Stephen Early

(It looks like I mentioned this in #60 but only the first of the two warnings got fixed.)

Not currently. We don't have a sequence for assigning addresses to control devices, we don't have implementations of the various parts 3xx of the standard, and we don't know whether...

The short answer is "it doesn't". The library isn't a complete system, at the moment it's more a component of a build-it-yourself system. There are lots of different ways to...

> I don't think we should talk about a "winning" API. Indeed. I don't want to mandate a particular API; I see the drivers as an "optional extra" to the...

Not currently, but the serial protocol described at **https://atxled.com/pdf/AL-DALI-HAT.pdf** looks reasonably simple so driving it shouldn't be too difficult.

I don't have any information on the USB protocol used by that Helvar interface (they don't appear to publish it, and I've never had access to a sample for reverse-engineering)....

Ok, so it's showing up as a generic HID device. To work out what the protocol is, I would install the DIGIDIM Toolbox software on Windows in a virtual machine,...

Sorry, yes, that's unclear! Yes, I mean the interface between whatever the code making use of python-dali is running on, and the physical DALI bus.

There's still an issue with dependencies. You unconditionally import `serial` in the driver.base module, which means that (for example) someone using the `driver-unipi` extra and not the `driver-quaddali` extra would...

The CI test does pass now, because it ignores extras and drivers.