Thanks Anthony! We do need this kind of discussion. So libcloud right now is an healthy project that continues to improve, mostly with external contributions. Which is nice! However it's difficult for me to keep up with those pull requests, so for any other developments we would need more reviewers.
I think we should be careful about asyncio. It's likely to become the next urllib or httplib: useful, but much worse than existing third-party projects. Indeed, it was set in stone when added to the standard library, but the async/await keywords changed all the best practices. So it's not clear that asyncio is the future of Python async, http://trio.readthedocs.io/en/latest/ seems more promising and more natural to use for example. I'm actually contributing to an async urllib3 port to asyncio, twisted and trio, while keeping sync support for Python 2 and Python 3, using a "bleaching" approach : see https://github.com/shazow/urllib3/issues/1323. Hopefully this will be ready for requests 3.0, and at that point the community will have learned how to make useful async + sync libraries. Until then, I don't believe we should work on making libcloud async or even async-only. Hope that helps, Quentin
