app development environments offered by OS providers have evolved, over the 15 or so years of their existence, to offer comprehensive support for a wide range of tedious, difficult, or costly aspects of developing, versioning, upgrading, and deploying apps across multiple device configurations. So much of the hard work of app development is “blackboxed”—especially for small developers, who can take many things off the shelf, as it were. There are many unrealized, largely invisible aspects of these hinterlands of app software and computing environments; and they are due to the growing complexity, scale, and yoking together of different technologies in the digital platforms. So a customer seeking to grow her or his business by developing her or his own app is reliant on an assemblage of digital media and communication. Thus the single app and its potential users and communities fit into a global picture, which is of course much larger. More on this later.
So far in this section I have been sketching an anatomy of an app. Apps are software that rests upon layers of other software, all ultimately written in code, and all collectively drive the machine of smartphones and other devices to undertake what Lucy Suchman famously called “situated actions” (Suchman, 2007). Among the many things that apps marshal, one that looms large is data. Data from smartphones have special links with personal and collective information.
If we recall the predecessor technology of the telephone, information about subscribers was most systematically known by the phone company, and it was gathered and made available in directories. The calling patterns were typically studied by engineers in telephone companies to inform their design and planning of network capacity and distribution. The content of conversations held during telephone calls and the calling parties themselves could either be overheard, on party wires or via the operator, or listened in through telephone interception or phone tapping (Goggin, 2006). Such interception was possible with mobile phones as well, although encryption made it more difficult. However, with mobile phones came the widespread sharing and collection of telephone numbers: the preciousness of this identifying personal information is underscored by its role in money transfer apps or in messaging apps such as WhatsApp or WeChat. As they evolved, mobile phones gathered and brought to maturation many other sources of data in the smartphone era.
Especially important are location data, which are obtained via technologies such as cellular network triangulation, GPS, and Bluetooth. Thanks to their portability and intimate relationship with their users, smartphones offer rich data for following these users’ daily journeys and for approximating their locations. Many apps have been developed to take advantage of location data, the general arc moving from dedicated apps (e.g. check-in apps such as Foursquare, or map apps such as Waze) to incorporation of location data features in a wide range of other apps, especially social media ones.
Then there are data about people’s bodies and bodily states. These are the kinds of data used by health and wellness apps. Such data are directly gathered from the sensors contained in smartphones, as we have just seen. Many of them are used inferentially—for instance in apps that monitor, gauge, and arbitrate sleeping patterns, the amounts and quality of exercise and physical activity, health, well-being, or any kind of behavior; and they often do so problematically (Barnett et al., 2018). During the COVID-19 pandemic, health researchers, medical practitioners, and developers sought to develop apps that would assist in the diagnosis of positive cases on the basis of data from sensors. Some apps encourage people to enter these data themselves, as in a diary or journal.
One of the major axes of smartphone data is the connection between the app that runs on the smartphone and what is accessed, transferred, and collected, be it via networks, via databases, or via people and other things. Apps have a communicative function, accessing data from elsewhere or sending their data to a server, a database, or a repository elsewhere—which is dramatized in the development and discussion of cloud computing. The rise of apps has been enabled by the rise of cloud computing. Many apps are designed for, and rely upon, the cloud (Sitaram & Manjunath, 2012). Key cloud-based apps include the Google suite of apps, Microsoft Office, and many health apps (Woodward, 2016). So the “appification” of mobile communication has been powered by the rise of the cloud (Stawski, 2015). There is a spectrum of such implementations, from the many apps running on mobile and other devices that draw data from and use the services and capabilities (e.g. machine learning, AI, virtual machines) of cloud-based platforms such as Amazon Web Services (AWS) (Mishra, 2018), to cloud apps and cloud app marketplaces (Nguyen et al., 2016).
Another category of data is transactional data, which are generated when we make a purchase or book a ticket. There are also data on the activities we perform with apps. Watching a video via Netflix on a smartphone or tablet generates data that are held remotely as well as locally, and are “synced up” (i.e. updated) with one’s account. More and more areas of everyday life require apps for participation: there is now, for example, check-in to places via quick response (QR) code, which is designed to enable infectious disease tracing in the COVID-19 pandemic through social media and search apps; and there can be requirements to book a swimming pool spot or do banking or money transfer via an app. Given such developments, many more data about people, their lives, and their environments are gathered by or pass through apps. This dataphilic quality of apps is not only defining, by now it is well nigh constitutional of apps.
Hence the constant struggle to staunch the flow of “leaky apps” (Ball, 2014; Cadwalladr & Graham-Harrison, 2014), and to put in place safeguards that can regulate the data gathering, data use, and data sharing done by people’s main devices or by apps operated by better known brands and by companies with third-party apps or providers. This was (and remains) the nub of the problem with the 2018 revelations that exposed Facebook’s sharing of user data with the Cambridge Analytica company. The Facebook scandal was but one of many instances of data breaches, poor practices, and lack of adequate legal and regulatory frameworks and redress that have made privacy and data governance a burning issue of our time. By turns, apps are at the frontline of concerns about both private companies’ and the government’s use of personal data for profiling, tracking, and surveillance.
Histories of Apps
So far I have provided a working definition of “app” and showed you what an app typically does and looks like. To gain a better appreciation of apps as a socio-technical system, it is useful to look at the processes by which they emerged.
The classic predecessors of smartphone apps can be found in the so-called handheld devices, which gained their markets in the 1980s and 1990s and persisted until the early 00s. Handheld devices can be seen as part of the long history of the emergence of personal computing, with its debates on the kinds of control that users and developers may have over these tools and the software possibilities of their systems (Ceruzzi, 2003; Kelty, 2008). Such devices include handheld electronic calculators (Hamrick, 1996; McGovern, 2019), which attempted to replace the slide rule. For instance, Hewlett-Packard’s HP-65 calculator of 1974 offered “full programmability,” featuring “interchangeable magnetic cards as storage media for factory and user programs” (McGovern, 2019, p. 300).
A direct descendant of the smartphone is the family of devices variously called “handheld computers,” “palmtop computers,” and “portable digital assistants” (PDAs), the last term being one coined by Apple’s CEO John Sculley (Sakakibara et al., 1995). The PDA was patented in 1975, and Toshiba is credited with bringing it to the market for the first time, in 1980 (Golder et al., 2009). The UK computer firm Amstrad introduced its PenPad in early 1993, just ahead of Apple’s Newton MessagePad launched later that year, which featured built-in apps with web, email, calendar, and address book functions (Sakakibara et al., 1995). The Newton is claimed by some to have made a breakthrough; it had attributes that anticipated the smartphone OS and apps environment (Foley, 2000). It gained a strong following, and its brand community persisted in using it with quasi-religious fervor even after the device was abandoned (Muñiz & Schau, 2005).
Three electronics companies known for their calculators also launched PDAs: Hewlett Packard, Casio, and Sharp—which by the early 1990s dominated the personal organizer market
