An Introduction to the Archaeology of the Desktop Environment
Smooth movement over Non-Place
Rightly credited for the rise of personal computing and a paradigm for interaction that has become the ancestor to a wide range of different computer interfaces, the Desktop Environment (DE) has for decades provided non-technicians with a way to interact with computers, thereby creating a huge market that has been capitalised very successfully. The DE established a Graphical User Interface in order to alleviate the previous requirement of knowing lots of text-based coding languages in order to perform computations, introducing iconography to the interface and opening up opportunities to thereby use computers in new ways. Once this move was made from a typographical space to a much more visually flexible one, interface designers were better able to bridge the chasm between public practices and niche tools that were mainly of use to mathematics and the sciences.
The principle design language of the DE was of the desktop metaphor, the once almost ubiquitous practice of designing a user interface that is visually reminiscent of a desk and the paraphernalia one might expect to find there. This essentially skeuomorphic approach offered a new paradigm that would make an easy transition for businesses to adopt the technology and its human-friendliness catalysed a veritable explosion of adoption throughout almost every sector and institution.
This essay will constitute a new introduction to an archaeology of the DE, one that is new by virtue of our moving away from the desktop metaphor as the essential characteristic of paradigmatic use. While it is certainly the case that the desktop metaphor continues to be prevalent, the movement of the computing industry away from this paradigm suggests that it isn’t the whole story in this archaeology. Tablet computers, smartphones and the increasingly wide range of other computing devices have almost entirely dropped the desktop metaphor, and many attempts are being made to reduce this skeuomorphism in contemporary desktop operating systems as well. If the practical uses of the desktop computer aren’t changing at a similar pace, it stands to reason that the desktop metaphor is not essential to these practices but was rather merely a useful yet contingent narrative for facilitating technology adoption — a market that is reaching saturation in many areas. In that case, our task is to reveal what else lies in the essential practices that are taking place here — not what metaphors are being used but what is actually happening in these interaction paradigms. The user may instruct the computer via this interface to allocate computational power to particular tasks using graphical metaphors but the practices of allocation are quite literal. Our analysis must address these literal practices that are common between interfaces that use the desktop metaphor and those that don’t.
To begin such an analysis, we must also treat these interfaces in the same manner by which we would treat the practices that preceded it. To unveil what is common between these practices and paradigms we must treat them as spaces, without distinguishing between virtual and actual. The modes by which humans behave in the DE-as-space and where else we find similar modes of behaviour will allow us to put together a genealogy of uses that is more expansive than that portrayed by only the desktop metaphor. As a space, we must recognise that behaviour in the DE is principally one of movement. While computers themselves are indeed computational devices, the commanding of these resources on the part of the human is primarily based on allocation — the strategic movement of resources across spaces.
It is worth noting that in this essay we will be using two, perhaps contrary, philosophical frameworks to guide our analysis — the phenomenology of Heidegger, specifically his writing in The Question Concerning Technology, and the technique of uncovering meaning-as-use from the later writings of Wittgenstein. Apart from the benefits of referencing from both sides of the continental/analytical philosophy divide, both of these philosophies are concerned with examining experiences and practices that are common, as they are found in typical behaviour in various contexts. Yet these philosophies are perhaps too abstract for a practical archaeology. We will therefore be referencing two other, somewhat parallel, frameworks — those of Deleuze and Guattari from The War Machine and Augé’s Non-Place. Deleuze and Guattari’s approach resembles Heidegger in that they are wont to speak about essential characteristics and properties, whereas Augé, as an anthropologist, is concerned more with practices and activities not unlike Wittgensteinian games. A deployment of both these approaches will be necessary to achieve a well-rounded framework for investigation.
Smooth Space
Space that facilitates movement is addressed in Deleuze & Guattari’s chapter in A Thousand Plateaus: Capitalism and Schizophrenia, titled Treatise on Nomadology — The War Machine. The concept of smooth space is crucial for us, which in Nomadology is set in contrast with striated space. Being mutually defining in a semi-dialectical sense, space that is striated is quantified and measured, particularly useful to the State as a system of control, whereas smooth space is more concerned with movement and trajectory. A typical example of this difference is between a city and an ocean — the former is often divided into addresses and static locations, the latter in an unstable plane where movement is inevitable and velocity more of a certainty than position.
[T]he difference between a smooth (vectorial, projective, or topological) space and a striated (metric) space: in the first case ‘space is occupied without being counted,’ and in the second case ‘space is counted in order to be occupied.’
We want to apply the notion of smooth space to the DE as a framework to reveal how its design encourages movement by virtue of its smoothness. If we think of the DE as desktop metaphor, the literally smooth surface upon which we place shortcuts and teacups alike also reveals something of the essential commonality between them. The planing of wood, moulding of plastic, etc is a process of bringing forth (in Heidegger’s sense) a space-creating vector that directs one in every direction along its plane. The balancing of these pulls from all around the plane generates a stable yet fluid space that discourages inertia and stimulates movement. One is therefore compelled to remain stationed and yet move in some sense — to be nomadic upon the stationary smooth desk space. This is achieved in both cases by moving things across the neutral surface and only really engaging with the desktop itself in order to make it smoother and more conducive to nomadic movement, such as moving a table nearer a window for more light, or adjusting the keyboard shortcuts for added convenience. The value of the smooth desktop in practice is in its neutrality, in the extent to which it does not assert its own being but facilitates other tasks.
“It is the cutting of the stone that turns it into material capable of holding and coordinating forces”, as it is a bringing forth from any material that enables it to create a smooth space. Yet in digital virtual space there is no cutting required — the material must be actualised as essentially smooth. It is easy to see this smoothness in the physical desk and how the desktop metaphor seeks to remediate this, but placing the desktop metaphor directly in the context of the DE raises an important question — is computational space not completely striated, Euclidean, digital and inherently measured?
In digital virtual space, every calculation, resource allocation and input is quantified within a meticulously striated framework, else no computational operation would be possible at all. Positioning a cursor on a screen gives it precise coordinates that can then be processed, and the ultimately binary nature of digital media ensures that nothing can be expressed or processed without such precise allocation. However, there is certainly scope to say that all communication, indeed even digital communication, takes place in inherently smooth space.
Let us take a concrete example — that of an error message generated by the DE in response to some human input. Every error report that is prompted and delivered proceeds through a number of different systems with the utmost expedience, as the larger the delay between fault and report the less inclined the user will be to associate the report with the fault and then act on it. There is no need to quantify this time in any other terms than ‘immediately’, which in effect means ‘as fast as an unhindered message can travel — however long that is’ (which is incidentally extremely fast). This is a fundamental of computing and is reflected in various aspects of the DE, but a common instance of this that is particular to the DE is the transition between the ‘empty’ DE space and loading other software.
To refer once again to the analogy of an ocean, the space traversed by this new software or error message, which is little more than an advanced system of wires, could well be seen as a fluid space that is continually in motion. There is little meaning to say that in its transition the message is ever at a particular point within such a wire, not only because of the great speed at which this happens but because of the dependency upon the source and destination of the message for its meaning — in transition the message is little more than an encoded signal that has yet to be decoded, without meaning or identity. How long does it take to load a piece of software? This is obviously contingent upon the hardware being used — processor speed, available RAM and so on. Yet this transition that a human makes from DE space to other software spaces is crucial and non-striated, as non-striated as the provision of an error message.
It is important to note that the speed of transmission in these cases is certainly measurable, as we often see the measurements of processor speeds in technical details for computers. Such speeds relate to the movement, allocation and processing of data internally and certainly affect the speed at which a DE can process something like an error message. Yet once given these mechanical constraints the speed of movement of any signal still occurs as fast as is optimal, and such measurement on the part of the user typically will only take place between sending and receiving such a message. In order to qualify as striated, measurement would need to take place within that transition, rather than across it — and if something can be measured but isn’t, it is effectively smooth abiding by our above definition (“without being counted”). When moving a notepad onto a desk on which to write, the time taken to retrieve and place this item is measurable but ostensibly happens as quickly as is appropriate and isn’t practically measured.
So we can say that both DE and desk are somewhat smooth, but a notable difference must be underlined — the desk is more comprehensively smooth than the DE. The DE is comprised of striated components and necessarily can only compute within striated spaces — in order for inputs into the DE system to be processed at all, they must be striated. The anatomy of the DE contains smooth spaces whereas the desk is itself smooth.
Applying such concepts as smooth and striated spaces to the digital virtual spaces of computing is thus fraught with problems. The biggest problem is that all smooth spaces are traversed veritably instantaneously (effectively invisibly fast) and all user input is striated in order to be processed as information — striation is hence paramount. It seems as though there is little scope available for the occupation of smooth space, due to this invisibility coming about from the high speed of data transmission and resource allocation.
The smooth spaces intrinsic to the DE are traversed so quickly that we may want to pause before calling them space in any understandable sense — when a distance is traversed instantaneously, what meaning does that distance thereafter have? This principle of speed overcoming space is reflected in McLuhan’s “global village”, noting how we are brought closer together by this speed of communication. Smooth space seems to be too narrow a concept to be applied to the DE, as there are a great number of ways in which the DE is more striated than smooth. We therefore need a conceptual apparatus that can treat the space of the DE as a whole, on the surface level where human behaviour takes place, as the computer itself pays little heed to any non-striated space.
Non-Place
Marc Augé developed the concept of non-place based on the work of Michel de Certeau, as the “traveller’s space”, “formed in relation to certain ends (transport, transit, commerce, leisure)”. Spaces such as airports, motorways and corridors are quintessential non-places, the purposes of which are ultimately to facilitate movement. Commonly these spaces are constructed with a certain minimalism, designed to communicate their function with relatively little aesthetic that might otherwise distract from and discourage movement.
If a place can be defined as relational, historical and concerned with identity, then a space which cannot be defined as relational, or historical, or concerned with identity will be a non-place.
This definition may appear a little over-stringent given our stated family of non-places, a good counter-example being the Dostoyevskaya underground station in Moscow that features sizeable artistry celebrating Fyodor Dostoyevsky. This would seem to contradict the notion that such non-places are a-historical, and parallels the tendency for DEs to feature background wallpapers. Developing Augé’s definition we may propose that such spaces are being ‘played’ with, offering a fleeting feeling of place-ness to balance the perhaps frustrating experience of naked transience that such non-places would otherwise offer.
As noted above, the meaning and identity that any signal or resource has is in its allocation and reception — its use once it is no longer in transition and movement. This space of movement, where it exists in relation to static and striated spaces, is our non-place, which is emphasised by the speed of transmission in this computational context.
Does the physical desk also qualify as non-place? Perhaps less obviously. When we consider the physical context of the desk, perhaps placed in a study or office environment, this doesn’t resemble non-place as purely transitional space — people remain in these locations at length and repeatedly, they are destinations in and of themselves. Yet abstracted from that physical context, we can see the desk itself as a stationary non-place where objects, tasks and activities transition across it without supposing the desk as its ultimate destination. We can see the desk as a conveyor belt of sorts, an open-plan motorway where resources are in movement across it and may pause along the way for some manipulation to occur (like within an airport check-in desk or a motorway service station). The DE also has its physical contexts that are quite similar to the desk, such as research or office environments, and similarly depends upon movement and allocation with opportunities to pause for some manipulation along the way.
Returning to Heidegger briefly, non-place is the space left behind after the bringing forth, not unlike a quarry, and non-place is not directly brought forth in the same sense as the components upon which our DE depends. Our DE consists of crafted packages that are brought-forth, but the non-place of the DE is the space that exists between these packages. A motorway or corridor is similarly comprised of crafted, often striated, elements but it is the space between them that is smooth.
With the conceptual framework of non-place, we are in a position to put together a Wittgensteinian family of resemblances, linking the DE with the airport, the quarry, the motorway and corridor. To solidify this family of resemblances, we can look to another quality of non-place that Augé illustrates — that “[t]he link between individuals and their surroundings in the space of non-place is established through the mediation of words, or even texts.”
There is a significant portion of the recent history of typography that concerns computing, with a number of examples of companies commissioning or designing new fonts themselves for use on the DE. Manufacturers of DEs have historically had an interest in typography that is optimised to the technology in question. One example is Microsoft’s utilisation of the Times New Roman font, originally created for The Times newspaper in the early twentieth century, that was adapted for this new use.
There are parallels here with the use of font in transportation icono- and typo-graphic ventures. A notable example of this is what was called the Underground Electric Railways Company of London commissioning the Johnston typeface for use on branded signage across what became the London Underground. Additionally, throughout many English and European cities the Helvetica font can be found on similar signage on roads and streets, which serves this similarly functional iconographic, navigational purpose.
Signage is a crucial element to non-place spaces that facilitate transitions, though it is often a symptom of striation (where the ‘there’ being pointed towards is quantified in a unitary fashion). In order to traverse smooth space effectively and not end up just anywhere, following signs allows a subject to navigate over smooth spaces in order to arrive at their striated destination. This technique is reflected in the DE quite directly and the similarities between road and desktop signage are clear, exemplified by the tendency in both to feature basic, common iconography underlined by a brief textual description.
We are now in a position to say that the DE space is smooth only by virtue of its non-placeness. It’s not enough to merely say that the DE, a motorway, and airport are smooth, because they have numerous components that are deeply striated. However there is a smoothness that comes about at the level of common (or average everyday) use, in the movement and allocation that is the practice of the humans using these spaces.
So by highlighting the non-place of the DE we have something we can point to that is smooth in Deleuze & Guattari’s sense. This smoothness is principally in the absence that lies in between the computer waiting for input and receiving it. In a system so heavily reliant on precise striation the only space available for smoothness is in between operations, and then only where human input is prompted. The smooth space in the DE, as with other non-places, is always traversed as quickly and efficiently as possible, but there is a distinction to be made between space that requires human input to traverse and that which doesn’t. The DE creates a space to receive input from the user, and it is this space that is smooth. In fact only for the human, to whom the computer is a mere tool or even ‘part of the furniture’, can this smooth space be occupied.
Current and future personal computing
In summary, we have concluded that the Desktop Environment as space can indeed be described as smooth by virtue of its being non-place — we must regard the DE as non-place in order to make such a statement about its smoothness. We are now also in a position to say that the desktop metaphor isn’t the whole story when discussing DEs in how they encourage and afford behaviour paradigms — design practices in building non-place spaces form a more crucial connection between such paradigms and will likely be used much further into the future. In fact on the level of common, individual use we might want to say that the DE has more in common with the motorway than it does with the desk.
This constitutes a clear example of the affordances that digital virtual spaces provide — that the visual metaphors for interaction don’t depend on an exact surface resemblance to what is being functionally performed. This is a consequence of what Hayles refers to as “[a]bstracting information from a material base [meaning] that information could become free-floating…” Looking ahead, as skeuomorphic references to the desk are progressively dropped, what is essential to the DE and to what it owes more genealogically is in fact the concept of non-place.
We have certainly not arrived at the conclusion to the genealogy of the DE, which could conceivably lose its reliance on the desktop metaphor completely as skeuomorphic dependencies are alleviated. Currently touch-based operating systems, which are predominantly moving away from the desktop metaphor, consist of another branch of non-place virtual spaces that perhaps won’t completely supersede the classic DE — indeed it seems likely that these two branches of virtual spaces will continue to co-exist and influence each other for the time being.
Additionally, the increasing expansion of cloud computing technology is enabling the further development of sophisticated voice-controlled systems, such as Siri, Cortana and Google’s voice interaction technologies. While we are not in a position to speculate about such systems existing without any visual interface at all (though some use cases do exist they are embedded into systems and devices that do rely on visual interfaces) the notion of an Audial User Interface is likely to become a new kind of digital virtual space where non-place and smoothness will take different forms, meanings and uses.
