Michael Carroll

> It would probably be good to test what's the benefit of asynchronous primary-secondary stepping in a more realistic simulation (not only physics), any recommended world to test that? As...

I think that you may have posted the same image twice in your histogram analysis.

This was part of some of our work. I don't think there are plans to pursue it as part of the current project, but I will land this PR.

Looks like this will also need a linting pass. I'm not sure why the Focal version isn't failing.

@mxgrey we have a whole slew of test failures here that aren't in `main`. I spent a little bit of time, but didn't make much headway.

> Let's start by removing the usage of SDF types inside components in favor of Ignition Messages. I don't think that's what we want to do. Generally, I think we...

I think based on the recommended guidance from the [Protobuf C++ Tutorial](https://developers.google.com/protocol-buffers/docs/cpptutorial), we're going to end up with another collection of objects one way or another. > Protocol Buffers and...

> Is SDF serialization performance significant? If it's being serialized to XML and back, I would say that the performance is not great. > Will the serialization issue be resolved...

> The reason to use SDF is to avoid ABI issues, follow guidance from the protobuf developers, and leave open the door for improving our cache performance. Unfortunately, most of...

> I know it sounds backwards not to use contiguous storage anymore, but as detailed on https://github.com/gazebosim/gz-sim/pull/856, that had been hurting performance, rather than helping it. I think you could...