My first experience with Tesla was more than two years ago, when they were still a small company that made novel roadsters. On November 8, 2010, a showroom Tesla Roadster on its way to NYC ran low on charge in Cambridge and pulled into Lowell House courtyard to borrow the laundry room’s power.

The next day, I made an excuse to meet the engineer that had driven it up, and I traded some product photography for a short test drive.

The Roadster felt like a glorified go-kart. Sure, it accelerated incredibly fast, but given its comically small cabin, bumpy ride (the Lotus frame on which it is based is also comically small), and lack of power steering (not so good for Cambridge streets), you almost felt as if the ungodly responsive acceleration was owed to you.

The Roadster felt like something someone with too much money would own for short highway drive or track days, but not really a car. I thought it was a really cool test bed for EV technology, but still an immature product. Now, in 2012, the Tesla Model S has fixed all of that.

The Tesla Model S is really the first Tesla car. Three rows of seats, 52/48 weight split, effective 89 MPG, and giggle-inducing performance that put an M5 to shame: the Model S is such a mind-blowingly impressive car that Automobile Mag named it its 2013 Car of the Year.

Except it has one major flaw.

The entire dashboard is one giant touchscreen. “But hey, that’s so modern. So Tesla,” you might say. And you would be right – but this is where Tesla’s desire to distinguish itself from any old performance sedan ended up shooting itself in the foot.

Think about a car’s dashboard for a second. It’s populated with analog controls: dials, knobs, and levers, all of which control some car subsystem such as temperature, audio, or navigation. These analog dials, while old, have two features: tactility and physical analogy. Respectively, this means you can feel for a control, and you have an intuition for how the control’s mechanical action affects your car (eg: counterclockwise on AC increases temperature). These small functions provide a very, very important feature: they allow the driver to keep his or her eyes on the road.

Except for a the privileged few that have extraordinary kinesthetic sense of where our hands are, the Model S’s control scheme is an accident waiting to happen. Hell, most of us can barely type with two hands on an iPhone. Now a Model S driver has to manage all car subsystems on a touchscreen with one hand while driving.

I’m not saying the touchscreen is bad idea. It’s a fantastic idea if implemented correctly. This means taking advantage of the screen’s digital nature with gestures recognized anywhere on the screen, and feedback in some form such as audio confirmation (preferably in an aloof, British accent). The driver can stay focused on the road, and the touchscreen provides functionality beyond the analog controls its replacing.

Instead, Tesla has chosen to use the touchscreen as an over-engineered version of the analog controls. Digital buttons and sliders make up most of the interface. A touchscreen will never convincingly replicate a button, but it doesn’t need to. By doing so, it neither fulfills its full potential as a next generation control system, nor does it accurately replicate what it’s trying to replace. Imagine if the inventor of the cello insisted that it play violin parts. It’s possible if you try really hard, but everyone was better off that the cello music exploited a cello’s defining lower register instead of crappily playing violin parts.

Tesla’s focus on the Model S touchscreen console was a major and potentially dangerous step in the wrong direction. BMW’s Heads Up Display got it right: use technology in radical new ways that analog displays couldn’t – by placing important data in the driver’s field of view so eyes could stay on the road. At least the M5 has something to be happy about.