Back in the 1990's (I think) Chrysler wired their stereo's into the vehicle's CAN bus. Problem with that was that when the radio failed, the whole vehicle shut down and needed a tow back to the dealer. Seemed like a good idea at the time.
Not only is it harder for DIY auto repair because of the proprietary nature of the components, but because these components interact in complicated ways.
Only legislation requiring manufactures to publish technical spec's on their components (and their component's intended/expected behavior) will improve the situation for the consumer. Thankfully we have the internet and youtube.
Here's a counter to your proposal for using inexpensive microcontrollers....
We were driving our 2000 Pontiac Montana minivan one day when somebody started flashing their lights at us from behind. We asked what was the matter and they said our brake lights were out. That was strange...in 40 years of working on cars, I had never had BOTH lights go out at the same time. Funny thing is that the high-mounted center brake light still worked.
I replaced the bulb on the right side and...NOTHING!!! Same on the left. We called the dealer and they laughed saying, "Oh that must be the controller board." Funny thing that, for over 50 years, a switch was good enough for controlling brake lights, but NOW you need a controller.
It is almost always "easier" to stick in a micro controller and put in some code than to do any actual engineering. No question about that. BUT the unintended consequences of code that it is unable to handle exceptions, or does not handle the way some users think, and is generally not exactly the best choice. That is the fundamental flw that I see in using software to implement many functions. Primarily code is only written to handle what the programmer believes will happen, which is usually a very small subset of the real world.
William K: I have to disagree with you there. For one thing, there are at least 50 microcontrollers in a car today. It is often much simpler to put in another microcontroller than trying to integrate software within a central processing unit. Since many of the components come from other manufacurers, or even other divisions of a car company, this makes sense. As for cost, even simple electromechanical are very expensive. My wife and I recently replaced the automatic door lock actuators in a minivan. I think they were between $50 and $100 apiece. And we installed them outselves. I was a fun project. On the other hand, these were simple relays. There are many reasons for the high cost of automotive parts, but not having access to the code is not one of them.
Inexpensive micro controllers will not ever, at any time, under any condition, improve automobile anything! BUT they will certainly increse the profits gained from replacing the failed parts, since the code will be unavailable to anybody except the auto companies repair parts makers. While emissions and economy have been improved through electronic controls, all of the other car functions have been rendered less reliable and nonrepairable by the use of them.
Recall that a few years back it was predicted that automotive electrical systems would go to a seriel buss and just a pair of power wires, in order to reduce the weight and cost of all those wires. BUT what we have now is a few central modules and huge bundles of wires. We also have a few accessory modules for those extra-priced expensive options, which only become more expensive each year.
An unfortunate number of things that sort of sound like good ideas are actually some vendors idea in search of acceptance, instead of something that actual customers would have any desire to implement.
Putting the bad switch back IN was a waste of time except as confirmation that it was mechanically bad. The failure was in a very small bounding box containing only the switch and its connector and the vehicle's harness side. Going to the dealer for a $20 replacement was also suboptimal since almost every vehicle has $8 aftermarket replacements available at any parts store.
I'm a volunteer tech specialist on a Chevy Trailblazer / GMC Envoy forum, and we get these sorts of questions monthly. Brake lights and ignition switches are well-known high failure rate items.
But I have to comment on the microcontrollers changing automotive manufacturer's design goals any time soon. The capability is there for plug and play peripherals on everything except vehicles. Install a new intelligent driver's door switch module (that dies over time because the switches are horizontal where rain and snow falls right on them with every time you open the door), or a 4WD transfer case control module, and you THEN must go the dealer, who has the "magic" Tech II tool, to download the personality firmware into the module. Minimum charge for this 5 minute process? $75-200!
GM and the like will NEVER do something to decrease the revenue stream for their dealer network for outyear support by assisting the DIY repair owners. I get more and more bitter when I see design decision made that thwart DIYers in order to drive the owners back into the clutches of often-abusive dealers. I know it's not the designer's fault - they are under management orders. But it's an economically-driven cycle, not technical, and therefore not open to the usual rational discussion we would have about other products and systems.
Just LOOK at the insane number of Delco/Delphi connectors and ESPECIALLY the little safety catch/latch mechanisms that thwart mechanics and diagnosticians at every turn. Irrational.
I had a minivan that I put many miles on doing field service work. Got to where the cruise control would not stay engaged one 11 hour trip. When I got back home (another 11 hours without the cruise control) I started poking around. Wiggled the brake switch and things behaved for about 2 weeks. Then I had a problem with intermitent brake lights. Finally replaced the switch. Turns out it was a few years of accumulation of crud and minor arcing that finally led to switch failure.
Solder resist/mask is not a mechanically robust component and is not designed to provide any mechanical isolation. As its name implies, it's designed to prevent solder from going places you don't want (e.g. wicking from pads).
If you cannot confine your enclosure's contact with the PWB to purpose designed areas of the board (e.g. a mounting pad), you need a mechanical component like an insulated washer or a coverlayer of kapton or equivalent.
On a tangent, for higher reliability, you should avoid via's (or plated holes) under mounting pressure anyways since the thru hole barrels tend to crack under the compression.
The specific occurrence was a small array of plated thru via's (to the main LCD flex ZIF) anchored through the board with annular rings. The annular rings were <.015" O.D. but completely exposed. Meanwhile, the transceiver module had a die cast zinc outer casing and laid directly over those via's, flush to the PCB surface.
There are nearby areas of other printed circuit traces also under the zinc casing, but they are under the layer of solder resist. (I don't get why those traces were covered but the ZIF vias were not; oversight by the PCB designer, I guess) But that's a moot point if your concern is true (which I do believe you).
Focus on Fundamentals consists of 45-minute on-line classes that cover a host of technologies. You learn without leaving the comfort of your desk. All classes are taught by subject-matter experts and all are archived. So if you can't attend live, attend at your convenience.