Python3 compliance

[mortlind]

It looks as if the code base could be immediately made Python3 compliant. From a fast read-over I only picked up a "print" statement in client.py which breaks Python3. If you wish, I could go through the code more thoroughly, and report back, or add a pull request if I can immediately update the code base. Please let me know.

[erensezener]

Thank you for your interest. I don't think it is very important to make the project Python 3 compliant at the moment. I think the XML parsing library could be an issue as well.

Have you used the project to control a robot or do you plan to?

[mortlind]

Not yet, but I am planning to use something like your project in the future. I believe we will be heading towards KUKA. Currently we are using robots from Universal Robots, because they have a really good real-time performance. However, they seem uninterested in supporting real-time interfacing in the future. The big question is, what are the characteristics of KUKA RSI 3 regarding real-time? Specifically, when we were using RSI 2, we measured a 120 ms control delay; the time it took from sending a given set-point on some axis and until the position had been opbtained. 120 ms delay is way too long for our real-time control applications. Do you have any experience with this delay over RSI 3?

Regarding Python3, I will port your code if I find it necessary, or see if I can use Python2 in isolation from my other code bases; for which I prefer Python3.

[erensezener]

As far as I remember the delay is 12 ms in RSI 3. I am no longer working on this project, other members of the lab took over. They will be the maintainers as well.

If you edit the code for Python 3 and test it, feel free to send a PR!

[mortlind]

12 ms is the cycle time. We measured 65 ms delay over RSI 3.1 in a quick test in the lab. I define control delay as the lag of the reported position trajectory after the commanded position trajectory. I will probably take a look at using your code some time early next year.