a decade back), I was moving from C and C++ into being a script writer, and I found it incredibly limiting. I struggled most with the almost COBOL like structure of statement-list coding style, which was a big jump from OO language. Libraries felt like hard work to build. The amount of work to switch between adding a keyword to my library and then write a step, then write the next keyword was so much context swapping it often made my head spin. Am I alone in having this experience of this process-automation tool?
We used to write tests that ran on two machines at once (RPC). One machine was a serial connection, the other was TCP, and the serial connection always ends up being slow. Has anyone managed to get robot to do nice remote (over RPC) parallel execution of steps working. Something I rely on a lot in Python in order to save time mostly is to make an RPC call to 2 machines and wait for them both to complete. I could never find a clear way to implicitly make “role” based testing that did not require me writing a library that did not hard code things like which RPC socket/port was doing what.
I’m hopeing this helps other people, but also that it clarifies the problem of deciding when a specific tool it the right tool for a job. I only used robot for a year, but I liked so much of how it worked that I incorporated keyword-driven into the tools we wrote in my next job, but we added a layer to handle the remoting and role problem.