Update soft-off.md

[GermanG]

Adding Split Peripheral Off On Press basic documentation.

[caksoylar]

I'd like to understand why this setting was implemented and defaulted to first, before encouraging people to turn them off. @petejohanson can you elaborate?

[GermanG]

I think it's related to the issue I reported today #2300

[petejohanson]

I guess I'm confused when you'd ever want to only put the central side into soft off, that you'd need this documented?

[GermanG]

While fixing the peripheral soft_off issue, I noted that the peripheral turned off as soon as I pressed the key, despite the hold-time-ms setting. Upon investigation, I discovered an undocumented option (possibly a workaround for the issue I reported in #2300). After disabling it, I encountered the aforementioned issue. I am unsure of the original author's intent with this option and can only speculate; therefore, I refrained from modifying it. Instead, I documented it and demonstrated what can be done if you prefer to disable it.

[lesshonor]

I guess I'm confused when you'd ever want to only put the central side into soft off, that you'd need this documented?

The way the docs for hold-time-ms read indicate that the keyboard only goes into soft-off mode if you hold down a button for X ms.

From what I'm reading here, how it actually works is that when you press the button, part of the keyboard totally ignores your settings and goes into soft-off immediately.

My presumption was (is) that the primary reason hold-time-ms exists is to delay soft-off so it can't be immediately triggered—so if (part of) the device is going to do the exact thing the user is trying to prevent, that seems counterintuitive enough to be noteworthy.

[caksoylar]

From what I'm reading here, how it actually works is that when you press the button, part of the keyboard totally ignores your settings and goes into soft-off immediately.

My presumption was (is) that the primary reason hold-time-ms exists is to delay soft-off so it can't be immediately triggered—so if (part of) the device is going to do the exact thing the user is trying to prevent, that seems counterintuitive enough to be noteworthy.

I agree with this part, but I still haven't figured out why split-peripheral-off-on-press was defaulted to in the first place. If it is needed for robust operation (e.g. to make sure peripheral receives the command before central sleeps), what happens when soft-off doesn't have a hold-time-ms? If it isn't needed for technical reasons, why default to it?