Hi! We’ll be getting an xArm 6 next week, so I’m thinking of spending some time testing the current integration on real hardware and potentially contributing a few additions.
A few things I’m currently interested in are:
- teach-point / coordinate capture and replay
- native trajectory recording and playback
- linear rail support
- fuller Bio Gripper support, especially grip/catch/slip detection
Before I start working on these, may I ask if anyone is already working on any of them, or if any of these would be better handled differently / are outside the intended scope of the current xArm integration?
Happy to coordinate so I don’t duplicate ongoing work. 
1 Like
tangental, I recommend looking at what we have been doing with the pf, it’s getting really nice
1 Like
I went through the PF implementation and it was really useful. one question for the xArm6 work: do you think it is better to keep using official SDK, or should we implement the communication through pylabrobot.io?
.io would give us byte-level capture and validation for testing and troubleshooting. but replacing the SDK transport would also mean more code for PLR to maintain, and it may be less reliable than the official SDK.
I’m leaning towards keeping the SDK behind a small async wrapper, perhaps with a small capture layer later if needed (an optional PLR capture layer that records SDK operations or protocol frames only when capture is enabled. communication still goes through the official SDK, not through pylabrobot.io.Socket. tho idk if this is really needed since SDK might be reliable enough). what do you think?
