Dan Allan

For reference, we had previously discussed adding "write retires" into ophyd. This is, IMO, a better way to achieve the same result. Putting the retry logic that far down hides...

Yes, the example above is lightly adaptive from one in this section of the tutorial: https://blueskyproject.io/bluesky/tutorial.html#add-metadata

> I see routine operations of the beamline via notebook as all risk and **negligible gains** [emphasis added] I think that is key: as someone who does not like to...

Thanks. @cryos and I also hit this issue recently but I may not have got around to creating an issue to track it. Should be simple to fix.

Looking at that final error, ``` RuntimeError: Another set() call is still in progress ``` I think this is a duplicate of #1102, which concluded as: > This is not...

The memory footprint of an object in Python is not a very well defined question, but we can track total process memory usage over time and the lines of code...

@prjemian: Abstract away from count time, exposure time to "time for measurement". Make it the responsibility of the ophyd Device to handle this rather than pushing it into plans. @danielballan:...

@prjemian: If interrupted mid-stream, human intervention is required. Automatic recover is not possible, too likely to result in undefined / unintended behavior. That would just be an unavoidable risk of...

All sound like good ideas to me. Seems like we should make the absolute/relative thing configurable too.

Yes, I think that is sensible early on, just as we currently package a bunch of unrelated drivers together in `databroker._drivers` with the intention of splitting them up once things...