Agreed . . . but as for:

Unfortunately, the argument that, if it helps, "Why not have one?" easily can - and has - led to calls for safeties, then mechanical locks, then embedded locks, then biometric safeties . . . and ultimately a battery-operated gun. I can't remember the last time I reached for a flashlight and it actually worked as I needed it to lol.

Personally, I need a stronger argument. And for me it is simple.

Pulling the trigger is what fires the (striker) gun. A heavy trigger pull is often cited as preventing accidental discharge, and I'm sure it helps.

But that same citation also implicitly recognizes that an accidental, light trigger impact is possible . . . and personally I'm not interested in estimating how heavy an accidental trigger contact is likely to be. Nor am I interested in a trigger-mounted safety for the very same reason.

I've heard the arguments that a safety is not required, training and safe practices are the answer. Again from my perspective, a second mechanical device (ie, a safety) requires two accidents to occur. Requiring two actions (eg, 2 signatures, 2 keys to launch, etc) is not perfect, but is a well-recognized approach to preventing "accidents" while maintaining operability.

As for the safety interfering with rapid use . . . its sort of hard to argue that training overcomes the need for a safety without admitting that training should successfully incorporate releasing the safety as a routine part of your "draw".

So . . . personally . . . safeties that are easy to operate as part of the draw . . . like, eg, the M&P9 that I own . . . are positive safety features. When they are available, I will take them.

Safeties like on the SR22P (also a decocker) do not qualify for me - I tolerate it only because the SR22P is not a defensive weapon for me. The absence of a safety on my Ruger LCP is barely tolerated - I like the weapon for pocket carry, but wish it had a safety like the M&P's.

As always, YMMV.