Very soon, every visitor to the Cooper Hewitt, the Smithsonian’s recently reopened design museum, will receive a giant pen. This pen is not really a pen. On the table, it looks like a gray plastic crayon the size of a turkey baster. In the hand, it feels pleasing, chunky, hefty like a toddler’s rubber ball. And at the museum, it does something magical.

Next to every object on-display at the Cooper Hewitt is a small pattern that looks like the origin point of the coordinate plane. When the pen touches it, the digital record of that object is added to the visitor’s personal museum collection. When they leave, they will have to return the pen, but information about and high-resolution photos of the object will be waiting for them.

To people who photograph placards when they visit museums—a group to which I belong—the pen is a godsend. It anticipates a need and executes it; it is a straightforward, useful object. But it’s something more. The pen does something that countless companies, organizations, archives, and libraries are trying to do: It bridges the digital and the physical.

Last month, the Cooper Hewitt welcomed visitors again after a three-year-long closed-door renovation. Its leaders were rethinking what a design museum should be. And a five-member team inside it—Cooper Hewitt Labs—was thinking about questions which the pen addresses, questions about how to bind the vast possibilities of the digital with the finite fact of the physical.

It is a good time for the museum to reopen, for design has rarely been so central to the American popular conversation. But its leaders have succeeded in something that should interest more than the chambray-wearing set. The Cooper Hewitt has transformed into an organization not unlike Wikipedia, Pinterest, or, for that matter, The Atlantic: Somewhere between a media and a tech firm, it is a Thing That Puts Stuff on the Internet.

Or, more precisely, A Thing That Puts Things on the Internet.

But to get to that point, the museum has made sweeping decisions about who it wants to serve, how it should serve them, and what that “service” should look like. The leaders of the Cooper Hewitt—a national steward of well-designed things—have ultimately had to shift their understanding of what a thing is in the first place.

* * *

Cooper Hewitt

Excepting the ones who are alive now, every human being who ever lived has died. Their belongings—the objects that filled their life and helped give it meaning—all met a similar fate. Some were destroyed. Some found new owners. And a tiny, tiny fraction of a fraction were saved. They would be preserved—they belonged in a museum.

If you wanted to see those objects, you had to go to that museum. In glass cases and on wooden shelves, you could survey the stuff that people made or used or prized. Saving this old stuff was so important that cities and states established institutions to do it, libraries and archives and museums that made sense of the receding past. Their job—to steward the products of the past into the unfolding present—was tough, but possible. As long as the government lasted, so would the objects.

“All jokes about politics aside, the United States isn’t going anywhere, and the Smithsonian is the national museum,” Aaron Straup Cope tells me. “By definition, it not only traffics in the past, it has to traffic in the near future. It has to keep an eye on it, it has to have some sense.”

Cope is the lead engineer at Cooper Hewitt Labs. For years now, I’d come across intriguing updates from him and his colleagues. Led by Seb Chan—a prominent Australian thinker on museums, who’s also a DJ and music journalist—Cooper Hewitt Labs seemed to be spinning a textile from that near future and releasing it strand by strand. A blog post here, a demo there: The ideas and software they put out represented such a complete vision for what a museum could be that I wasn’t able to grasp it all at once. There was a sophisticated idealism in their output—an idealism about what the Internet could be, especially for public institutions hooked up to it—that felt rare, precious, and vital.

The museum’s magical pen was made by GE and other firms. (Cooper Hewitt)

I, at last, got to visit the museum’s physical premises in November. It was in its frenzied final days of hibernation, and it smelled like sawdust. The Cooper Hewitt resides in Andrew Carnegie’s old Manhattan mansion, an idiosyncratic testament to industrial domination that’s now next door to the curvaceous Guggenheim. The building sports all the marks of its fin de siècle manufacture: carved ceilings, rococo columns, marbled steps. But it is also a recently renovated contemporary museum, so sometimes you climb those stone steps (carpeted with maroon velvet) and find yourself in a hall of pristine white drywall splashed with huge sans serifs. It’s steampunk, but it’s also Eamespunk.

In this labyrinth of hardwood and cement, the Labs office looks like any other crammed New York tech outfit. Sheets printed with 64-point black Helvetica were taped to the wall and whiteboard. Desks had the regulation iMac, hand-scrawled notes, and (sometimes) a vinyl figurine or two. There was a DSLR and a half-disassembled Raspberry Pi.

The Labs team is small, just five people. Cope has had many lives, including a stint at San Francisco star firm Stamen Design and a “self-imposed sabbatical” where he made parallel copies of private social networks, but his most famous gig is as a senior engineer at Flickr. It was his code that let users geotag their photos and hook them into third-party apps like Foursquare. Seb Chan came from the Powerhouse, Sydney’s premier design and science museum. Katie Shelly produces films and multimedia for the museum; she’s also made an all-graphic cookbook. Micah Walter was at the Cooper Hewitt first, the museum’s old webmaster, but before that he was a photojournalist in the Middle East. And Sam Brenner’s projects have ranged from an RPG about LiveJournal to an Internet-connected fantasy football trophy.

When I visited, I talked to the Labs team in their office and then toured the then not-quite-finished mansion. We talked about the museum first—the physical one we were in. Unlike leaders of other New York museums, who are investing in events, Chan (and the Cooper Hewitt generally) believe the heart of the museum is in its collection and its visitors. In other words: its stuff and its people.

“They don’t want to have the burden of this preservation forever,” he said of the increasingly event-focused Museum of Modern Art, 40 blocks south. “The beauty here is: We’re the Smithsonian. We don’t have a choice. No matter what other staff in this building might say, we don’t have a choice but to keep all this stuff forever.”

The museum will forever be committed to its stuff. But it has to have a more enlivening presence, he believes, than placards and shelves. Cope held up his smartphone at one point and pointed at it.

“Everyone walks in with one of these,” he said. “We’re gonna have to find a different way to air-quote compete, so why don’t we just try to meet people halfway? All the visitors arrive with superpowers.”

* * *

Tall Green Bloom, a 3-D printed urn which joined the museum’s collection in 2013 (Cooper Hewitt)

When visitors come to the Cooper Hewitt, what do they want to see? This is the question the entire museum staff has asked itself. It has decided to offer them something that can’t be found anywhere else.

So it gives them, first, yes, physical objects from the collection—the artisanal and industrial specimens that distinguish the Cooper Hewitt alone. A 3-D printed urn, Milton Glaser’s Dylan poster, and small model staircases built by Victorian woodworkers greet the visitors on the mansion’s first and second floor. (“People love those freaking tiny staircases,” Shelly said.)

The Cooper Hewitt also wants to teach visitors about design-as-craft, so it presents science-museum-like interactive exhibits on the design process. It also has an expansive upstairs exhibit on tools and tools-making, that includes a “sand selfie” portrait-making robot.

But the real treats are in the museum’s interactives that draw from its collection. There’s an “immersion room,” which projects patterns from the museum’s expansive wallpaper archive on the wall. Visitors can also draw their own patterns in there too, which tessellate on the projected walls like the original historical decorations. There are also large, “social” touch-screen tables—think of giant iPads—that let people alone or in groups sort through and look at objects in the collection. These have special search and manipulation features: Someone can draw a shape on the table and see what items in the collection fit it. And the pen—the jewel of the museum’s collection-based interactives—will function as a pen on these touch surfaces. The pen is the exact kind of object that the museum hopes to deploy in the mansion, as it augments a smartphone without requiring one.

All three of these tools—the pen, the touch-screen tables, and the immersion room—were designed and manufactured by outside firms like DSR and Local Projects. But they were created in collaboration with the Labs team, and—more importantly—they used an infrastructure developed by the team. It is the infrastructure that lets the museum plan for the near future, that lets it bridge digital and physical, that lets it Put Things on the Internet: the API.

API stands for “application programming interface.” When combined with a network protocol like HTTP, an API lets computers talk to each other. Often, that talking looks like fetching information from an online database in a repeatable, organized way. Many big websites have APIs, because they let the website play nicely with the rest of the web: Twitter, for instance, has one that allows you to ask, for example, for a user’s 10 most-recent tweets or for all the tweets tagged in a certain geographic region. Flickr’s old API, which permitted far more, was considered so special that fans asked it be considered a National Historic Landmark.

The most powerful, most important thing that Cooper Hewitt Labs has done for the museum is build one of these. No wonder it's being called “the API at the center of the museum.” The Cooper Hewitt’s API connects to the museum’s two operational databases—its vast collections database and its complex customer and ticketing databases—and fuses them. Then it makes the collections part public and accessible.

What the API means, for someone who will never visit the museum, is that every object, every designer, every nation, every era, even every color has a stable URL on the Internet. No other museum does this with the same seriousness as the Cooper Hewitt. If you want to talk about Van Gogh’s Starry Night online, you have to link to the Wikipedia page. Wikipedia is the best permanent identifier of Starry Night-ness on the web. But if you want to talk about an Eames Chair, you can link to the Cooper Hewitt’s page for it.

Cope explained the importance of these permanent links.

“If you and I have something we can share, then that starts to give that object weight and mass in the universe. It starts to exist,” he said. In other words, these shareable and permanent URLs start to stand in for the locked-away object.

“It’s a proxy for sure,” Cope said, “but because we can’t let you run through the actual warehouse, the choice is nothing or something. And we choose something.”

“What 'digital' in the museum means is really that everything is available whenever you want. Wherever you want, whenever, however,” said Chan. Then he asked himself the follow-up question: “How does that play out in the museum, physically? How does that begin to change the exhibitions and the ability of the exhibitions to do just different sorts of things and different ways of presenting the collection?”

For his team, it’s meant teaming up with design firms to build the pen and the immersion room. The API’s fusion of collections and visitors is what lets the pen function. The API lets users look at their favorite objects after they leave the museum. It lets users link to those objects on Facebook and talk about them.

The museum’s stack, with its two databases at the bottom and the pen at the top (Cooper Hewitt)

The museum has built itself what programmers call a stack. At the bottom are two big, proprietary servers: the database that knows about the collection and the database that knows about the visitors. In the middle is the API. On top of the API is the website, where visitors can learn about the collection and buy tickets. (It’s the website, therefore, that transfigures users into visitors.) And at the very top are the gallery interactives and the pen.

Notice the trick the Labs team has completed. The API seems to be first for users and developers. It lets them play around with the collection, see what’s there. As Cope told me, “the API is there to develop multiple interfaces. That’s the whole point of an API—you let go of control around how people interpret data and give them what they ask for, and then have the confidence they’ll find a way to organize it that makes sense for them.” But who is doing the most work around the collection—the most organizing, the most-sensemaking? It’s the museum itself.

“When we re-open, the building will be the single largest consumer of the API,” said Chan.

In other words, the museum made a piece of infrastructure for the public. But the museum will benefit in the long term, because the infrastructure will permit them to plan for the near future.

And the museum will also be, of course, the single largest beneficiary of outsider improvements to the API. It already talks to other APIs on the web. Ray Eames’s page, for instance, encourages users to tag their Instagrams and Flickr photos with a certain code. When they do, Cooper Hewitt’s API will automatically sniff it out and link that image back to its own person file for Eames. Thus, the Cooper Hewitt’s online presence grows even richer.

The Cooper Hewitt isn’t the only museum in the world with an API. The Powerhouse has one, and many art museums have uploaded high-quality images of their collections. But the power of the Cooper Hewitt’s digital interface is unprecedented. There’s a command that asks for colors as defined by the Crayola crayon palette. Another asks if the snack bar is open. A third mimics the speech of one of the Labs members. It’s a fun piece of software, and it makes a point about the scope of the museum’s vision. If design is in everything, the API says, then the museum’s collection includes every facet of the museum itself.  

* * *

The museum’s social tablet tables (Matt Flynn/Cooper Hewitt)

Why can the Cooper Hewitt build something like this? It comes down to money.

Most museums—or arts groups, or public institutions—work on the grant model: Organizations apply for a large sum of money to be spent over a finite period of time. For example, a museum team might get 18 months to spend $500,000 on a big, one-off exhibit about Monet. In-house curators hire external developers and a design firm, both often discounting their time as pro bono work. The outside developers make a fancy Flash app about why Monet’s brush stroke is different, and the designers produce elegant signage to teach visitors about just what they’re seeing in the museum’s halls.

Then the money is spent, and the famous design firms leave.

But this guess-and-fund model doesn’t always work. If contractor code breaks, there’s no money left over to fix it. Consider that the grant model is not only how many big exhibits get funded, but also similar to how Healthcare.gov was funded: “Expert” consultants got the specs, deployed a broken site, and departed.

“We banked on being able to outspend consumer technology by just this much,” Cope says. “So what we do is we raise a huge amount of money, and we go and we buy some very, very fancy ooo-shiny, and we put it in the galleries and amortize the cost over three or four years, and that gets people in the door, and then we do it again.”

“The museum I was at before this was a science museum, and in 1988, it launched with the most interactive experiences of any museum in the world… in 1988,” said Chan. “Well, before I got there, most of those had stopped working.”

The museum’s marvelous tiny staircases (Robinson Meyer/The Atlantic)

Even if things do work, the model turns museum websites into museums themselves, catalogs of once-snazzy apps built for special occasions before being discarded forever. Exhibits go away, but those apps never do. A museum’s website—the primary face of the museum to the world—winds up looking like a closet of old prom dresses.

When Bill Moggridge became the Cooper Hewitt’s director in 2010, he wanted the museum to make its digital infrastructure more thoughtfully. Moggridge, it should be noted, is a legend. He helped design the first laptop computer. He founded the world-famous firm IDEO. And he invented the term “interaction design.” Moggridge died in 2012, not living to see the renovation project he began.

Moggridge created Chan’s position and hired him for it. And while Chan could have kept outsourcing projects to big outside firms, he instead lobbied for funding and hire a staff. The museum’s digital work was too important. It had to have in-house experts. “There's a lovely phrase we use a lot,” Cope said. “The guy who invented the Perl programming language talked about Perl as being there to make easy things simple and hard things possible.”

“That’s how we try to think about this. Not everyone’s gonna understand what we’ve built or the potential of what we’ve built right away. It’s gonna take some of the curators longer than others to figure it out. But the minute they get it, they should be able to turn around and be like, 'What if…? Can we do…?'—and if it’s easy, it should be live in 15 minutes.”

“It’s capacity-building for people to imagine new things, beyond a book, beyond what they thought possible,” said Chan. “For people to go, Wow, I’ve always wanted to do an exhibition about this and now it’s possible to do that.”

* * *

Robinson Meyer/The Atlantic

The team has accomplished so much largely by accepting imperfection. When the Labs launched the API, it was missing a lot of information. Cope called the quality of its metadata at launch “incredibly spotty,” before Chan clarified, “it’s terrible.”

But that was on purpose. Better to put the museum’s grand imperfection and incompleteness out in the world and let people make of it what they will, the team decided, then wait for it to be perfect. “It was a tactical play to say, don’t obsess about that stuff, because its what people do with it that matters,” said Chan.

“We could spend the next 50 years trying to make that data perfect and it still would not ever be perfect. There was 70 years of collecting that had different documenting standards. Museums only started collecting policies in the eighties and nineties. How can you retrospectively fix everything? It just can’t be done. So let’s move on and figure out what we want to do with it,” he said.

This attitude—popularized by Steve Jobs with the phrase, “Real artists ship”—extends to how the team thinks through media production, too. “I can’t sit on a video for six months, making these minute edits. I have to pitch it out door, so we can say: This interview got this many views, this thing got this many views, let’s keep going with this,” said Shelly.

The Labs’s work, as a whole, is an investment in a particular idea of cultural democracy. It’s a view where imperfect speech can always—and will always, and should always—be augmented by further speech. It trusts in the discourse over the perfection of the original work.

And that piecemeal vision of how culture should work is borne out by how the museum actually came together. The Cooper Hewitt is an institution agglomerated. The Hewitt sisters assembled their collection of designs and patterns as an educational resource. Like their grandfather, Peter Cooper—father of the Cooper Union, which was until recently one of the last free undergraduate art schools in the country—they hoped to tame the rugged American craftsman by exposing him to the Continent’s fine decorative arts. And the collection only resides in the Carnegie Mansion now because both it and the Hewitt collection happened to enter the Smithsonian’s stewardship at the same time—and both were already in Manhattan.

A screenshot from Planetary, the first app the Labs team collected (Cooper Hewitt)

Maybe that’s why, when the Labs team collected its first app, it open-sourced it. Planetary is an iPad app that let users explore their music collection like it was a galaxy of celestial orbs, with stars as artists, planets as albums, and moons as songs. Thanks to iOS updates, it hasn’t worked perfectly for a couple years.

“We didn’t want to make the first thing we collected different from everything else except it was a different medium. It was: No, this is fundamentally different, because you have access to it,” Chan told me. “The ability to see how a thing’s changed through its development process is something that is potentially unique in that regard.”

And that open-source code has begun to restore the old purpose of the Cooper Hewitt: a kind of teaching database of physical objects. Writing about the museum in the early 20th century, Eleanor Hewitt made it clear that her main goal was not to save the objects in the collection, but the information within them:  

Naturally constant use will have a tendency to damage, even destroy certain objects (many of course, are indestructible), but irreparable damage could not be accomplished under a hundred years, and if in that time an artistic tradition passed on from father to son, as in Europe has been created, the existence or non-existence of these objects will not seriously matter, and during all that time the Museum will have been fulfilling its destiny.

From the beginning, then, the Cooper Hewitt has prized information over object, discourse over perfection. And while it can no longer permit artists to destroy the physical objects it holds by copying them—the Smithsonian’s central job is preservation—it can allow them to mess with the digital versions. Hence the API, hence the stable URLs, hence the open source code.

And perhaps already, the Labs team believes, that digital information will be inextricable from the physical object. The Cooper Hewitt has long collected napkin sketches of famous logos and inventions. If it wants to collect the rough thoughts of today, it will have to work fast, because napkins last longer in files than sketch files do on iPads.

“To collect a Nest absent of any data, what does that tell you?,” asked Cope.“It tells you it’s a beautiful piece of industrial design. Well, maybe the museum should start thinking about some way of keeping that data alongside the object, and maybe it doesn’t need to be privileged in the way the object is.”

In the past few years, technologists have mourned the “web we lost.” Most websites used to have open APIs: Now, the gardens of user content on which the modern web was founded have been walled away. What the Cooper Hewitt suggests is that public institutions might take up this mantle, making their considerable extant collections public and beginning to preserve new ones. The Library of Congress already holds the Twitter archive. In two decades, that collection—a remarkable reserve of American speech circa the 2010s—could look like a souped-up version of the contemporary Cooper Hewitt.

But the team’s aspirations are, for now, more commonplace. As I was leaving, Cope recounted how, early on, a curator had asked him why the collections website and API existed. Why are you doing this?

His retrospective answer wasn’t about scholarship or data-mining or huge interactive exhibits. It was about the web.

“I want people to link to this,” he said. “We’re the national design museum. Ninety percent of the essays that are written online about the Eames chair or whatever link to Wikipedia, and good for them, they earned it. But people should link to us, because we’re the Smithsonian. We should be that good. When it comes to social stuff, if we can provide that stability, if all it is is just people linking to us, to me, that’s enough.”