TURTLE,GREEN,RAW: Depending on External Data Feeds

I'm working on a project that involves nutritional data, and was happy to stumble upon the USDA's database of Every Food Known to Man. The database has more information than one could possibly wish for, including the amount of Retinol and Ash in common foods. And then, at the bottom of the list, is a record for TURTLE,GREEN,RAW. Which, for those who are curious, has 19.8 grams of protein per 100 grams of... turtle, green, raw.

There are probably a few good reasons why that, along with other gems (like PECTIN,LIQUID; SQUIRREL,GROUND,MEAT (ALASKA NATIVE); and BEAR,POLAR,MEAT,RAW (ALASKA NATIVE)) ended up in the data source. I'm more interested in what happens to the various products we build when we all rely on the same data feed, APIs, and links between services.

Here are just a few places you can go to find out the nutritional data in TURTLE,GREEN,RAW—25,900 results, all referencing the same data set. Programmable Web doesn't list it as a source, but does list some other places that are using the same data set and charging for access to it. There's both a nice sense of collaboration, as we reference open-source data provided by our government, and also a strange feeling of the blind leading the blind. We're all in it together, building structure after structure upon the same foundation. Who checked the foundation to make sure it's solid?

We can probably pretend, for a minute, that the data coming from the USDA is accurate (although I'm aware that, even with the rigorous scientific standards in place, various lobbying factions have likely influenced the methods used to identify the ash content in an ounce of turtle). A larger concern I have is regarding the lack of skepticism we have, both as designers and consumers, about how individual products and services are using data to shape and shift behavior. I bought Jess a fitbit for her birthday. A saga for a different time is how the device has broken three different times in three different ways; my point now is that I find it both troubling and also extremely predictable that this made it into a shipping product:

I'm most intrigued by how things like this happen in the first place. Did a nutritionist go to Indonesia, have a delicious meal of turtle with a side of tater tots, and then decide it needed to be nutritionally analyzed? Did the Polar Bear Company of Greenland lobby to get their meat added to our database? I asked the USDA, and got a response from Dr. Exler, a nutritionist. He explained that "The entry for Turtle, green, raw was in the 1963 edition of Agriculture Handbook 8, Composition of Food: raw, processed, prepared. We retained it for our current database, but have not gotten any new data to update the item."

A good fiction typically tells a story of the world in a way that we can relate to, but that has twists and turns that are surprising, challenging, or emotional. Part of a good design fiction is that it's a believable story about our future uses of technology. And so I'll tell you a quick story of a future where, each time you eat a meal, your phone keeps track of it automatically. It knows what you ate, where you ate it, how much you had, and it compares it to nutritional data to build a model of your health. And then, when you are at the gym, it organizes a workout tuned just for you, to track to your goals, based on your diet. It's called iHealth, and it's simply amazing: everyone has one. And over time, slowly but surely, everyone keeps getting fatter. Because the data's wrong. Some people question it, because they pay attention to their body. Most don't, because they just trust that it works, and because most people don't really question anything. And because it has such wide adoption—it is, after all, an Apple product—slowly, we all become obese.

That's extreme and an obvious ending—like the hero getting the girl—so I suppose I could do better. We're at a place where photographs aren't real, manufacturing is digital, and everything is digitally connected to everything else. And so, we all fail by the weakest link. What happens when my friend shares a popular 3D model of a chair with me, and so I print it on my fancy new printer, and—because it's TURTLE,GREEN,RAW —it's not structurally sound? What about when I see a stock photo in a news piece about a famous politician and it's TURTLE,GREEN,RAW—an entirely digital creation of something that didn't happen, that's been perpetuated without question? It's our modern day equivalent of war of the worlds: our inability to know what's real, based on layers of technology that we don't understand or question.

When the world is physical, and something is strange—say, a story about a guy who woke up in a bathtub missing a kidney—my ability to spread the strangeness is fairly linear, and so both exists slowly and changes slowly. I might tell 10 people about the Niemen Marcus recipe in my whole life, and I certainly wouldn't build a business on top of it.

But when the world is digital, I can spread strangeness easily. What's more, I can do it without knowing, and I can build an entire business or product on top of peculiarities that are invisible. We aren't that far away from open-source software for your car, and at one of my previous jobs, I watched a car company drive a car around a parking lot with the driver in the back seat, using an iPad. What strangeness will be perpetuated in the OS of these vehicles? We're still living with design decisions made by Bill Gates and Ivan Sutherland, and these were purposeful decisions. What about the subtleties of the digital medium that have crept in by mistake, or on a whim, or even as a joke, and then embedded in the systems we use and rely on?

Originally posted on Tue, 22 May 2012