>> An electric only latch would reduce parts count

I disagree. Dozens of parts are needed simply to bring an electric wire into a door. Every penetration needs grommets and strain relief. A flexible weatherproof system needs to get the wire from the door to the rest of the car. Relays. Fuses. It all adds up. I cannot believe that a manually-actuated lever would require any more parts.

There are two kinds of engineers in the mechanical and electrical departments. One says "If we just run wires / automate / motorize it will be easier", and they see the direct path from start to finish that matches the requirements.

Then there are those who say "every wire complicates compliance testing, every automation complicates failure modes, every custom part complicates supply chain", and they are the ones that prefer in-house mechanical parts with conservative materials and tooling.

I think it's an age thing, maybe it's a hipster vs greasy-elbows-shop-floor guy thing, but honestly it can be hard to tell who's right.

There's another angle: system independence.

When things go wrong, i.e in a crash, you still want to guarantee an undammaged door latching mechanism can still be operated independent of other parts of the car being damaged (likely). To satisfy this for a purely electrically actuated latch you would need a battery and independent controller in every door.

It's hard to beat the reliability of a mechanical latch and a human hand to satisfy that requirement.

I think it's mostly an issue of experience. Younger engineers tend to make things more complicated than they need to be and are unaware of reliability issues. Older engineers better understand tried-and-true systems and are more likely to want to stick with things they know work well.

Obviously I'm speaking generally and individuals vary greatly.

>> Every penetration needs grommets and strain relief.

You mean the grommets and strain relief that are already present for windows/speakers/locks? In most cases it's probably not even additional wires as power and control (either can bus or lin bus) are already established from the door to bcm (body control module).

Sure but in most cases, there's ALREADY wires going to the window motors and locks. In this situation, removing the mechanical door handle/latch operation does indeed simplify things.

You already have power in the door, because a car without power windows is a non-starter these days. And power locks.

So using the electricity already present for the door handles has to be entirely tempting.

Thay come with that jigsaw puzzles anyway. Manual release is required.

>>Manual release is required.

Is it?

Happy to be proven wrong since I'm far from an expert on FMVSS, but what I've seen specifies capabilities like ability to lock and amount of force it must endure, not the actual implementation (mechanical/electrical/etc).

Regardless the article only tells us the "emergency override mechanism did not work". Whether that is electrical or mechanical we can only guess (one would hope mechanical).

TIL, [0]:

  FMVSS No. 206 does not directly regulate either interior or exterior door handles. It does, however, impliedly require both interior and exterior door handles or other release mechanisms in its regulation of door locks (see generally, 49 CFR 571.206, S4.1.3 and S4.4.2). 

Also UNECE R11[1] soft-require mechanical external door handles.

I think there are also other implied requirements from other sections/directives, such as in crash safety that the doors are required to be locked while crashing, unlocked after a crash, can be opened but not unlatched and flopping. Meanwhile, electrical systems would be required not to pose hazards. I’m guessing those together are ruling out all but flap type manual handles, at least on front side doors.

0: https://www.nhtsa.gov/interpretations/24393rbm

1: p.23 https://unece.org/fileadmin/DAM/trans/main/wp29/wp29regs/201...

If my car is on fire I sure as heck don't want to rely on an electrical door release