Now DDF is taking shape and new sample files keep popping up here and there I was wondering: will there be a large scale conversion to add common devices to the deconz base install, or is the community supposed to take care of this? Thinking of devices that many people use, for which the auto-generated DDF seem to work fine. I think it would make sense to include these as DDF files at a cetain point? Or is the idea to only include more complex DDF files and generate the “basic” ones on-the-fly? And will generic/basic Zigbee devices also be added as DDF (tunable/dimmable/color bulbs, smart plugs, etc.)
I created a few DDFs, but to be completely honest I have no idea what properties to include and what to configure exactly for reporting settings etc… Just trial & error does not sound like a very good idea. Some additional guidance from deconz for certain device types would help?
I think there is a gain to translate to ddf files since the legacy implementation is less optimized and thus better performance can be attained by moving to ddf
Wondering myself too. I know that Swoop is doing some conversions for his implementations.
@de_employees feel free to comment here
The goal is to transition all devices to DDF. This is gonna happen piece by piece, especially since already tested and working devices (hard coded in C++) need to be tested again to not create regressions.
The transition allows us to remove old C++ code and getting the code base way more maintainable, robust and clean.
Does that mean that batches of new “Silver” DDF files will be added to deconz, which can be tested by the community and will become “Gold” in later builds when are confirmed to be working as expected?
The process itself is challenging
Lot of “Silver” DDF are here since months but who could confirm that all possibilities offered thru DDF are used, clearly tested or optimized ? On the other side, many DDF have been delpoyed in “Gold” for facility without more test than the one that need some parts of it in third party app.