Combo-layer hybrid keymap

As a programmer and avid user of a 32 key layout (see keyboard), I’ve spent quite a lot of time optimizing my symbol layer and how it’s activated.

There are two main problems with symbols on low-key-count boards:

1. False negatives. Regardless of how the symbol layer is activated (HRM or dedicated key), the activation must occur before symbol access. Why is this a problem? Well, if you spend enough time coding, you get really fast at typing symbols. This rapid layer switching will often result in false negatives because, unless you have unnatural consistency, the symbol key could be pressed prior to the activation key.
   
   
2. Combos are not a perfect solution. Typical fix to problem #1 is to map every symbol as a combo. This is exactly what urob has done and people like it quite a lot. It’s not a perfect solution though because we have common symbol pairings like +=, -=, >=, <=. These pairings are best mapped as inner rolls for speed and comfort; not really possible with combos.

What I’m proposing is a hybrid approach. The concept is very simple: symbols are mapped on a layer and each symbol also gets an equivalent combo (with a pseudo-mod-key). How does this work? When you type a single <, you hit the combo. When you type <=, you hit the combo for < but keep holding the pseudo-mod-key and press the = key. The combo works as expected, but while the pseudo-mod-key is held, the symbol layer is active.

Why am I calling it a pseudo-mod-key? Because it works similar to a dedicated mod/activation key, but the activation of the layer is only available through a combo + holding action. Every symbol is mapped as a combo with the same pseudo-mod-key (per layer). The timing does take some getting used to, but after a couple of days I’ve found it to be very fast and natural with virtually zero misfires. This approach also works very well for num layers.

This neat functionality can be achieved with a parametrized ZMK macro:

macros {
	clh: combo_layer_hybrid {
		compatible = "zmk,behavior-macro-two-param";
		#binding-cells = <2>;
		bindings
			= <&macro_param_1to1>
			, <&macro_press &mo MACRO_PLACEHOLDER>
			, <&macro_param_2to1>
			, <&macro_tap &kp MACRO_PLACEHOLDER>
			, <&macro_pause_for_release>
			, <&macro_param_1to1>
			, <&macro_release &mo MACRO_PLACEHOLDER>
			;
	};
}

And we use the macro in our combo binding (example for = and key #14):

//define SYM_PSEUDO_MOD_KEY and SYM_LAYER somewhere

combos {
	compatible = "zmk,combos";
	
	combo_equal {
		bindings = <&clh SYM_LAYER EQUAL>;        //use macro here
		key-positions = <SYM_PSEUDO_MOD_KEY 14>; 
		layers = <DEFAULT>;
		timeout-ms = <40>;
		require-prior-idle-ms = <150>;
		slow-release;
	};
}

See my repo for a clean implementation. My personal keymap is unorthodox but in principle any keymap layer can be converted to a hybrid one.

Personal keymap:

• cross-hand layers for symbols
• I also use this hybrid approach for the num layer
• my pseudo-modifiers are on the home-row.
• I use urob style HRM’s for everything else

Thanks for reading!