[freehorse]

> But HID had not been designed for the possibility that packets could get lost

Doesn't the same problem exist for USB dongles with proprietary RF protocols?

Logi Bolt is a good solution. But ime most other USB dongles are terrible. I have had a lot of bad connection issues with such USB dongles, and never with similar bluetooth devices. USB dongles also use the same 2.4GHz band, and even more they are prone to interference from nearby active USB ports [0]. If you have ever had a "jumping" mouse while transfering big amounts of data through a port neighbouring your mouse's USB dongle, this is likely the reason.

[0] https://www.usb.org/sites/default/files/327216.pdf

[hmry]

The proprietary protocol can use absolute positions between device/dongle, and then the dongle can translate to relative positions at the edge, by returning the difference since the last poll

[Findecanor]

Precisely. That is how I would have designed a wireless mouse protocol: using wrapping counters and sending the counter values. The HID protocol does not support an input value that is Absolute/Wrap (although it could be extended to do so, and I think that it should)

I'd think it would also be possible to get around congestion problems by using tricks such as multiple channels and/or interference detection on top of BLE. But only Logitech knows how Bolt actually works.

[freehorse]

Is position estimation from the signal that accurate for that?

[tehbeard]

They don't mean the absolute real distance between dongle and mouse.

They mean the mouse communicates an absolute position (relative to some arbitrary 0,0 the mouse decides upon) instead of a relative direction.

Dongle can then take latest coord packet and diff it against previous coord packet to get a relative coord to pass via HID to the system.

If the RF packets are lost, some latency occurs but the dongle still has the previous mouse coord and can make a fairly accurate correction once a packet gets thru (get's from A to D, but might skip points B+C).

[pwg]

What happens with that "absolute position relative to some arbitrary 0,0 picked by the mouse" when the user picks the mouse up off the table/pad/etc. and repositions it (i.e., they hit the edge of the pad and now "re-center" to continue moving left (or right) on screen). The mouse loses its 0,0 point reference as soon as it is picked up.

It could send a "reset 0,0" packet of some form in this case, but now reception of that packet becomes critical to continuing to properly communicate motion to the attached computer.

[hmry]

That's not how mouse input works though, right? If I move my mouse cursor to 10,10, and then pick up the mouse and set it down somewhere else, it's still at coordinates 10,10. You don't need the mouse's physical absolute position, but just the cursor position (which is the sum of all the relative movements)

[pwg]

> That's not how mouse input works though, right?

No, mouse input data does not work that way.

My reply was referring to @tehbeard's suggestion that the mouse could be modified to send absolute coordinates instead, and I was pointing out a reason why that modification would not work out so well.

[justsomehnguy]

> It could send a "reset 0,0" packet of some form in this case, but now reception of that packet becomes critical to continuing to properly communicate motion to the attached computer.

And those "how I would have designed a wireless mouse protocol" guys are back at the square one.

[freehorse]

That sounds like a software problem to me, not one that requires a hardware solution. There is nothing in what you describe that cannot be performed through bluetooth packets.

I am not sure which dongles make these corrections, but my experience with dongles is worse than bluetooth. Typically, a mouse is very close to the bluetooth antenna of a computer, and I have not really experienced any sort of connection issues due to missing packages etc. In contrast, I have had tons of issues with usb dongles due to usb interference.

[numpad0]

No, mice sensors are far from being SoTA linear encoders. They return approximate instantaneous movements somehow when read request is received. Mouse controller chips(USB or RF or Bluetooth) pack and report up the movements, that's it.

(psa: none of Chinese ADNS-2610 clones have the raw pixel output debug command. Maybe security implications or maybe something else, either way, mouse-as-microscope hacks don't work on sensors extracted from e-wastes)

[frollogaston]

In all the years I've used wireless dongled mice, I've never had an issue. And all my stuff is bottom of the barrel unbranded from eBay or Amazon.

Bluetooth mice and keyboards always have trouble pairing, or there's input lag, or sometimes I can't use them to wake the computer. And if you ever want to hold a startup key...

[numpad0]

Nintendo Wii Remote used almost completely bare standard HID over Bluetooth Classic and it just worked flawlessly. It needed Wiimote aware apps and special procedures for use with non-Wii devices but it had none such problems(I've had those too) in gaming. It could wake up consoles too.

The problem is in 5W1Hs between combinations of (Windows, PC, Bluetooth), not the protocol. How should pairing keys be retained, etc.

[frollogaston]

Yeah the cross-platform situation makes things hard, but in a way it's the protocol's fault if it's hard to agree on things.

Also, I really remember Wiimotes failing to pair a lot of the time, but it was so long ago that I could be wrong. All I know for sure is last time I was at someone's house and they wanted to play Switch, it took like 10 minutes to get the controllers set up, but most of that was bad UI.

[frollogaston]

Clarification, the dongled mice are cheapo, the Bluetooth mice are expensive and still bad.