A version of this post originally appeared on Tedium, a twice-weekly newsletter that hunts for the end of the long tail.
How much credit does Apple deserve for introducing its App Store concept to mainstream consumers? I mean, how obvious is the idea, anyway?
This is a question that seems worth asking as modern app development has become so centered around building applications that exist in digital storefronts like Apple’s App Store, the Google Play Store, Steam, and numerous others. What about app stores that existed when the proverbial dodo was still around? There are more than you think.
“A trip back in time reveals at least one popular one-click store which predates Apple’s attempt by at least 5 years. I know because I built it.”
— Michael Robertson, the software developer best known for his creation of MP3.com, in a blog post discussing his work on “Click-N-Run,” an early attempt at creating a digital download store along the lines of the App Store in the early 2000s. Click-N-Run (CNR), which was an aspect of the commercial Windows-like Linux distribution Linspire that Robertson helped build, was a commercial GUI-style interface for Debian’s apt package manager. It was eventually made available to other distros to much interest, though the results were reportedly a mixed bag. While no longer made, Linspire’s work on CNR (one of a few stabs at the GUI-based software distribution interface in Linux) likely inspired the graphical package managers now commonly offered with many Linux distributions, which largely work the same way.
Let’s play a game: If you were to access a piece of shareware circa 1991 and wanted to unlock the full version of that application with your computer alone, how would you do it?
No web. Perhaps no Windows. (Perhaps GeoWorks.)
Sure, there were probably lots of ways to download an application, if you had the tools to do so—as in, a modem. Perhaps you might grab it on a BBS, or through a service like Compuserve. Maybe Usenet binary groups were your preferred strategy.
But still, you’d be stuck with a nag screen. Your copy of WinZip would just be annoying you every time you started it.
See, the issue with the distribution of software via computer was never about the download part—that part was figured out relatively quickly. It was the paying part that proved difficult.
Think about how it might work compared to a store: You choose a physical object of interest, you give a physical object of value (i.e. money) in exchange, an intermediary (i.e. a cashier) confirms you bought it (by scanning a barcode), and you walk out, without an object of value but with an object of interest.
Which is why every time you read a story about some shareware pioneer, like the developer of Paint Shop Pro or Tim Sweeney’s efforts to sell folks ZZT, it’s always paired with a story about these developers literally taking checks in the mail, despite the fact that it was entirely possible to purchase things fully electronically with a credit card by this point.
Without the retail element, people were kind of stuck distributing software without a way to easily purchase it. (What’s the big deal, the open-source folks say.) There was no way to secure the process, so therefore, fraud was prevalent. And if people are distributing through multiple systems, the experience of downloading becomes annoying, because it feels wildly inconsistent.
This is the problem a few entrepreneurs worked to solve starting in the mid-to-late 1990s, with varying levels of success.
The year that Tucows, a well-known repository of downloadable software, first went online. The repository was built by Scott Swedorski, a technology enthusiast in Flint, Michigan, who spotted a need for a central resource for basic internet software. Swedorski’s work, while not initially intended to be commercial, proved the basis of a long-running company that came to prove an essential part of the early internet. After all, we needed software to get on the internet, right?
An example of the Electronic AppWrapper, a program for NeXT computers that could be used to distribute software digitally. Image: JohnWayneTheThird/Wikimedia Commons
The app store that generally gets the nod for being first dates to the early 1990s, on a platform that was responsible for a lot of software firsts. I’m, of course, talking about NeXT, the platform on which the World Wide Web came to life and whose object-oriented approach to programming became a big part of what has made iOS so successful.
That software is called the Electronic AppWrapper. (See, it had “app” in the name way back in 1993!)
It did not exactly have the same kind of distribution method that one might expect for a modern app store, however. Really, what it did that is incredibly clever was that it took the shareware CD-ROM and made it into something that allowed for trials.
NeXT was a good platform for this in part because of its initial target audience—since the NeXT Cube tended to focus on educational and research settings, those settings were often networked well before PC and Mac equivalents, so people were able to purchase software online before other markets.
In fact, Electronic AppWrapper developer Paget Systems, which initially built a print version of the AppWrapper, made this very point in a 1992 article first proposing the idea:
The NeXT community is a perfect testbed for electronic distribution. The market is still small; we know where almost all of the computer owners are, and the community is more fluent with networking than most. And we have more than our share of creative people willing to tackle problems in new ways.
Also helping matters: Since NeXT systems were rare, users of this platform didn’t really have the advantage of being able to go to Radio Shack to purchase software, so it was either do everything through the mail, or go electronic.
Paget Systems’ great gift to the app store concept was the process it enabled. A 1993 issue of NeXTWORLD described the benefits of the tool like this: “To order by e-mail, just click a button; the application automatically displays an order form, asks for your credit-card number, and sends an encrypted message to Paget.”
That sounds pretty simple, right? It was, and it's not all that dissimilar to how we do things today.
(Side note: Jesse Tayler, who helped develop Electronic AppWrapper, has put up an informative documentary website highlighting the history of NeXT and the innovations the company helped to enable—including this. A highlight is Tayler’s discussion of successfully demoing the Electronic AppWrapper to Steve Jobs.)
“Since virtual shelf space is much cheaper than a storefront, Online can represent thousands of products. We can carry Microsoft Word and hundreds of related add-on products, while traditional re-sellers can barely find shelf space for mainstream software.”
— Tim Choate, the president of the firm Online Interactive, discussing the company’s atOnce online software store, which was one of the first examples of a traditional app store for Windows computers. As NetworkWorld explained in 1996, the atOnce software store was something of a test, complete with Microsoft’s blessing, to see if application distribution of commercial software over network mediums was even possible. The process required a more secure approach, at Microsoft’s behest. It should be noted that atOnce provides an interesting case—as it was effectively the app store for the AOL era (along with a very early web presence), though it quickly went away.
The Digital River website, as it appeared in 1998. Image: Internet Archive
Joel Ronning probably doesn’t get the credit he deserves in the grand scheme of things, but he is a figure in this discussion that matters.
In the 1980s, he spent time focused on the Macintosh market, selling software and distributing white-label accessories as well. This gave him an understanding of the digital market, so he could see all the flaws of the retail approach up close.
Around 1993 or so, he had a revelation that proved prescient: downloadable commercial software was going to be big. Really big. Big enough that he should spend the next few years developing processes for making the ideas around secure downloadable software workable, and patenting them. And building a company around them. And turning them into something that other companies would likely want to use.
Ronning’s work led to the creation a dozen patents—and a company called Digital River that could handle the encryption and distribution of applications. Not that anyone knew how to properly contextualize the idea at that early stage. In a 1997 profile with the Minneapolis Star-Tribune, columnist Dick Youngblood tried, and came up with this:
What Digital River has created is an enormous virtual warehouse containing tens of thousands of software products offered by hundreds of developers and retailers through their individual Web sites.
In simple terms, the system gives customers the ability to download their software choices with a minimum of fuss and a maximum of security for the credit card numbers and other personal data required for the transaction. I like to think of the operation as sort of the Supervalu of electronic software wholesaling.
Imagine having to describe something that people do over the internet on a daily basis without being able to use the terms “app store” or “cloud,” nor the frame of references that come with those terms, and that’s probably what you might come up with.
A patent drawing by Digital River that showed an early example of its digital-distribution technology, which included encryption functionality. Ronning has a lot of patent filings to his name. Image: Google Patents
The approach Digital River took in this early stage is actually quite similar to what we think of as an app store today—including the idea that the middleman is going to take a cut. (Digital River took just 20 percent, rather than Apple’s infamous 30 percent cut.)
But one difference is that the company represented a provider of purchasing services—i.e., it built the tools for individual companies to create their own storefronts, rather than becoming an app-store player itself. This model worked for them. By the year 2002, Digital River (a still-active company!) had more than 32,000 customers according to NetworkWorld, with roughly a third of those representing three quarters of Digital River’s sales.
“Year over year we continue to see more products purchased digitally.” Ronning said in a NetworkWorldinterview. “People are getting more comfortable with getting a digital file than they were one, two or six years ago. That’s good news because it allows us to deliver a product halfway around the world in a matter of seconds.”
And hey, because the company was in a position to provide the technical know-how of running an app store, there was at least one case where Digital River was tapped to manage someone else’s app store—the creation of Research in Motion’s Blackberry App World in 2009.
Download Warehouse, a.k.a. what app stores would look like if we made no changes to our approach between physical and digital distribution. Fun fact: This is what the atOnce website evolved into. Image: Internet Archive
Not every site was a winner, of course: When looking for info about app stores, I ran across this website on the Internet Archive that literally sold digital software as if it was still in a shrink-wrapped box, which made me crack up so hard. The backend? Digital River.
In many ways, the success of the app store was just as much about the packaging—i.e., the way consumers were pitched about the idea, rather than the shrink wrap—as the commerce itself.
The year StarCode Software, a developer of software for the BeOS operating system, was formed. The company built PackageBuilder and SoftwareValet, which combined together to become one of the first graphical package managers purpose-built for an operating system—and one Be acquired in 1998 and integrated into the operating system.
Stardock Central, a digital distribution service produced by the customization-focused company Stardock. Image: via the Stardock website
- Steam. The digital distribution service started by Valve in 2003 was effectively the app store that proved the model to the world in a big way. There’s a reason why Steam remains so dominant in the PC gaming space, and it’s because they nailed it the first time—to the point where many of its competitors directly mimic the service.
- Windows Marketplace. This circa-2004 online store, targeted at consumers, was an attempt by Microsoft to centralize the often-confusing app distribution options for Windows software. It wasn’t successful, but it helped set the stage for later digital storefront successes for Microsoft.
- Club Nokia. This online store for Nokia’s early mobile phones provides a really interesting example of a service that was essentially a direct analogue to the modern iOS App Store, but in a situation where the carriers, rather than the phone-maker, holds all the power. This service, founded in 1997, became controversial as ringtones became more popular, and Nokia eventually folded to pressure from mobile carriers and scaled back its service in favor of the mobile providers’ options. Could you imagine Verizon and AT&T doing this to Apple today?
- Xbox Live Arcade. Launched in 2004, this represented an important formative effort in the attempts to bring digital download services to a large group of people. One secret to the success of Xbox Live Arcade was its piggybacking upon what Microsoft was doing elsewhere; it leveraged the existing Xbox Live service to sell people more simplistic games. (Apple later replicated this by using its mechanisms for the iTunes Store to sell apps.) It later proved the starting point for the company’s Xbox Live Marketplace, which could distribute full shrink-wrapped games to consumers.
- Stardock Central. I’ve mentioned them in Tedium before, but Stardock is an interesting company historically because of the fact that it was early to a number of important trends that have become even more essential today. One of those trends was customization; another was digital distribution, which it first dipped its toes into with Stardock Central, an app-distribution service from circa 2001. It worked particularly well for Stardock in part because it offered apps in a variety of verticals, including games.
The NTT DoCoMo NEC N904i, a phone that supports the formative i-mode network, is considered the first mobile app store ecosystem. Image: Wikimedia Commons
Thinking less about the maker of the phone itself or its general functionality, the mobile phone represents something about technology that a regular desktop computer doesn’t—it is intended as a full package, one that is hermetically sealed and managed by its maker and distributor.
Because it started from a different place, the expectation is different. That expectation creates tension for power users who wish that their phones or tablets worked a little more like their laptops, but on the other hand, the devices aim broader for a reason.
And one can point to the reason for the expectation that applications would be managed by the provider. After all, the first mobile app store that feels like what we got with the modern App Store or Play Store came from a mobile company—Japan’s NTT DoCoMo released i-mode, a service that offered access to digital services through their phones. It worked particularly well in Japan because home internet was rare at the time when i-mode was first released in 1999, meaning i-mode was many Japanese users’ first experience with an internet-style service.
As The Japan Times noted in 2011, the reason i-mode succeeded (and spawned many imitators) was the tight integration of payments and software:
For those of you who may not know, i-mode is the mobile Internet-access service built into cell phones from Japanese communication giant NTT Docomo. It costs ¥315 per month to use and includes the i-mode network, which is Docomo’s closed system, separate from the Internet at large. Within this network there are “official” i-mode sites, which are only accessible from an i-mode enabled cell phone. On sites such as these, users can purchase goods and services and have the payment appear on their cell phone bill. This cell phone-integrated-payment is what makes the i-mode system so special.
Other companies tried to do this same thing during the early 2000s, including firms like Nokia, to mixed levels of success, but the connective tissue was that the phone was treated like an integrated experience of purchasing, distribution, and usage, rather than a vessel for applications.
Perhaps this integration explains, in the present day, why Apple has ramped up attacks against sideloading (or allowing the installation of external applications outside of an App Store experience), something that nearly other mobile platform (including Android) has long allowed. As it wrote in a white paper it recently released:
Allowing sideloading would degrade the security of the iOS platform and expose users to serious security risks not only on third-party app stores, but also on the App Store. Because of the large size of the iPhone user base and the sensitive data stored on their phones—photos, location data, health and financial information—allowing sideloading would spur a flood of new investment into attacks on the platform.
Mobile phones have been built with this expectation that the whole experience is seamless and managed by the hardware developer—and at one point, the mobile provider even played a significant role. In some cases, it still does.
But one wonders how strong Apple’s case against sideloading will actually be, given that, y’know, it also sells desktop computers that allow sideloading … or as we call it over that way, downloading and installing apps from the Web.
It’s long been said that Apple, when it released the iPhone, launched a device so compelling that it made people forget that there was years of prior art that predated the moment.
In many ways, the App Store made people forget about app stores. It was such a brilliant concept, idea, and execution that when Steve Jobs announced it in 2008, people basically ignored the nearly two decades of prior art that wasn’t even particularly well-hidden.
In some ways, the move to centralization was arguably disappointing, because it wasn’t perfect, and it put a middleman in control. Apple’s approach to the digital storefront had flaws—most notably the size of its cut (which companies like Microsoft are now explicitly counterprogramming against) and the weirdness of putting a single company’s moral compass in charge of the apps that people downloaded.
But we can look at the positives of their approach as well, and sort of the element that they nailed that few others were able to in quite the same way. The integration of the App Store into the operating system made both better; the integration of commerce into the App Store using a common system solved the problem of having to give a credit card number out every time you wanted to download an app; and the integration of a development strategy that worked in tandem with the App Store gave (and still gives) Apple a reason to constantly improve its programming interfaces so they remain at the top of their class.
No developer of a prior app storefront had been able to nail down quite this mix (with Steam possibly getting the closest), which explains why it was so effective when Apple did it.
But prior art is prior art, and one hopes that the technology industry takes a step back to learn the lessons from both the Apple App Store’s strengths and weaknesses going forward. After all, so many others got there first.