Emily Love Watson

I don't think it's a great idea to do this on the premise of performance. The real underlying problem is just that our recorder is really slow, and if we...

> But I think that we should not aim to build a full data warehouse into our recorder. Care to elaborate on this? I'm not sure I understand why we...

Alternatively, if we really want to store completely de-normalized results for each query we support, we should probably not try and fit it into a relational database, and look for...

As we're discussing this, I'm thinking we should consider NoSQL document store options such as MongoDB, etc. It seems like we have a use-case that's exactly suited for those types...

Yeah that's possible, and I'd be open to a slightly fuzzy equality check myself, maybe we keep it if it's within 1 or 2 seconds of the previous value to...

Just a thought: What if the hass backend always returned standardized metric units, and we made unit localization a frontend concern?

Link to where the dependency was introduced, for reference: https://github.com/gentoo/gentoo/pull/34447