Reconsidering Typica's configuration system in light of some new capabilities that I've been experimenting with. I already like what I ended up with by the end of 1.x better than what I've seen elsewhere, but I think there's a lot of room to improve on feature discoverability, expressiveness, and making it harder to mess things up.
A lot of software in the space has fundamental issues here. In most cases, tuning the signal path just isn't a thing. Someone picked your settings and that's what you get. It's simple, but the person who picked your settings has never actually roasted a batch of coffee and what they picked is extremely suboptimal.
The other class of common issues would be that someone had a clear vision of what they wanted and came up with something sort of sensible for that, but then more features got bolted on around that and nobody ever went back to reconsider a more general approach so if you're not doing things exactly like you could have back in 1.0 things get very strange very quickly.
Some day I would love to have a chance to work on a line of coffee roasters where all of the details are optimized for joy. No cutting corners to shave the BOM, all the modern amenities, fast tactile controls for everything, integration with surrounding workflows, everything tuned to make it easy to get the coffee to do whatever it was you intended to do with it and repeat that reliably. There's only so much I can do strictly in software with no budget.
If anybody out there is working on that and needs an internationally recognized expert on coffee roasting who has extensive experience developing software for coffee roasting operations, feel free to reach out.