Yes, I'm familiar with those tools, and used some of them in the past, I was just wondering if you built/used an actual real world trading system written with Python + a messaging middleware solution, and what were the reasons developers picked that combination instead of C++ or Java?
The place where I used to work had python and a custom messaging bus. For my own personal system I did look at python plus rabbitmq (rabbitmq is pretty neat IMHO), however I don't need especially low latency so I just write everything into databases and then pick it up with another process (it needs to be in the database anyway for historical / diagnostics so why not just use that). I think python is a more natural language to manipulate data in than c++. It's more natural if you've come to systematic trading from a trading / research background, rather than from a programming background.
Back on topic, the problem with R is it's very easy to write bad, dangerous, code in R (in particular its way to loose with type checking; and almost randomly returning stuff in different types on a whim. You need to litter your code with manual type checks and conversions. It's also really easy to write eg loops that suck up the whole memory. Okay theres a lot more I could write...).
It's fine for research but I'd be terrified to use it for a live trading system. I've run systems running 9 figure FUM written in R, and it wasn't a pleasant experience or something I would do again.