This would come from a tradeoff between going very bold with something that won't have a chance in a million to become the future, or staying very close to the present and showing an iteration on what could be the next year's model.
I can't stop laughing at the one that looks like someone said "Hmm, I like this low profile front end, but it needs something else...I know! I'll melt another front end on top of that!"
The little egg car seems almost futuristic today, the only giveaway being the narrow steering column.
The one with the steering joint in the middle of the car was like every last place British entry on Full Metal Challenge. Turns out that's a terribly impractical idea.
This concept basically invented the minivan/MPV (no sliding doors, but most European MPVs don't have them).
After Fiat/Lancia passed on putting it into production themselves, multiple manufacturers took note and began working on their own production versions. The Japanese were actually the first to market with the 1982 Nissan Prairie and 1983 Mitsubishi Chariot. The Americans and Europeans were next: Chrysler launched the 1984 Dodge Caravan and Plymouth Voyager, while Matra and Renault launched the 1984 Renault Espace (it was designed by Matra and originally intended to be sold by PSA as a Talbot... but PSA killed the Talbot brand and with it their partnership with Matra at the last minute, so Matra took the design to Renault).
The Scion xB is one of the best vehicles I have ever owned. I traded in my 2001 Jetta VR6 (mystifying mass airflow sensor issues VW could never fix for more than a couple weeks) for an xB in 2004. It had its flaws, but for a daily driver, I think you'd be hard pressed to find a better car than those first generation xBs. Mine is currently owned by my parents and has 300K+ and the only major service has been a new clutch at 175K.
This should be a slide deck for startup school, or any product management course.
So many of these seem like obviously terrible ideas, and yet, these are what happens when you design products without customers.
You can see where some of them were just derivations of other companies design languages, where others were the expression of one meaninglessly novel idea (the catamaran car). Some of them you can see the thing they thought would be good (big windshield, aerodynamics, ornamentation, minimalism, quirkiness, etc), but when you tried to wrap a car around that idea, it was uncanny and un-car-like.
For each of these abominations, there are software and startup analogs, I'm sure of it.
I also like to think they were built thinking outside the box, because that box didn't exist yet. I think we should stop from time to time and think about our designs, is this the only and perfect way to design it? or maybe I'm just copying what everyone is doing?
Now days most cars, web pages, apps, phones, TVs, look the same, yes that has its benefits and maybe customers just want the same old things, but I always wonder how different designs would be if apple made a different iPhone, Ford a different model T, xerox a different UI...
At least with cars, I don't think you can trace most of their features back to the Model T. The model T had a completely different control scheme, wagon-like suspension, band brakes (for parking only), and was only available in black.
What I often like is when production cars have a concept-car like look with features you usually do not have with other cars. (Citroen C4 Cactus just comes into my mind)
The US has different legal requirements for cars to the ECE standards used by most of the rest of the world (US protectionism? Surely not). Unless Citroen / PSA renter the US market and make a special US variant, they won’t be legal until the car is 25 years old and qualifies for an exception.
I was always bummed the VW one liter never materialized. At first, it was supposed to get insane gas mileage. Though I did see in a magazine that the new microbus is supposed to come in a few years. It looks pretty cool.
I wonder how much of the dashboard could really be relocated. Bosch had just began developing the CAN protocol that year and chips wouldn't be available until four years later. There's an illustration I found [0] that shows how the modules connect and based on the history of the automotive databus, I don't think it would have been able to give you a dashboard as populated as the image in your link. There would need to be a pair of power pins along with a pair of pins for every value you would want to read out (either for analog voltages or RS-422 since SPI/I2C/CAN had not been invented yet). I'm guessing there might have been hard restrictions on some modules like the stereo and speakers, why would you want the speakers to be anywhere else in the car? Going off of the typical dashboard in the 80s, you have speed, tach, fuel, battery level, oil temp, radiator temp, and the odometer and a trip meter. Maybe integrate the trip meter with the odometer value to reduce the 3-4 wires needed for displaying/resetting it. The clock would only need power. Then add in the climate control, that has to be at least a couple more signals for temperature and fan speed and maybe an on/off for the AC if it was available at the time. Don't forget all the status lights, that is going to add a lot more pins to these connectors. The actual pictures don't show any of how they connect. The illustration shows what looks like two connectors. Maybe that's how they did it, two 25 pin d-subs would probably be enough for running everything out to every module dock, impractical but possible. You could accomplish this today with a single 4 pin connector: power, ground, can+, can-.
My favorite from the first few pages is the '55 Gilda. It's hella swooshy but still looks like it'd be a practical car to drive as long as you glued some mirrors back on.
http://oldconceptcars.com/category/1930-2004/seat/