5 ms·
That kind of generalizing isn't helpful and isn't true, either. I am sorry you didn't have a "Bluetooth anything" that worked reliably. I did. In fact, everythi
by jwr 2y ago
That kind of generalizing isn't helpful and isn't true, either. I am sorry you didn't have a "Bluetooth anything" that worked reliably. I did. In fact, everything that I own that uses Bluetooth works very well.
There are two important things about Bluetooth.
1. There are actually two kinds of Bluetooth. The "traditional" mostly connection-oriented Bluetooth and BLE (Bluetooth Low Energy). They share very little with each other, except for the marketing name. BLE works way better in practice. Many people still remember their old headsets that used Bluetooth and that took ages to connect to their phones and associate those with the "Bluetooth experience".
2. Bluetooth is complex. Many manufacturers get it wrong. But the problem is not with Bluetooth. You could invent a different set of protocols, and if they were to do everything that Bluetooth does, they would get very complex as well, and we would have the same problem, except in a less popular and less broadly adopted protocol.
- ryandrake 2y ago> Many manufacturers get it wrong. But the problem is not with Bluetooth. To the end user, this is kind of a distinction without a difference. The end user doesn't care whether it's Bluetooth, The Standard that's bad or Bluetooth, the Implementation. He just knows that when he has to use Bluetooth, he's probably going to have problems. I've worked with BLE implementation developers and I've heard the stack described as a "Layer cake of sorrow." There are major problems with it, and by "It" I mean the entire bag containing the specification, all the various implementations, and the hardware ecosystem. All of these things combined define Bluetooth in people's minds.
- darkerside 2y agoHN is traditionally the kind of place where unpacking that nuance is rewarded, not dismissed
- ryandrake 2y agoYes, it's useful for technical folks to understand the real difference, but I'm saying the end user doesn't care.
- hedora 2y agoBut if the spec as a whole is bad, all layers of the spec are bad, the implementations are all bad, and the user experience is bad, then in what way can Bluetooth be described as anything but bad?
- jwr 2y ago> if the spec as a whole is bad, all layers of the spec are bad, This discussion is intriguing. I wonder how many people commenting on "the spec" have actually read any part of "the spec" in question (I have). And since we're posting anecdata: I can't think of a single problem I've had with a Bluetooth device in recent years, and I use a lot of Bluetooth devices. So this kind of generalizing doesn't help. As a counterpoint, every embedded device that uses Wi-Fi promises a world of pain. The bizarre pairing procedures, connecting to temporary access points, entering passwords — it's all a combination of pain, timeouts, problems, and resets. But that doesn't lead me to state that "wifi is bad, the spec as a whole is bad, all layers of the spec are bad, the implementations are all bad".
- scarfaceneo 2y agoYes, but if we’re discussing solutions, maybe it would be relevant to understand where the problem stems from
- catlikesshrimp 2y agoSometimes it is the hardware, sometime it is the software. Two xiaomi phone models with different BT issues: One couldn't connect to two chipolos, the other loses wifi stability when connected to A2DP (ping raises, connection stutters and as the BT devices get closer, connection is lost randomly) Samsung phone: no issues [detected so far] Then I have some Sennheiser headphones that I can get stuck and need to be physically turned on / off when it loses connection due to distance and there are other phones nearby (I roam around the house without the phone. I know that's partially on me) I remember the time there was for windows a "BlueSoleil" BT stack besides the "Broadcomm" stack. Blue soleil was more stable, supported more profiles (pan, a2dp, etc) on more devices. And, all those BT security issues over time? It makes me feel BT protocol cockroaches always come back. Apple has no excuse since they control both hardware and software. They are dropping the ball.
- thfuran 2y agoBut saying that Bluetooth is fine, it's just that every implementation is broken is a bit ridiculous. Sure, there's some value in distinguishing between irreconcilable issues with the standard and implementation bugs, but if those bugs are ubiquitous, the protocol as a whole is broken.
- Dalewyn 2y agoThe moral is: If everyone hates Bluetooth, the problem aren't the haters. Personally, my experience with Bluetooth has been mixed. It certainly works, but audio quality is bad for headphones, latency is bad for mice and keyboards, sometimes the connection requires an intervention from the Norse pantheon, and so on.
- thenthenthen 2y agoMac and bluetooth… surely a mixed experience. Most issues seem to boil down to my devices connecting to my Macbook while it is “ASLEEP” (aka lid closed). Somehow the Macbook is very eager to take over/overpower any other device (phone, other computer) to claim my BT headphones. This can be fixed in the terminal or with the Bluesnooze app [0] [0] https://github.com/odlp/bluesnooze https://github.com/odlp/bluesnooze
- thaumasiotes 2y ago> It certainly works That's going a little too far. If you're using bluetooth headphones, and you enable a microphone anywhere, you'll stop receiving sound from everything else on your computer.
- Boldened15 2y agoAll Bluetooth technology is licensed, no? Then it's still the fault of Bluetooth if not only is their protocol so difficult for many or most manufacturers to get right, and their qualification process doesn't enforce correctness. But yeah I mean Bluetooth sucks, everyone has had the experience of trying to connect to a device repeatedly that keeps trying to autoconnect to some other random device that was used before. My usual interaction with Bluetooth is to always enter pairing mode and just be done with it, but hardly feels futuristic to manually hold down a button press every time you want to use a device. > except in a less popular and less broadly adopted protocol Isn't the usual way this problem is solved that the big players (Apple, Google, Microsoft, etc.) come together and create a consortium for a new technology? So you get widespread support in all new devices and within 5-10 years most people are using it over the old one. If a better protocol is possible I don't think this would be the bottleneck.