Читать книгу Being with Data - Nathaniel Tkacz - Страница 9

What kind of thing is a dashboard? Beyond an initial definition (data, display, cognitive function), what category of thing does it belong to? What kind of conceptual apparatus is able to handle its specificity; its diverse historical threads and peculiar ontological status? Because a dashboard mediates our relation to data, should we approach it as a medium? Or is it only one component of a medium or larger system? Is it to be understood instead as an interface, or perhaps as software? A dashboard may very well fit definitions of all these things and I will draw on them throughout the discussion as appropriate.

But while a dashboard mediates and this term (‘medium’) can act as an inclusive shorthand for all the things that constitute a dashboard, the term ‘medium’ no longer affords the analytical precision it once did. During the twentieth century, one could speak of ‘radio’, ‘television’, ‘film’, ‘print’ and so on, with some degree of confidence regarding the ontological distinctness of each. These diverse, mostly analogue technologies were given further coherence through the standardization that came along with mass production (and consumption). As is taken for granted in the field of media studies, the arrival of the computer as a communication medium, or what is broadly referred to as ‘the digital’, radically transformed understandings of media.¹⁶ The computer is able to absorb and ‘remediate’ the majority of previous media, altering their ontology (or materiality) while preserving their function and adding new qualities. As Lev Manovich has noted, the computer has become something of a ‘metamedium’, having subsumed all previous media.¹⁷ This metamedium is generative of dynamics that did not exist in previous media and/or is able to translate techniques from one ‘medium’ into a generic property of many others (such as ‘copy/paste’ or ‘zoom’).

The rise of the computer as a metamedium has also resulted in a shifting of focus regarding the territories of inquiry, which no longer adhere to the categories of analogue media (print, television, radio etc.). Manovich and others suggested a focus on software, while hardware, infrastructure, networks, platforms and interfaces have also all emerged as distinct areas of inquiry. These have not emerged through a master plan and thus they criss-cross and overlap in any number of ways. Software may include or even be an interface, while a platform may include hardware, software, infrastructure, interfaces and so on, plus a business model. I have no interest in redrawing the boundaries of these territories, but I do want to stress that despite any overlap, each does different kinds of intellectual work and makes possible forms of inquiry that at some point go in different directions. Each has its own historical trajectories, is attached to different forms of expertise and has its own privileged objects. All this is to note that there are methodological ramifications for assigning these terms to the things that spark our interest.

Contemporary dashboards are typically digital media and thus fall under the many logics of the ‘metamedium’. They are comprised of software and hardware; they have an infrastructure, including the aforementioned user-facing software and hardware as well as ‘background’ information systems or cloud services; they are typically networked; they may be found on platforms, be fed by platform data or otherwise be part of platform strategies; and they are constituted by a number of interfaces, with the graphical display the most visible and relevant for the current inquiry. Each of these elements of dashboards is significant in its own right and I will lean on them from time to time.

When I first began studying dashboards, I did so primarily through approaching them as a specific category of software or as a type of interface. Together, these were generally adequate, but neither quite did the work that was needed. There were two problems. Dashboards clearly exceeded the definitional limits of both of these terms, and neither got to the heart of precisely what dashboards are doing to data and the cultures in which these data circulate. For example, the majority of contemporary dashboards are interfaces and the people who make them either use ‘interface’ as a native term or are at least familiar with the notion. Interface positions the dashboard within the history of computation and general reflections about humans and machines. While Branden Hookway rightly locates the origins of the term ‘interface’ in thermodynamics,¹⁸ its history really gets going (for my purposes at least) with mid-twentieth-century talk of ‘man–computer symbiosis’,¹⁹ ‘man–machine interfaces’²⁰ and the general emergence of human–computer interaction (HCI). Hookway’s claim that the interface is first and foremost a relation, where two things are brought together and where specific qualities of each are privileged and ‘augmented’, no doubt explains a lot of what dashboards do. However, if I was to take this as my primary orientating term, I would have to begin my study with dashboards as they appeared in computer systems, and this didn’t happen until the 1960s, which is much too late. I would also have to account for how the term dashboard ‘moved’ from carriages and cars to computers and devices, most likely by reference to metaphor or perhaps simulation. Part of what I want to suggest about dashboards, though, is a continuity across different media over time that is not mere metaphorical transferral. While interfaces dominate today’s dashboard landscape, the relation is not one-to-one over time. One could of course simply refer to pre-computational dashboards as interfaces, but to do so would seem overtly anachronistic (although I do recognize one cannot escape this issue of language and history entirely). Likewise, while the term ‘software’ does a good job of pointing to the contemporary analytics industries (with their different software offerings) and the technical specificity of most dashboards, it too has this problem of the dashboard’s pre-computational history.

As the inquiry developed, I increasingly found myself thinking with another term, not unrelated to ‘interface’ and ‘software’ or the other terms mentioned, but one that seems untroubled by the dashboard’s peculiar history or by the fact that dashboards can be comprised of radically different configurations of technology. More importantly, the term more strongly connotes what it is a dashboard does. Whatever else a dashboard is, I want to suggest it is a format. A dashboard is a format and the work it does is one of formatting.

What is a format? My understanding of format draws on three of its most common definitions:

1 the way in which something is arranged or set out;

2 the shape, size and presentation of a book or periodical;

3 computing: a defined structure for the processing, storage or display of data.21

A format is an arrangement, a way of arranging things. This arrangement includes shape, size and presentation, though (in our case) not limited to print media, and involves the processing, storage and display of data. While the third definition of ‘format’ also comes about with the rise of computation (like ‘software’ and ‘interface’), the other two definitions predate this one and together already contain much of the specificity I am looking for. Etymologically, ‘format’ comes from the Latin formatus, a past participle of formare (‘to form’). In book publishing, it originally referred to the size of a book (folio, quarto, octavo) and the ‘leaf’ folding and print strategies required for each size.²² While acknowledging this historical specificity, I will use this notion of format to approach the dashboard, and thus my emphasis is on the arrangement of elements (including size, shape, presentation and more) and the structuring of data through processing, storage and display, with a particular focus on display. I favour the use of ‘format’ over that of ‘medium’, ‘interface’ or ‘software’ because I want to highlight these questions of arrangement, presentation, structure and display.

A format is distinct from the medium, software, interface or other technology that realizes it. While some formats are limited to a particular medium, many even proprietary formats are compatible with different medial configurations. Common digital text formats such as .docx (Microsoft Word Open XML format) or .pdf (Portable Document Format), for example, can be read by many different kinds of software and their files ‘opened’ on different operating systems and devices. A format is thus dependent on some configuration of media elements, but it is not reducible to these elements. A format is ‘medium agnostic’.²³ To focus on formats thus opens onto different historical trajectories where the advance of specific technologies need not take priority. That is, the history of a format is distinct from that of any particular medium. It may for a period run in parallel to the development of a medium – and we will observe this with cars and computers – but the format may just as well part ways with one medium and take up with another or persist across different configurations simultaneously.

This understanding of format (as distinct from medium) is in part indebted to Jonathan Sterne’s excellent work on the MP3 format, where he similarly writes that ‘some formats may offer completely different inroads into media history and may well show us subterranean connections among media that we previously thought separate’.²⁴ Indeed, while ostensibly a book about the MP3 format, Sterne’s study opens onto a much longer history of sound compression and perceptual technics from which the MP3 format becomes possible as both cultural and technical artefact. In the pages that follow, I too give a lot of space over to historical discussion, to what I call format archaeology, charting a course across technological hosts of the dashboard format in different times and places.

Further commenting on the relationship between format and medium, Sterne writes, ‘Format denotes a whole range of decisions that affect the look, feel, experience, and workings of a medium’, and adds ‘it names a set of rules according to which a technology can operate’.²⁵ He thus raises questions regarding the dependency of media on formats and wants to highlight their aesthetic, phenomenological and functional importance. Sterne’s ‘format theory’, as an alternative to established ‘media theory’ (and the other terms mentioned above), has been very helpful in thinking about dashboards for these reasons. However, my usage diverges in a number of ways and contains a number of different emphases. The object of Sterne’s analysis, the MP3 format, is much more strictly codified. It is a format to be read by software. Thus, when he writes about format as a set of rules for which a technology can operate, he means this in a rather strict sense as code. In this way, his working definition runs very close to the third definition given above, where format relates to computers and specifically to the storage and processing of (sound as) data. However, Sterne is also clearly drawing on yet another definition of format relating specifically to sound, which I did not include in the initial three common definitions: ‘the medium in which a sound recording is made available’.²⁶ Obviously, this definition further muddies the waters regarding format and medium, but it clearly informs Sterne’s work, concerned as it is with sound. While Sterne’s ‘format theory’ is a conceptual elaboration on this (fourth) definition of format, my own account draws more on the other three and especially the first, most general one: ‘the way in which something is arranged or set out’.

As software, a dashboard may very well contain coded elements, but the notion of format that informs this study (that is, the dashboard as format) is generally looser, as likely to be a matter of interpretation as of technical rules for operation. As format, a dashboard is a way of arranging or setting things out. Exactly what is arranged, how and for what purpose changes over time. The development of a format weaves a history across different technologies (horse and carriage, car, reports, computer systems, among others); some elements of the arrangement persist, either in terms of function or connoting a ‘look, feel [or] experience’, while others fall away.

Format also has an interesting ontological status, seemingly as a thing in itself and as a kind of container for other things, which, when formatted, take on specific qualities yet without losing their prior identity. To put it another way, the same thing can be formatted differently. Data can be formatted in different ways, in a table, in a graph, scribbled in a field diary or displayed on a giant sports scoreboard, without losing their status as data. And yet, the format is not without ontological and epistemological influence. In fact, I hope to convince the reader that format matters deeply with regard to these things. Understanding how the dashboard as a format, through its arrangements and set-ups, its presentations and displays, works over the data that pass through it and in doing so constitute it, underpins much of this inquiry.

This ontological blurring resulting from the dynamic interplay between data and format is partly captured in how ‘format’ acts as both noun and verb. A dashboard is a format, but one can also format data into a dashboard. As a verb, ‘format’ suggests an active reworking, a transition in state; it is the work required in order for ‘format’ to function effectively as a noun. As a rule, there is no format without a priori formatting. This book takes a strong interest in the formatting of data through dashboards, but other formatting work is also going on. The dashboard is layered with perceptual orientations, cultural significations and epistemological qualities that format not only the data and elements that pass through it, but also the bodies and organizations that make use of it. How far one can stretch the notion of formatting beyond a direct encounter with a format and what kind of force one can grant it once placed alongside competing forces – however conceived – are empirical questions that I wish to explore but do not expect to resolve. This bidirectionality of formatting, such that what is formatted is not only data but also the people, things and situations within which a format is at work, brings us close to the use of the term by Bruno Latour and Michel Callon.

Over thirty-odd years, the notion of formatting pops up in several places in the work of Latour. In an early co-authored piece with Callon, the two use the concept of formatting in at least three senses in order to develop their own approach to the study of capitalism, markets and economic science (as a discipline).²⁷ They use the idea to describe the performativity of economic categories: ‘the notion of formatting refers to the always ongoing and effective performance of economic categories, which are definitely real as long as they are networked to other mechanisms that not only do not hide them but actively contribute to their reality’.²⁸ They use it to refer to the performativity of the practical methods of economic science:

even the most theoretical economic science is constantly performing this formatting operation, for its work is much more practical, down to earth, and effective than it would admit. Indeed, economic science extracts from the mobilization of people and things the necessary elements to make the exchanges calculable – and it is precisely this set of operations that we have in mind when we use the notion of formatting.²⁹

And they use it as a way to get at the relation between capitalism and markets: ‘capitalism refers to the formatting of markets’.³⁰ Overall, we get a sense that formatting refers to the ongoing performativity of things. It is the ‘how’ of performativity. It refers to the method of framing some thing or activity in a specific way such that it is amenable to a pre-existing epistemic order (i.e., economic science). And it is used to define how one thing relates to, or acts upon, another.

In later writings, Latour uses the concept once again to highlight the active, performative role of knowledge construction, but now as part of a negative contrast: ‘The notion of “fact,” let us recall, had the disadvantage of not taking into account the enormous work of shaping, formatting, ordering, and deducing, needed to give the data a meaning that they never have on their own.’³¹ Here, the idea refers to the active work needed to give data meaning; work that can go unnoticed once something attains the status of fact. Latour tends to use this notion of formatting to point to the practical elements of knowledge construction that are easily overlooked, though not invisible or hidden. Part of what the term ‘formatting’ does, as a verb, is ascribe these practical and mundane elements some force and an active role in the production of realities.

Like Latour and Callon, I too see formatting as a performative notion, able to describe the ‘how’ of one thing’s relation to another. I like the way Latour places it within a context of knowledge production and the overlooked (but not invisible) elements that feed into and define more finished products (like facts). However, my own use is more literal and specific than Latour’s. I want to stick much more closely to the notion of format when I invoke formatting. Formatting refers to the bidirectional performativity of a format as it is enacted (or ‘used’, in more common parlance). It refers to how a format acts on the data and other elements that pass through (and thereby constitute) it and on the people, devices, and organizational time-spaces within which it is deployed. To describe the active work of a format as formatting suggests a specificity to its performativity. Formatting is not a free-for-all; it has limits, its ‘tendencies’ in terms of what it can achieve in any given situation. What a format does is carry along these specificities, limits and tendencies; they are in part what constitute it as a format; they give it its consistency and make it recognizable. Understanding these specificities does not allow us to grasp the entirety of what’s going on when a format is at work (a difficult and questionable aim in any case), but can serve as a privileged point of orientation. It enables contrast and comparison of different empirical instances and permits one to rise up, if cautiously, and posit certain generalities and regularities (carried forth as tendencies in a format) among such instances.

To get a sense of the work of formatting and how significant a format can be to societal-level organization, consider what is in some ways a precursor to the dashboard: double-entry bookkeeping (DEB). While DEB is known as a method for keeping accounts, it also serves as an excellent example of a format (or rather, a number of closely related formats). Consider the striking parallels between how I have defined a format thus far and how DEB is defined by Hans Derks:

a way of organising, arranging, registering accounting data of ventures (public and private) that informs several users, in particular the owner(s), to improve decision-making, to uncover gain/loss, to track the entity’s rights and obligations, to recall past activities, and to achieve a greater control or surveillance of transactions internal and external to the firm or other institutions.³²

As the name connotes, DEB involves entering financial transaction numbers into account ledgers (a format), with every transaction entered twice, appearing as a debit entry in one account (in a leftsided column) and as a credit entry in another account (in a rightsided column). Through entering every transaction as both debit and credit, the system should ‘balance’ and this acts as a safeguard against errors. Ostensibly, DEB involves arranging numbers for the purposes of keeping account. The numbers are arranged and ordered (formatted) through the ledger, and this imbues them with a specific value and meaning. The numbers become ‘accounting numbers’, representing financial values (credit and debit) in the context of business. However, a number of historical studies have shown there is much more to this format.

While the double-entry of numbers is core to this method of accounting, it is but a component of a larger system comprised of different books or journals (typically three). Historically, DEB made use of a daybook, a journal and a final ledger of accounts, and one could approach each as a format in its own right, but for the sake of brevity I treat them together. As the names suggest, each book or journal serves a different purpose within a larger system, moving from more descriptive records of daily transactions (daybook) to more abstract forms of numerical representation (ledgers). In her excellent history of the modern fact, Mary Poovey has argued that DEB played a crucial role in the becoming factual of numbers; that is, in numbers’ capacity to appear beyond matters of interpretation.³³ Poovey was precisely interested in this question of how it came to be that some things appear as available for interpretation while others do not, or as she puts it, in ‘how description came to seem separate from interpretation or theoretical analysis; the story of how one kind of representation – numbers – came to seem immune from theory or interpretation’.³⁴ DEB is crucially important for Poovey as it produced a kind of proto-factual form of numerical representation. This type of numerical representation was achieved by the parsing of economic activity through DEB’s system of books and from more detailed to abstract modes of writing numbers. This movement, which Poovey describes as one from narrative to number, imbued these numbers (and their related accounting practices) with specific capacities and a kind of authority. While I cannot convey the full extent of her argument here, Poovey positions DEB in contrast to Ciceronian rhetoric (influential in the late fifteenth century) and shows how the form of numerical representation found in DEB diverged from and challenged this dominant form of rhetoric. That one could move from narrative to number (and back) in DEB, for instance, ‘seemed to make writing transparent instead of performative, as rhetoric so obviously was’.³⁵ She adds, moreover, that DEB ‘helped undermine the epistemology and social hierarchy associated with rhetoric’ by challenging Ciceronian scepticism through its system of representing numbers (that is, through its format). This was observable in the apparent transparency of the numbers themselves, but also, for example, in how DEB enabled ‘writing positions’³⁶ which were not reliant on the same social hierarchy as that of rhetoric for their authority.

Poovey further argues that DEB didn’t do away with rhetoric entirely, but rather repositioned rhetoric as a component of the system, visible in different elements of its formats: ‘The system of double-entry books displayed virtue graphically, by the balances prominently featured at the end of each set of facing ledger pages and by the order evident in the books as a system.’³⁷ The idea that late fifteenth- and sixteenth-century DEB was structured as a work of rhetoric is supported by others, such a James Aho, who has pointed out that everything from the ‘properly humble and thankful Christian salutation’ atop the ledger, to the ‘procedure to invent material for inclusion in the ledger’ or how the elements of the accounts relate to each other, all correspond to specific components of (largely) Ciceronian rhetoric.³⁸

We see in DEB a specific formatting of economic activity; one that produces distinct kinds of numbers with an epistemological value that seemingly no longer requires interpretation and appears transparent to those who cast their eyes upon them. We also see, however, that whatever formatting work DEB accomplished, its formats are themselves pulled in different directions and take on different characteristics over time. The early DEB that Poovey and Aho consider was strongly influenced by rhetoric; it was literally arranged as rhetoric. Oddly enough, it was this arrangement that laid the grounds for numbers’ apparent transcendence of the epistemological conditions of rhetoric. While Poovey doesn’t straightforwardly equate DEB’s numbers with facts, she claims that in DEB the modern fact existed in ‘embryonic form’. DEB’s formatting produced a way of relating to numbers that foreshadowed the facts of modern science. While I find Poovey’s work compelling, one does not need to accept her argument about DEB and modern facts to get a sense of how a format is able to produce, through its arrangements, presentation and display (among other things), numbers with a distinct epistemological value. One can also see how a format is itself shaped by the forces of the day (in this case, Ciceronian rhetoric), in ways that might not be immediately obvious but clearly influence its arrangements and the qualities of things being arranged.³⁹

For Poovey, the numbers found in DEB are significant in their role as precursor to the ‘epistemological unit that has dominated modernity’, namely modern facts.⁴⁰ Well before Poovey’s epistemologically orientated inquiry, however, many others had similarly argued for the significance of DEB vis-à-vis the broad trajectory of modern society. Max Weber identified DEB as a key stage in the development of modern rationalization and a necessary ingredient of capitalism. Joseph Schumpeter and Werner Sombart agreed with Weber on DEB’s rationalizing function and its importance for modern capitalism. Out of the three, Sombart in particular accorded DEB the highest significance, declaring that ‘capitalism and double entry bookkeeping are absolutely indissociable; their relationship to each other is that of form to content’.⁴¹ In a passage that reads like a precursor of German media theory, Sombart adds the following regarding the notion of capital: ‘The very concept of capital is derived from this way of looking at things; one can say that capital, as a category, did not exist before double-entry bookkeeping.’⁴² Since these Germanic reflections on DEB, more weight has been given to its theological and rhetorical role as opposed to its assumed rational and technical superiority, though its significance as system, practice and indeed format is not disputed.⁴³

If Poovey considered the modern fact as the dominant epistemological unit for the period of modernity, besides playing a crucial role in this history, the influence of DEB extends well beyond Poovey’s concerns, underpinning centuries of commerce while also providing a moral basis (through its rhetorical function) for commercial activity itself.⁴⁴ We can see that formats have the power to radically alter epistemological systems; not just what is considered knowledge (or not) but more fundamentally how one approaches and perceives items of epistemological significance. As Poovey shows, formats can also lay the ground for further transformations in the things that have undergone formatting, in her case the unleashing of a new potential ‘facticity’ within numbers that would ripple across early modern science. The importance others have placed on DEB in relation to rationalization and capitalism further confirms this capacity to format the social.

I introduce DEB to help flesh out my conception of format (without anticipating the material to come) and to communicate the significance of certain formats regarding the contours of societies past and present. While it would be naive to consider DEB as only a format or number of formats, or to attribute its historical role to the work of formatting alone, it would be equally naive to underestimate the significance of arranging numbers (or code, data, text, images) time and time again and in standardized ways; ways that can travel and insinuate themselves into the routines of daily life, habituating forms of being and knowing, interpreting and perceiving.

DEB also provides an interesting contrast to the dashboard as a format. Like DEB, the dashboard has found a home in commerce and, in particular, the management of large and medium-sized organizations. As we shall see, with at least one of the historical trajectories belonging to the dashboard format, dashboards are presented as a development from mid-twentieth-century accounting, specifically in relation to the rise of ‘industrial accounting’. One finds within the dashboard format a new relation to numbers and data, one that includes not only the financial numbers typical of accounting, but also non-financial numbers and forms of representation no longer in the service of balancing the books or auditing. Like DEB, dashboards are also loaded, or layered, with epistemological and broader cultural significance. If DEB served a rhetorical function, adding legitimacy to commerce (at a time when profit on lending money or ‘usury’ was prohibited) through the production of numbers that were always ‘in balance’, the dashboard too is caught up with technological and managerial ideals. And if DEB provided ‘writing positions’ and accompanying forms of authority derived from its system, so too does the dashboard privilege certain ways of knowing and doing, certain types of organizational arrangements, and certain ways of representing data and numbers. The dashboard is no straightforward replacement of the formats of accounting and I will not speculate as to whether it will attain the historical significance of DEB. The dashboard criss-crosses with accounting in certain historical moments, but also contains within it layers from different historical trajectories and present tendencies that reach far beyond the numerical representation found in accounting ledgers. The dashboard has already been with us for over a century, slowing, stabilizing and recursively adapting to new technological configurations, absorbing the visions of its designers, engineers and the cultures which put it to work in predictable or unexpected ways. If DEB produced a form of ‘noninterpretative’ numerical representation that was able to travel and become a bedrock of modern science, it is not so much the travelling of data in dashboards that is of interest, but rather the travelling of the dashboard format itself.

The notions of format and formatting provide the conceptual underpinning of this study of data and dashboards. They are how I hold the two together (data and dashboards) and inform many of the specific observations I will make about each. As much as this is a study of data and dashboards, it is a study of formats and the work of formatting.