I’ve heard multiple times that apparently a lot of proprietary Bluetooth stacks are filled to the brim with bugs. I’m not sure how complex the Bluetooth spec is but I wouldn’t be surprised if it were true.
It’s also part of why I can’t stand the trend of remove the headphone jack from devices.
I've read chunks of the bluetooth spec and while I didn't wind up spending very long actually using it, the vibe I got was that of a committee composed of architecture astronauts who pumped out pages by the truckload and applications engineers who could never agree on anything. The standard is very heavy on abstract cruft and very light on constraining decisions, which is probably why 30 years later we still have pairing issues with headphones.
BLE actually fixed that and it uses less energy. I have used it both on desktops and on embedded devices, in C and in scripted languages, and it has been a pleasure.
The original design was utterly terrible. A low data rate frequency hopper with long packets. Which meant the RF performance was terrible and non state of the art. By design. I think they thought spec'ing low data rate frequency hopper meant automatically low cost, low power.
They messed up the low power part by spec'ing a complicated baseband that had everything but the kitchen sink. Basically duplicated the USB spec but then added authentication and encryption on top of it. Which is even more complicated because everything runs over a unreliable frequency hopping rf channel. Frequency hopping messes with a lot of things.
It’s also part of why I can’t stand the trend of remove the headphone jack from devices.