I know about transistor radios, never heard the term transistor battery.
Notice my "" around the 9V "capacitor". That's an important detail. Batteries actually pack some punch. A battery doesn't get it's energy from thin air. You have plates and a dielectric. In order to supply current, a battery must have charge (Q). The equation is Q = C * V and derive from there.
A fish diagram is what we use when taking chunks of a system drawing and represent what's hanging off the car electrical system. Look in a car electrical manual and you'll see why we have to do that since things can get crowded when viewing so we break things up, etc.
I've attached a simplistic drawing. Here's how the system is at risk using a 9V hold device. The battery is hooked hooked up and at 12.8V. 9V battery is attached with a diode to prevent the car battery from charging and destroying the 9V. The diode to the 9V battery is reveresed biased. All is well. Joe tech removes that car battery, does his thing. Radio presets are held. Battery is put back in car and hooked back up. When the battery is connected, the system is at the 9V (or there abouts) and the car battery is at 12.8V. That is a deltaV. So where does it go??? It's doesn't just magically disapper as you have Ic = C * dV/Dt to contend with (remember how a battery is a capacitor?). That is a surge and the current has to go somewhere until the diode can shut off (it's not instantaneous). In our designs we try and protect for both negative and positive spikes which is exactly what happens on the system. If a spike is too quick than you can cause damage to the input of devices connected to the 12V system.
Keep in mind semiconductor devices are subject to degradation with power cycling, temp cycling, changing moisture levels that the mold compound sees. Also, a lot of stuff is relayed on and off the electrical system for reason such as noise, loading, and transient protection reasons.
I am NOT saying 9V device = automatic damage. I am saying there is definite risk. Those transients can be quick. Faster than the proection devices can respond. How many people have damaged something when they jumped a car (correctly) or accidently shorted a screwdriver across the battery? Trust me when I tell you that you can damage parts of the car that can be latent issues causing parts to fail early or not as intended. Six months later when something goes because you whacked it with a surge you'll never even think is was from that time you jumped the car, etc. We have a non PC phrase we use when this condition occurs.
Lastly, I have attached Littlefuses (not the company I work for) arsenal of automotive electrical protection devices. You asked why there are not widespread issues with car electronics going bad, this is why. A car electrical system is quite nasty when you put an oscilloscope on it. The protection devices you see in that document are everywhere on you car because people plug in things like 9V memory hold devices
Yes, the techs are being saved from themselves. In the end it's up to you folks to decide how to work on your car. I'm just trying to warn people that the potential to damage components in your car with a 9V battery gizmo (even the one's you buy or diode protect) exists.
I design power devices designed for automotice and appliance applications for a living and I would never use a 9V memory hold doo-dad.
Carry on. Last post on this topic.