LINDA HILFLING RITASDATTER
UNWRAPPING COBOL
Lessons in Crisis Computing
The performance lecture UNWRAPPING COBOL: Lessons in
Crisis Computing, unfolds through the thesis’
s interdependent con-cepts of Crisis, Execution, and Maintenance, covering key aspects of
COBOL ’S development and use. More than a history of this suppo
-sedly outdated programming language, the lessons serve to outline the underlying frictions and asymmetries in global flows that are
otherwi-se thought of as
seamless. The lessons developed iteratively between 2012 and 2020 and take
the form of performative lectures as well as a variety of other formats including slide-shows, magazine articles, installations, and videos. The lectures have been presented at locations and events such as
Winchester Gallery in 2018, T ranspiksel South American Tour
, Peru in 2017, Overgaden Institute of Contemporary Art, Copenhagen in
2016, Archaeologies of Media and Technology Research Group at
Winchester School of Arts in 2016, Academy of Media Arts Cologne in 2014, Paletten Magazine in 2013, the Goethe Institute Bangalore in
2012, and the transmediale festival, Berlin in 2012. MALMÖ UNIVERSIT Y 205 06 MALMÖ, SWEDEN WWW.MA U.SE isbn 978-91-7877-115-8 (print) isbn 978-91-7877-116-5 (pdf) SC HOOL OF ART S AND C OMMUNIC A TION DISSERT A TION SERIES
Doctoral Dissertation in Interaction Design
School of Arts and Communication Dissertation Series Faculty: Culture and Society
Department: School of Arts and Communication, K3 Malmö University
CC* Linda Hilfling Ritasdatter, 2020 Designed by Linda Hilfling Ritasdatter Layout by Nina Gribat
Artworks and photos by Linda Hilfling Ritasdatter unless stated otherwise Copy Editor: Janet Leyton-Grant
Supported by grants from The National Dissertation Council and The Doctoral Foundation
ISBN 978-91-7877-115-8 (print) ISBN 978-91-7877-116-5 (pdf) DOI 10.24834/isbn.9789178771165 Print: Holmbergs, Malmö, Sweden, 2020
* This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International.
LINDA HILFLING RITASDATTER
UNWRAPPING COBOL
Lessons in Crisis Computing
I once worked on a mainframe computer system whe-re the fan-folded listing of my COBOL program stood as high as a person. My program was sixteen years old when I inherited it. According to the library logs, nine-ty-six programmers had worked on it before I had. I spent a year wandering its subroutines and service mo-dules, but there were still mysterious places I did not dare touch. There were bugs on this system no one had been able to fix for ten years. There were sections where adding a single line of code created odd and puzzling outcomes programmers called “side effects”: bugs that come not directly from the added code but from some later, unknown perturbation further down in the pro-cess. My program was near the end of its “life cycle”. It was close to death.
Yet this system could not be thrown away. By the time a computer system becomes old, no one completely un-derstands it. A system made out of old junky technology becomes, paradoxically, precious.
ACKNOWLEDGEMENTS ... 11
INTRODUCTION ... 17
APPROACHES ... 33
LESSON I: EXECUTION ... 57
1.1 INTERFACING EXECUTION,
BANGALORE SEPTEMBER 2014 ... 65
COBOL, A COmmon Business-Oriented
programming Language ... 67
The Inherent Paradox of User-Friendliness ... 68
Commanding without Command ... 75
Cobol as a Panoptic Tool ... 80
1.2 DISPLACEMENT ... 83
Automatic Management ... 86
1.3 EXECUTING/EXECUTED COBOL ... 89
Perform Until ... 89
Lack of Charisma ... 90
LESSON II: CRISIS ... 97
2.1 BUGS IN THE WAR ROOM ... 103
Bugs ... 105
War Rooms ... 107
Anxious Flows ... 117
Crisis in Progress ... 122
Economies and/of Execution ... 124
CONTENTS
LESSON III: MAINTENANCE ... 129
3.1 PROLOGUE: HUMAN FACTOR ... 135
Conflations ... 135
Human Factor versus Human Factor ... 138
3.2 A DRAB; A DRIVE: RE-VISITING STUDIES ON
MAINTENANCE . ... 149
That Which Sustains ... 149
Improvisation, Creativity, and Complexity ... 151
Beyond Debugging ... 154
Eternal Revisions ... 156
Enacting Ontological Order and Stability ... 157
An Inherent Paradox ... 159
Maintaining Networks and Data ... 162
3.3 BROKEN / UNBROKEN ... 167
Surfacing Maintenance Work ... 167
On Being a Favorite Example ... 170
Which Broken World? ... 171
3.4 MAINTAINING LEGACIES ... 179
Escaped Obsolescence ... 179
Media Undead ... 183
Undead Systems—A Majority of Silence ... 184
Maintaining the Undead ... 186
Zombies and the Silent Majority ... 191
3.5 MAINTAINING FLOWS ... 201
On the Production of Spaces of Flows / Global
Space ... 202
Global Education Center, Mysore ... 203
Universal, Virtual and Specific Bodies ... 205
Bodies in Universal “Real” Space ... 207
Institutionalization ... 211
3.6 MAINTAINING DEVELOPMENT ... 219
IV: CRISIS COMPUTING ... 233
4 CRISIS COMPUTING ... 243
Dreams of Full Automation:
The Fourth Industrial Revolution ... 247
Encoding Displacement ... 251
Internalization ... 255
Asymmetric Arrangements ... 257
Unsettling Ontologies ... 261
REFERENCES ... 267
A COMPLETE INDEX OF ALL ELEMENTS LEADING
TO THE END OF THE WORLD (IN THIS THESIS) ... 291
ACkNOWLEDGEmENTS
For this project, I am foremost indebted to my COBOL teachers: the educators, engineers, and students of COBOL in India who have shared their expertise, experiences, and stories of COBOL with me. I would like to thank my supervisors Susan Kozel, Maria Hellström Reimer, and Florian Cramer for their tremendous support and en-couragement throughout this thesis project’s different stages. Susan for your precise comments and engaged critique that has been instru-mental to the continued development of my arguments. Furthermo-re, I am thankful for your gentle push to make me take on teaching in embodied interaction design, which I enjoyed and learned so much from. Maria, when I started as a doctoral student, you encouraged me to approach the PhD as a kunskapsresa (journey of knowled-ge), which I should embark on and see where it would take me. I’m ever thankful for your encouraging of this freedom of perspective, without which the project would have been unfeasible. Florian, thanks for your immense support during all these years, which star-ted back when you were the supervisor for my M.A. studies in media design. I have always enjoyed your uncompromising and ever enga-ging insights and feedback, not to mention your unflinching belief in my projects that has helped me keep up the motivation throughout this extensive process.
Big thanks to Matthew Fuller, for your generous commitment when undertaking the role as opponent for my 90% seminar, your thorough, engaging and inspiring comments and critical inputs at that stage were defining for the wrapping-up of the project. For early stage feedback, I would like to thank Temi Odomosu and Tobias Denskus who provided important reflections during the first semi-nar presentation of this project. I am also thankful to my green-light readers, especially Tina Askenius for the detailed feedback.
Gratitude goes out to my fellow PhD candidates at the K3 School of Arts and Communication, Malmö University: Jacek Smolicki, Eric
Snodgrass, and Mahmoud Keshavarz for the vibrant discussions and talks and seminars we shared in the early stages of my project, to Anuradha Venugopal Reddy for your ever thoughtful and engaging remarks and comments, Michelle Westerlaken, Marika Hedemyr, Er-liza Lopez Pedersen, Alicia Smedberg, Johan Farkas, Therese Hell-berg, Veera Virmasalo, Sara Gottschalk, Hugo Boothby, and Roel Roscam Abbing for your wonderful support and for creating such an accommodating work environment.
Many thanks to the many other inspiring colleagues with whom I have taught and learned from over the years: Kristina Lindström, David Cuartielles, Nikita Mazurov, Temi Odomosu, Gunnel Petter-son, Tony OlsPetter-son, Daniel Spikol, Anne-Marie Hansen, Pia JönsPetter-son, Håkan Magnusson, Elisabeth M. Nilsson, Berndt Clavier, and Jonas Lövgren among many others. And great thanks to my students, who-se challenging reflections and commitment have been an absolutely life-saving contrast to the isolation of writing.
Thanks also to Per-Anders Hillgren and the supervising committee. In this context I would like to thank the K3 management who has been generously supportive and shown dedication to us doctoral students: Sara Bjärstorp, Åsa Elgameil, Susanne Lundborg, Cecilia Hultman, Erik Källoff, Fredrik Lindström, Carina Listerborn and especially Ulrika Sjöberg. Moreover, I thank Simon Niedenthal and Micke Sve-demar for introducing me to your colleagues in Bangalore and the IT staff for your patience with my Linux partitions. Also, many thanks to Janet Leyton-Grant and Nina Gribat, who both came in at a late stage for vital copy editing and layout respectively.
Outside the walls of my home institution, I am grateful to Olle Essvik and Joel Nordqvist from Rojal Förlag, for our invaluable collabora-tion on the Endless Endtime—A Complete Index of All Elements Le-ading to the End of the World book project. Olle, your bookbinding is marvelous, and I have enjoyed your wonderful company during our workshops and presentations. I am looking forward to this pro-ject’s infinite continuation!
As the reader will notice, the concept of execution is central to my artistic practice and research and the initial reflections on this were triggered early on in the thesis project at The Participatory Informa-tion Technology Centre of Århus University. Great thanks to
Christi-an Ulrich Andersen Christi-and the Centre for the kind invitation to present my work there in the summer of 2014. At this seminar, I had the pleasure of meeting Winnie Soon, with whom reflections on execu-tion as a means for radical artistic intervenexecu-tions flourished and the ideas of developing a collaboration between Århus University and K3 emerged. Back at K3, these ideas were developed even further in conversations with my PhD colleague Eric Snodgrass and together with Helen Pritchard, Magda Tyźlik-Carver, and many others, these conversations evolved into what became the Critical Software Thing Group. I am deeply thankful to this group for the conversations and discussions we have had among us and with various audiences. Thanks also to Geoff Cox, Søren Pold and Christian Ulrich Andersen at Århus University as well as Bo Reimer at K3/Medea for supporting the group. Furthermore, thanks to Wendy Hui Kyong Chun, Susan Schuppli, and Femke Snelting who as part of the group’s “*.exe” events have all contributed to my research with much treasured re-sponses and comments. Especially thanks to Susan Schuppli, your incitement for me, during a stroll through the city of Malmö, to ar-ticulate a concept of my own in my thesis has been decisive for how I have ended up enunciating this research through the concept of Crisis Computing.
The Bugs in the War Room exhibition was hosted by Overgaden Institute of Contemporary Art in Copenhagen in 2016. In relation to that I would like to thank Merete Jankowski, Anna Holm, and Trine Bork for such a supportive and engaging opening of the gallery doors to some Crisis Computing. In 2018, the Winchester School of Art hosted the third installment of the Endless Endtime project with my exhibition Now we have Proof! Here, I would especially like to thank the head of the Winchester Gallery, August Davis along with Ryan Bishop, Ed D’Souza, and Noriko Suzuki-Bosco for your ama-zing support and commitment. Also, thanks to everyone who over the years has supported and encouraged artistic installments of this project including especially the Piksel Festival (Gisle Frøysland and Maite Cajaraville) and the Asimtria Festival in Peru, Sensorium LiU with Norrköping Konstmuseum. Mathias Kokholm with Antipyrine Bookshop at Århus Kunsthal along with many others.
A very special note of gratitude goes out to Zeenath Hassan for your invitation to Kristoffer and myself to join you at the Srishti Institute
of Art, Design and Technology in Bangalore back in 2010. You intro-duced us to Meena Vari and Prayas Abnihav who became paramount to the initialization and realization of OutResourcing. This was an artistic research and exchange project which supported the develop-ment of the thesis project in its very early stages and also involved collaborations with the Goethe Institute in Bangalore and transme-diale in Berlin.
Thank you Padmini, for your enthusiastic involvement and unbeata-ble COBOL expertise and to your family for opening their home to me. Also, thanks to Bhargava, for your openness, passion and love of COBOL: I sincerely regret your much too early passing.
A substantial part of this project was written and developed in my studio in Gerichtshöfe, Berlin and I would like to thank my studio comrades, especially Wolfgang Spahn and Antti Pussinen for our many chats and coffees together. A further thanks to Christian Fau-bel, Darsha Hewitt, Brendan Howell, Sreejata Roy, Joy Mrityunjay Chatterjee, Sebastian Lütgert, Tsila Hassine, Tina Tonagel, Prof. Di-nesha, Pelle Ehn, Nikita Mazurov, Leslie Dunton-Downer, Archana Prasad, Jo Zahn, Lotte Løvring, Prayas Abnihav, Martin Conrads and Fredrik Svensk for inspiring conversations and inputs at diffe-rent stages of the project.
Furthermore, I am deeply grateful to my dearest friends Nina Gribat, Sheena McGrandles, Vanja Larberg, Laila Lørup, Anders Aarkrog, and Line Strandgaard for always being there, and for the unconditio-nal support from my lovely family: Mom, Dad, Rita and Preben, my brothers, Lars and Niels and wonderful nieces and nephews, Sofus, Signe, Sigurd, and Asta-Fernanda.
And last but certainly not least, I thank the two people who are the-single-most important for this project’s realization, Kristoffer and Wilbur. Wilbur: you have been subjected to my COBOL lessons your entire life, starting from the moment you were only a tiny seed of life in my belly when I engaged with COBOL for the first time back in the fall of 2011 together with my students at Srishti. Later, in 2018, you even became a participant in the extensive research travels yourself. So it should not come as a surprise that as you started to read, you claimed that the word GLOBAL must be a misspelling of
COBOL. Just as this research project affected you, it has also been affected by you, hence it has flourished through your changing inte-rests and hobbies from zombies to Scratch programming. Kristoffer, my love and life-companion: our conversations and laughter have gently carried me through this project. I do not dare to think of all the times you have been forced to listen to my analysis and theories of COBOL, or I have asked you for a comment or to double check my texts for typos or grammatical errors, or expected you to share my own excitement of the discovery of yet another term which ac-cording to the Endless Endtime algorithm added up to 666 (I guess it is not a coincident that Terence Hill was one of the first terms I discovered :-). Thank you Kristoffer for your infinite patience, your ever-shining brightness, and not the least your wonderfully skewed humor. Thanks for being.
INTRODUCTION
Finalizing a thesis on the frictions of global flows could hardly have taken place under more dramatic circumstances than during the spring of 2020 as the global world was hit by the coronavirus cri-sis. Symptoms of the new coronavirus, which induces the respirato-ry illness, COVID-19, first appeared in a province in Central China around New Year, but spread quickly to the rest of the world over the course of the following three months, resulting in worldwide lock-downs during March 2020 in an attempt to limit and prevent further spread of the virus. Countries and borders closed down, regions clo-sed down, cities and their workplaces and even homes cloclo-sed down. In other words, flows ceased to flow and everyday life as we had known it came to a standstill.
As part of the lockdown there was a global call for doctors and nur-ses, intensive care beds, hazmat suits, face masks, respirators, hand sanitizer, but also a rather unanticipated call for a certain type of engineer, namely those familiar with the programming language CO-BOL. On April 4, the state governor of New Jersey, Phil Murphy, held a press briefing during which he made an urgent call for what he mistakenly called Cobalt programmers. Murphy obviously did
not mean the silvery-blue metallic chemical element Co with the
ato-mic number 27, which is used in lithium-ion batteries. His call was for programmers of COBOL. It is not without historical irony that Murphy would not remember its name correctly, as COBOL as a programming language has been regarded as outmoded and dead for many years.
The call for COBOL programmers came as a surprise because
al-though we are all taught the importance of doctors and nurses, not many of us would even be acquainted with a COBOL programmer, let alone expect COBOL programmers to be key to the up-keeping of society in a state of emergency. To understand this better, we need to take a closer look at the history of COBOL and its implications
in the hidden human-computer-interaction, politics, and economics
The overall aim of this thesis is to investigate frictions in such flows through the notion of Crisis Computing, focusing on material mani-festations of execution, crisis, and maintenance. It is the result of more than eight years of engagement with COBOL, an encounter which has led me to reflect on the relations between code, execution, and maintenance on a global scale. Computational execution may appear to be a smooth, seamless process of automating commands, but as I argue in this thesis, such presumptions are based on a uni-versal vision that omits the underlying, blackboxed entanglement of execution with crisis and maintenance. COBOL is a case in point for this entanglement and thus this thesis is structured through a series of lessons, reflecting my own process of learning and reflecting on, and in, this supposedly obsolete language. These lessons are an ex-tension of my artistic research practice, the outcome of which is this doctoral thesis as well as several exhibitions, interventions, perfor-mative lectures, workshops, and an “infinite” publication.
COBOL
COBOL is an acronym for COmmon Business-Oriented Language, a high-level programming language developed by the US Department of Defense at the end of the 1950s. The language was designed to be easily accessible even for non-programmers, through its explicitly “people oriented” user interface design. Good intentions notwith-standing, it is no exaggeration to say that COBOL has, over the years, ended up becoming one of the most detested programming languages of all time. By way of example, the COBOL entry in The New Hacker’s Dictionary (1996) reads:
:COBOL: /koh´bol/, n. [COmmon Business-Oriented Language] (Synonymous with {evil}.) A weak, verbose, and flabby language used by {code grinder}s to do boring mindless things on {dino-saur} mainframes. Hackers believe that all COBOL programmers are {suit}s or {code grinder}s, and no self-respecting hacker will ever admit to having learned the language. Its very name is sel-dom uttered without ritual expressions of disgust or horror.
From the outset, COBOL was met with great skepticism and the in-dustry was hesitant to support the new language, which was viewed as clumsy and too verbose by expert engineering culture. Conse-quently, in 1975 the famous Dutch mathematician Edsger Dijkstra boldly stated, that “COBOL cripples the mind; its teaching should, therefore, be regarded as a criminal offense.” (Dijkstra 1982, p. 130) And, although COBOL has become one of the most widely applied programming languages, such discontent holds true even today. This became evident to me in the initial phase of my PhD studies through the following incident: On one of my daily train commutes, I happe-ned to sit next to a highly esteemed professor emeritus in the field of interaction design. As I had just been at work, the professor politely asked about my research and the topic of my thesis, and when I pas-sionately started to describe the case of COBOL and my analysis of it, he gave me a puzzled look. After a short bewildering moment, he said: “I have to confess something to you, COBOL was actually the first programming language I ever learned.” He went on to explain that he had been subjected to COBOL programming during his years of compulsory military service. After the detailed story, he suddenly added, “I have never told anyone that I know COBOL.”
Similar instances to the above have occurred during the course of my research. Another highly acclaimed professor, this time at a Univer-sity in South India, revealed a similar story to me: After pursuing a PhD in physics, he was employed by a private company which offe-red to pay him three times more than he could earn as a researcher but not to work with anything related to his doctoral degree in phy-sics. Actually, the company’s manager didn’t know what a PhD was, but since the professor had graduated from a well-known university, the manager reckoned that the professor was highly intelligent and thus a perfect fit for the job of learning to write COBOL applica-tions. For both the Scandinavian as well as the Indian professors, our conversations seemed to be the first time they ever talked about their hidden COBOL competences. After all, COBOL is not a skill you put on your CV. At least not if you are a well-respected professor of either interaction design or computer science in Sweden or India. In spite of this omission, or perhaps, exactly because of it, COBOL, to me, takes an uncharted, yet primary, position in the history of interaction design. Despite its history of neglect and dismissal, the
language may be the first example of a deliberately user-friendly de-sign interface. In fact, COBOL inhabits a classic dilemma of interac-tion design where on the one hand technologies are blackboxed in order to be as user-friendly as possible, and on the other hand, the blackboxing results in the user experiencing a lack of control and lack of understanding of the actual logic of the technology. Not to mention that COBOL was the first programming language specifical-ly tailored to expanding the domain of computation and making pro-gramming and automatic management integral to everyday life. The language actually presages both the vision of full automation and the drawbacks of the supposedly transparent and user-friendly design of digital interfaces today. There is increased focus on the need to understand the algorithms that are the motors of such interfaces and thus interaction design as a discipline also needs to turn its attention away from front-ends to such otherwise hidden digital materialities. As part of this endeavor, incorporating the legacy of programming languages such as COBOL might be a crucial critical revision in the field, as my analysis will show that it offers a prototypical case study in the early history of user-friendly design.
Furthermore, despite its odiousness, COBOL is still around. And re-gardless of the language being disregarded and broadly considered a legacy (old but still running) or even obsolete language (with no one actually using it to develop new applications), the amount of active lines of COBOL code continues to increase every year, simply due to the maintenance of its legacy systems (Micro Focus 2020). This has led some sources to claim, as late as in 2017, that seventy percent of all business transactions are still executed through COBOL applica-tions (Micro Focus 2017). Moreover, according to Reuters ninety-fi-ve percent of all ATM transactions rely on COBOL, forty-three per-cent of banking systems are built on COBOL and eighty perper-cent of all in-person-transactions are done using COBOL (Hartman 2017). A point in case is the recent surfacing of the importance of COBOL as part of the COVID-19 crisis. As a result of the standstill imposed through the national lockdown, the US unemployment rate escalated rapidly within a very short time span. People were being laid off due to the slowing down or complete halt of business operations. Hence, from a historic low unemployment rate of 3.8% in the US in
February 2020, the number rocketed to 14.4 % two months later in April, which is the highest number recorded since the Second World War (Kochhar 2020). This in turn affected the information systems administrating the unemployment benefits. In the US, more than half of all states have key systems like unemployment benefits written in COBOL (Sullivan and Marte 2020). These systems may be more than forty years old and the exploding amount of new unemploy-ment claims as part of the COVID-19 crisis became a struggle for the information architectures administrating them.
For example, the Department of Labor in Vermont would usually be handling around 3200 claims at a time. Their system was designed to handle a maximum capacity of nine thousand, but in four weeks, requests jumped to seventy thousand (Heintz 2020). In New Jersey, unemployment insurance claims in the second week of March 2020 numbered around nine thousand, but the following week there were around 116,000 claims, and in the last week of March the numbers had rocketed to 206,000 claims per week (Miller 2020). Such enor-mous increased pressure on the systems led to breakdowns, errors, and delays, leading to the state governor, Phil Murphy’s public anno-uncement asking for COBOL programmers.
As the above example shows, COBOL is indeed alive and kicking. Accordingly, there are still people adding COBOL skills to their CVs, but as the sudden COVID-19-related demand for COBOL pro-grammers demonstrated, those people are most likely not Western engineers. Given COBOL’s bad reputation and neglected status, the programming language has since long been omitted from computer science education curricula in the Global North. Subsequently, the COBOL workforce of today is mainly to be found in the Global South, which means COBOL’s survival not only challenges linear no-tions of development, but also points to asymmetric power structu-res and interdependencies of technological development and labor manifesting themselves globally. This thesis will, in the course of a close reading of COBOL, expose socio-economic as well as geo-po-litical relations between software, hardware, and global flows, and, not least, inherent social, economic, and cultural divides within those flows.
An engagement with COBOL allows for a glimpse into global infor-mation systems and architectures that are otherwise concealed and inaccessible. This undertaking in COBOL reveals a hidden side of human interaction within information architectures: a form of hu-man-machine-interaction, which paradoxically does not come about due to the type of design in which the human is supposed to be the subject of the design process. Instead, it is a hidden and blackboxed form of human-machine-interaction, taking place at the back-back-end of systems. There, humans carry out tedious labor processes that ensure the systems are continuously executing and performing in what appears to be frictionless flows. The workforce is, for instan-ce, programmers working with the maintenance of legacy systems like COBOL, and thereby ensuring the continuous flow of global businesses i.e. bank transactions, logistics, retail, ticket reservation, and so forth. One could argue that these workers are keeping up glo-balization (Henke 1999). Furthermore, this process could be descri-bed as a continuous re-enactment of technological stability (Houston 2017), or as termed in this thesis, “Crisis Computing.”
By calling attention to hidden and opaque aspects of labor that are sustaining our information architectures, my research relates to re-cent studies that, similarly, are unearthing the role of human labor as part of the up-keeping and sustaining of media flows (Roberts 2019; Gillespie 2018; Irani 2016, 2015, 2014; Irani and Silberman 2014, 2013). Through studies of, for instance, Commercial Content Mo-deration or the optimization of AI data-sets through crowd-sourced workers, these research projects have, in different ways, pinpointed the crucial importance of foregrounding the human labor processes taking over where algorithms fall short. Equally important work has been carried out by stressing the underlying exploited labor submit-ted by users within social media platforms (Terranova 2004; Fuchs 2014; Scholtz 2013; Zuboff 2019; Chun 2016).
While building on this wide body of research, a slightly different angle is taken in this thesis. While also being critical of it, my thesis relates to so-called “broken-world thinking” (Jackson 2014) as well as Paul Virilio’s notion of the “integral accident” (2007, pp. 10–11), by arguing that crisis is integral to the execution of all supposed-ly automatic computational processes. Automation, in order to not
break down and stop automating, needs to be continuously maintai-ned and sustaimaintai-ned as an ongoing and continuous process that I dub “Crisis Computing.” This is an intersectional analysis, because of the multi-level nature of the keeping-up of appearances. Subsequently, in order to understand and analyze COBOL, as it is situated within intersecting and multi-directional power structures, a focus on “mul-tiple layers” of oppression (Combahee River Collective [1978] 2014) through analysis of intersections of domains of power has been gre-atly important to this research (Crenshaw 1991; Collins 2000, 2016, 2019). Accordingly, rather than focusing on human labor in itself, the thesis turns to intersecting and multidirectional power structures reflected in the entanglement of underlying material conditions, tech-nological infrastructures, histories, socio-economic, geo-political, and cultural aspects, in which such back-back-end human labor is embedded. In the following, I examine questions of power in relation to the neglecting of the “keeping up” as they manifest through global information systems.
Execution, Crisis, Maintenance
In this thesis, COBOL becomes a case of what I call Crisis Compu-ting through a process of “unwrapping” that involves three inter-secting concepts: execution, crisis and maintenance. “Execution” is commonly defined as an act “of carrying into effect” (OED, n.d.), but it also brings associations of death and punishment: killing in accordance with the law. And it is regarded as central to power and command, as exemplified by the executive, the one who rules and manages. Code can manifest such power as a performative spe-ech act, through for instance Alexander Galloway’s statement that code is language doing what it says (Galloway 2004, pp. 165–166). However, to me code does not simply enact or execute power but is also continuously executed by its context, for example, its use and its socioeconomic, cultural, and geo-political relations. Therefore, by stepping away from etymological roots and allowing myself to re-read, perhaps even misread and explore the concept of applied execution, I find myself with a rather different understanding of code and execution. Despite visions such as Franz Kafka’s ([1914] 2011) depiction of the perfect execution machine in his short story In the
Penal Colony, the one who executes is seldom the one deciding a sentence or commanding. Rather, between the order and the actual execution lies a space of possibilities, of mis-communication, nego-tiation, disobedience, and/or desertion, in other words, a space of crisis, and a possible turning point.
My point is that execution is neither autonomous nor automatic. It is not a one-directional power, but is, on the contrary, constituted by its use/operation and maintenance. Hence, execution is constantly on the verge of breakdown due to its operationality, and needs to be sustained and maintained in order to continue to execute at all. Crisis, execution, and maintenance are in this way part of a tightly interwoven triadic formation, which I dub “Crisis Computing.” The concept of Crisis Computing thus deviates from Galloway and oth-ers’ understanding of code as a performative speech act, and expands on Wendy Chun’s (2008) turn to the machine’s performativity by in-cluding Nathan Ensmenger’s (2009) attention to the social relations of which the code is part of.
Thesis Format
Throughout the thesis, the concept of Crisis Computing is examined and discussed in the course of an iterative close reading, a diffractive “re-turn,” to use Karen Barad’s concept (2014), of the curious case of COBOL, which forms a series of LESSONS. Each lesson looks at one of the main components of Crisis Computing and interrogates this concept through an analysis of different aspects of COBOL.
This proceeds through a two-way implementation of the lesson for-mat fleshed out here: on the one hand, these LESSONS reflect my own subjection to COBOL, and document my research journey to India in order to learn the language. On the other hand, through the-se LESSONS, I am also able to subject you, the reader to this other-wise neglected “criminal offense” of teaching you about COBOL. In this way, I adapt American artist Adrian Piper’s “Funk Lessons” methodology of using an ambiguous and underrated cultural
pheno-mena as a starting point for discussions, reflections, and
confronta-tions of asymmetric power structures.1 Piper turned to funk music
and explored the lesson format as a means of discussing and con-fronting complex racial and cultural barriers. There is of course no way a lecture in a programming language tailored to writing payrolls could compete with a funk dance lesson. However, I hope that these COBOL lessons will still open up discussion and debate around neg-lected and dismissed narratives and power structures within global digital cultures and their messy and sometimes contradictory sites. The thesis takes the shape of an ambiguous, double-sided artifact: a book to be read from two sides. The book has two front covers (which also make up the back covers), which are almost identical: they share the same title, but the description differs slightly. If reading from one side, the Lesson is literally a lesson, meaning its one hundred and thirty-one pages consist of a series of slides and transcriptions of a performative lecture, unravelling the histories of COBOL.
If read from the other side, you are reading this introduction along with the rest of the traditional thesis text.
This design choice is a way for me to implement my artistic research practice within the more traditional thesis format, as being not only complementary, but also entangled. In other words, this format is an attempt to avoid a split between practice and theory and follows Hannula’s warning that artistic research should not mean an artist who transforms into a scientist analyzing her own work (2005). I see my research practice as a different means of diffraction through matter and a way to conceptualize the issues that the thesis opens up to and also to reach different audiences, opening up the questions discussed on different levels of association and expertise. This allows for different ways of entering the issues and different mindsets to enter through different materialities and contexts.
The thesis is wrapped in a dust jacket that is a minimally re-app-ropriated cover of my first COBOL textbook, handed to me by my
1 Adrian Piper’s Funk Lessons (1982–1984), was a participatory performance project where the artist, through lessons in funk dance and funk music exposed her audiences to funk music and its neglect and thereby opened up discussions on class and race (Piper 1996, pp. 195–216). The lessons started out as intimate dinner events, but developed into large scale events hosted at UC Berkeley (November 1983), the San Francisco Art Institute (February 1984), New Langton Arts (1975–2009), and San Francisco (March 1984).
teacher in Bangalore, and which is analyzed in the MAINTENANCE lesson. In order to read the “real” title of the thesis, whether the per-formative lecture-front-page or the thesis-front-page, you will have to literally un-wrap the textbook dust jacket.
Thesis Structure
The thesis is structured according to lessons in the three main con-cepts EXECUTION, CRISIS, and MAINTENANCE, which form the triadic constellation of CRISIS COMPUTING. Each lesson exami-nes one of the main components of crisis computing in relation to COBOL and its entanglement with the other concepts. The MAIN-TENANCE lesson occupies the most space, as it is the core area of analysis and research, having a bearing on the other lessons. A fourth concluding section transports my findings from the first three lessons to a discussion of automation in a contemporary setting.
The thesis opens with the APPROACHES section, which presents an overall introduction to the main methodological and theoretical app-roaches that have informed this research. Here, I discuss unwrapping more specifically as an analytical tool, not primarily for revealing, but for diffraction, where I consider the programming language COBOL from multiple, interconnected angles. Thereby, COBOL’s multiple entangled histories are themselves unwrapped and with them politics and discourses of power through different situations and at different times. I discuss how this unwrapping is informed by the apparatus of observation, which then also includes my own position within the research. This includes a consideration of the Lesson format of the dissertation. Here I review my own experience of subjecting myself to COBOL lessons during my research journey to India, and how be-ing in this learnbe-ing-situation was affectbe-ing the relationship between me as a Western PhD researcher/artist and the engineers I met. From there, I turn to questions of how material engagement on a granular level, as part of my artistic research practice, may be understood as a form of expanded maintenance where I alter and modify the structu-res and systems which I am examining. This is an alteration not in or-der to propose solutions, but which is trying to modify the structures by inserting moments and possibilities for self-critique and analytical
understanding of the issues at hand. From there, the thesis sets out on a critical examination of computing within COBOL through the Lessons: Execution, Crisis, and Maintenance.
The first lesson, EXECUTION, examines utopian visions of auto-matic management as they were manifested in the initial ideas and early implementations of COBOL. The main focus is a close reading of COBOL, and the visions behind it, as it was developed in the late 1950s and beginning of the 1960s. The section looks at code, but not in order to single out the code as an authority of power, but rather to investigate the double movement between execution consti-tuted through code, as well as code consticonsti-tuted through execution. My point is that execution is neither autonomous nor automatic. It is not a one-directional power, but on the contrary, is influenced by its context, its maintenance and its operation. Execution can in this way also be understood spatially, as noted by Wendy Chun, as a site which occupies “the gap between code and result, the space for and of execution. A space of determination and thus a space of political agency and action” (Chun 2015). Throughout this section I unfold this spatial notion of execution in contrast to the absolute approach to space as enacted through utopian visions of automatic management. In this context COBOL plays a significant role, since it was specifically tailored for widening the domain of computation and thus include automatic management in various mundane and everyday services and functions, in contrast to previous computa-tion practices which had focused more purely on number crunching. The section critically reflects on the understanding of execution and automatic management as performative activations of abstract sym-bols of code and how this relates to universal world perspectives. Through the use of critical spatial theory, this universalism under-pinning execution is ultimately revealed as a form of continuation of colonial power structures of oppression.
In the final part of the execution lesson, I turn away from the initial utopian visions of COBOL as an interface of frictionless, automatic management. Turning to material conditions such as the influence of hardware, as well as COBOL’s cultural context, I address how COBOL applications are not only executing smooth automation processes, but are themselves being executed. This is introducing the
questioning and critique of the dominant discourse of seamless auto-mation, which I will interrogate further throughout the other parts of the thesis.
From the critical look at transparent, user-friendly interfaces and frictionless flows attended to in the first lesson, the second lesson,
CRISIS, looks at the moment failure appears and smoothness and
flows are interrupted. Through a close reading of the Y2K bug as a prime example of technological crisis, the lesson argues that crisis is integral to execution. The Y2K bug was depicted as something to be put right once and for all, but as this lesson examines, the bug oc-curred as a result of another method of problem solving, that is, the most rational and economic execution possible. The supposed tech-nical error of the Y2K bug was initially itself a clever solution to a problem, which with time turned into a new problem. The Y2K bug exhibits how the bug, and with it crisis, is an integral feature of tech-nology’s situatedness in the world, and thus something that needs to be continuously maintained in order to execute at all. As argued by thinkers from Heidegger to Latour, technological systems only be-come visible to us when they fail and break down. Similarly, a crisis like the Y2K bug, not only exposed the entangled relations between execution, crisis, and maintenance, but furthermore, through its con-nection to the outsourcing boom in India, it also exposed the power structures integral to global information flows. Hence, the CRISIS lesson unwraps neo-colonial divides within global information flows by focusing on the discourses surrounding the Y2K bug as well as its material manifestations in software infrastructure and labor, and how it exposed neo-colonial divides within the global information flow.
The six sections in the third lesson, MAINTENANCE, examine the paradox of maintenance and linear technological development by turning to that which is most often an overlooked part of compute-rization and automatic management, namely the back-back-end ma-intenance of the systems themselves. In line with American perfor-mance artist Mierle Laderman Ukeles, who in her 1969 “Manifesto for Maintenance Art” asked, “after the revolution, who’s going to pick up the garbage on Monday morning?” I am questioning who is going to be “taking out the garbage” after the digital revolution. In
other words, I am concerned with the question of who will be main-taining the automatic systems when all maintenance has supposedly been abolished and replaced by automation. I will address the notion of maintenance in different ways: In the prologue and the two first sections from a more general perspective and in the last three sections more specifically through a tracing of topologies of maintenance in the context of software through a close reading of the programming language COBOL.
The lesson opens with a prologue presenting a general discussion of maintenance through a comparative investigation of human factors in technology. From there the first section moves on to critically re-evaluating the notion of maintenance as an object of study within different fields (from philosophy, sociology, ethnography, science, and technology studies to computer sciences). In the second section a post-colonial reading of maintenance is presented and here I examine studies dealing with the “surfacing” of maintenance work, while at-tempting to critically scrutinize intersecting power structures at play when we as researchers are turning to that which is hidden. Here I fo-cus on the politics of the concept of “broken-world thinking” when this concept is relocated from its initial context of a Western design research environment and instead applied to a third world context. Possible pitfalls of “broken-world thinking” are discussed and an alternative intersectional approach is introduced.
This part relates to the conditions of my own particular case study which involves studies of maintenance of global information archi-tectures carried out through outsourcing for Western companies by IT consultancy companies located in India. These studies are presen-ted as typologies of maintenance throughout the last three chapters. Section three, “Maintaining Legacies,” consists of an analysis of the material, cultural, and structural domain of the maintenance of le-gacy systems exemplified by COBOL applications that are still in use despite their invisibility and supposed obsolescence. The chapter looks into how they are still running and asks why they are still run-ning as well as under what conditions they are still runrun-ning. Further-more, I ask what it means for our preconceptions and discussions of the supposedly linear technological development. These examples are discussed in relation to the concept of linear technological
deve-lopment as integral to capitalist ideology, as well as in relation to me-dia and design theoretical approaches which diverts from teleological and evolutionary linear development.
The lack of programmers with COBOL skills today is a big issue. In section four, “Maintaining Flows,” I turn to the meta-spaces of ma-intenance, more specifically to sites where the maintainers themselves are being “produced.” In India, like the rest of the world, COBOL is not part of computer science curriculums at universities. Noneth-eless, the large Indian software consultancy corporations have their own training institutes, where newly hired employees are trained. In this context, I examine the corporate campus as a “global space,” which appears to be a virtual space but which, through my analysis, turns out to be a colonial space. As a further reflection on these spa-ces, I discuss and criticize Thomas Friedman’s concept of
“Globali-zation 3.0” (2005).2
In the final section of the chapter, “Maintaining ‘Development,’” I turn to the meta production of global flows as seen through the per-spective of the teaching of COBOL and I ask what it actually is that these COBOL engineers are being educated to become. Here I draw on Christian Fuchs’s analysis of the racist manner in which space is being re-organized through outsourcing, in order to exploit Indian software labor in a global context (Fuchs 2014, p. 205), as well as Doreen Massey’s analysis of globalization, as a form of “claustrop-hobic holism” characterized by a teleological developmental project (Massey 2005, p. 77). My own analysis, however, shows an inherent contradiction to this aspect, as the Indian engineers are maintaining business critical legacy systems of the supposedly developed Global North.
The fourth and final chapter of this thesis, CRISIS COMPUTING, takes the “lessons” learned from my engagement with COBOL in the previous parts of the thesis, and transports them into a broader
2 Even though this is a concept stemming from a neoliberal business journalist, and not a proper research framework, I turn to Friedman because his vivid accounts of visits to, for example, the In-fosys Bangalore headquarters have been central to the conception of a “flat world.” Friedman’s ideas have become integral to production of the imaginary of India’s tech boom and his books cannot be avoided. In Bangalore they are in every bookstore as well as being sold on street corners by local street vendors.
discussion of Crisis Computing in relation to current understandings of automation. The concluding chapter is an investigation into recent years’ buzzwords and terms such as the “Fourth Industrial Revolu-tion,” “Second Machine Age,” and “Industry 4.0.” It examines how the claims inherent in these terms—of new eras promoting automa-tion, boosted by neural network-based machine learning AI, as key to fundamental economical and societal changes—are grounded in utopian visions of automation and execution, which neglect crisis, maintenance, and human labor as integral to the execution/automa-tion processes themselves. From there, I turn to the un-wrapping of different sites of Crisis Computing where human labor is blackboxed in the name of automation. The first example examines the back-ends, which (similarly to the overall systems maintenance taking pla-ce at the back-back-ends) are blackboxing human labor in different ways. This ranges from back-ends literally being hidden spatially by displacement, either through outsourcing to places far away from the users or to the far end of high-tech campuses where the workers are employed. They can also be functionally hidden, by being black-boxed behind wrappers (APIs) which are turning human labor into scalable infrastructure (Irani 2014). In the second example, I turn to the front-ends and examine how labor is internalized as users allow themselves to be subjected to omnipresent tracking and surveillance. Finally, I examine how a term like “Fourth Industrial Revolution” is being hyped in developing countries, and how this furthermore re-sembles the way developing countries were marginalized in relation to the Y2K problem at the turn at the millennium, as examined ear-lier in the thesis. A call for designers to, through the concept of Crisis Computing, engage in a more situated practice with the awareness of the socio-political and spatial inequalities of users, maintainers, and developers alike, wraps up this final lesson.
APPROACHES
The main object of research for this thesis, COBOL, is a contempo-rary anachronism, which is both culturally, technically, and structu-rally difficult to access and encounter. Cultustructu-rally, COBOL is by and large regarded as obsolete and thus no longer existing. As technical infrastructure, COBOL applications are blackboxed behind smooth, up-to-date, front-end user interfaces. Moreover, structurally, pro-grams written in COBOL are shielded behind the corporate gates
of the powerful Fortune 200 companies,3 which are still dependent
on them, along with the correspondingly solid walls of the software consultancy corporations who are taking care of the displaced main-tenance of the programs in India and elsewhere.
Anthropologist Brian Larkin states that, “[b]ecause the basic object of infrastructure is so diverse and can be analysed in so many different ways, the choice of methodology is a theoretical question” (Larkin 2013, p. 338). Following Larkin’s line of thought, my engagement with COBOL has forced me to theorize and reflect on what it means, as an artist researcher, to unwrap and reveal that which is otherwise hidden and concealed. Throughout this chapter, I outline these in-tertwined theoretical as well as methodological reflections, beginning with a discussion of notions of power in relation to my methodology of unwrapping. Here, I first discuss Bruno Latour’s theorization of the concept of “blackboxing,” which involves exercises of mapping and tracing, and an encouragement to “follow the action,” as part of a reverse-blackboxing of how systems and structures are construc-ted. This at first appears to come close to “unwrapping.” However, I critique and extend Latour’s model, which I argue is too linear, first through expanding upon how power, with Foucualt, is always multi-directional and transversal. I propose intersectional theory as means of analyzing these complex, multidirectional domains of power and I discuss how the researcher’s own role and position is influencing the
3 “Fortune 200 Companies” are companies ranked in the top 200 on Fortune Magazine’s annual list of the 500 largest U.S corporations in regard to their yearly revenues (Fortune 2020).
research process itself. This leads me to Donna Haraway’s diffracti-ve model, rather than Latour’s reflectidiffracti-ve model of blackboxing, and how this has guided my research process.
With this chapter my own position as part of the research should be-come clearer to the reader, in particular my immersion in and subjec-ting to COBOL, through engagement with COBOL lessons in India. Subsequently, I discuss the learning situation as a model of ethno-graphic engagement where lessons may become means of infiltration and amusement, albeit also some embarrassment. With inspiration from Adrian Piper’s Funk Lessons, I discuss how the lesson-model, as part of my own artistic research practice, is a way to subject my audiences to the otherwise neglected and dismissed histories of CO-BOL. In relation to this discussion, I unravel my own take on artistic research practice, which I have articulated as a means of diffraction through matter: where the unfolding of the research through a diver-se range of materialities and formats (code, exhibitions, written the-sis, workshops, performative lectures, interventions etc.), opens up the research and allows for it to be accessed in different contexts and on different levels, and thereby reach disparate audiences. Further, I discuss how such practice may be regarded as a way of maintai-ning the systems and structures that are being studied. Not in order to restore them, but rather through modifications and interventions that challenge and reflect on and through them by offering means of misusing their domains of power.
Unwrapping Black Boxes
In “The Ethnography of Infrastructure” (1999), American sociolo-gist Susan Leigh Star makes “a call to study boring things” (Star 1999, p. 377). Star builds on Bruno Latour and Emilie Hermant’s 1998 study of the Paris sewer system, “Paris: Invisible City,” which foregrounded the importance of studying the otherwise neglected networks of sewers and power supplies in order to understand key aspects of power in relation to city planning (Star 1999, p. 379). Star stresses the need to use a similar approach in the study of informa-tion architectures. She notes:
Study an information system and neglect its standards, wires, and settings, and you miss equally essential aspects of aesthetics, justice, and change. Perhaps if we stopped thinking of computers as information highways and began to think of them more mo-destly as symbolic sewers, this realm would open up a bit. (Star 1999, p. 379)
Regardless of whether they are depicted as highways or sewers, the-re athe-re difficulties getting access to and studying back-ends of global information flows, which tend to be, as Star puts it, “buried in inac-cessible electronic code” (Star 1999, p. 379).
Although COBOL is said to be taking up key functions within cont-emporary global information architectures, it is hard to get a glimpse of it. Not only is the programming language buried away in inac-cessible code, as suggested by Star in the above quote, COBOL is in itself inaccessible code. COBOL was one of the first high-level programming languages, which meant that the source code could run on any machine as long as it had a COBOL compiler instal-led. Nevertheless, as the source code compiles into binary machine language, it turns into object code no longer readable by humans. Reverse engineering attempts can be made, but it is impossible to be
entirely sure.4 The source code is thus blackboxed as the program
becomes executable. Today, COBOL is a legacy programming lang-uage, which means that it is outmoded, but still performing critical functions. However, this contributes to the language becoming even more inaccessible: COBOL programs are often enclosed by wrapping the existing old program within a new program and thereby reducing its functionality to in- and out-puts, which are then called by other,
more modern programming languages5 (see the fourth section in the
MAINTENANCE chapter for a detailed discussion of the practice of wrapping). Wrapping can be thought of as a form of blackboxing the legacy COBOL code inside a more modern language.
The concept of “blackboxing” has been adapted by Bruno Latour from a cybernetic scheme of reduction in order to model feedback
4 Unless, of course, the source code is intentionally made accessible along with the object code as
in Free and Open Source Software.
processes. In cybernetics, the concept is used as a way to deal with incomprehensible complexity by blackboxing this in such a way, that it is possible to accept its in-and out-puts without having to pay at-tention to that which is happening inside the box (Latour 1987, pp. 2–3; Ashby 1957, p. 86). Latour uses the notion of “blackboxing” in two instances: Firstly, as a way to illustrate the social construction of science in itself, when it goes from being “science in the making” to “ready-made science” (Latour 1987, p. 4). Secondly, Latour uses the concept of “reversible blackboxing” as a way to trace the network of agents (objects and actors) making up a technological apparatus, in other words as a research technique for opening up the machine (Latour 1994, p. 37).
In his 1987 book Science in Action, Latour deploys his concept of blackboxing to describe how science develops from being “in the making,” a process depending on resources, participants, time, ti-ming economy, equipment and even luck, to be “ready-made” and thus settled as truth. In the book’s introduction Bruno Latour sets out to open “Pandora’s Black Box” in order to analyze how scienti-fic truth is being constructed through socio-economic processes (La-tour, 1987, p. 1). Such processes of social construction are revealed through Latour’s long-term ethnographic field work in laboratories. Latour exemplifies his analysis with a depiction of science as a Janus face consisting of an old and young researcher (Latour, 1987, p. 4). The older is a bearded, successful researcher, whose findings have been settled into truth. The younger, on the contrary, is not there yet, as he is still “in the making” and in order to succeed, his findings have to become as settled as the older researcher’s.
On a different occasion, Latour uses a broken overhead projector in order to exemplify the idea of what he calls “reversible black-boxing” (Latour 1994, pp. 36–37). Here Latour describes how the apparatus, upon breakdown, changes from being an enclosed object to a network of different agents—actors and objects—making up the notion of the overhead protector, for example: bulb, Fresnel lens, and the repairman entering the room in order to fix the broken machine. Thus, through the process of “reversible blackboxing,” in this case through a failure that didn’t allow the machine to function as
expec-ted, the user’s attention is directed from the system as an enclosed object to an awareness of the different parts that make up the system (Latour 1994, p. 36).
I am focusing on Latour’s theory of blackboxing here because it constitutes an example where the researcher sets out to uncover, trace, and reveal how what we perceive as given truths, or “enclo-sed objects,” are constructed. This is also important to my own method when attempting to reveal the hidden workings of COBOL. However, when considering the blackbox approach, it is important to be mindful of the limitations associated with a totalizing approach to revealing vis-à-vis a more contingent and relational one. I unfold this critique in the following section.
“Obviously, he had not made a note of when he had made a
no-te!”6—The Problem of Revealing
In order to exemplify “reversible blackboxing,” Latour’s choice of an overhead projector seems surprising. After all, the apparatus is not a convincing example of a black box. On the contrary, on a technolo-gical level, an overhead projector’s open surface and seemingly easily accessible inner components are hard to compare to that of the com-plexity found in, for instance, networked information architectures. Such simplification is, nevertheless, eventually what makes Latour’s example of “reversible blackboxing” work: after all it is fairly easy to “follow the action,” and thus trace and map the different elements making up the “system” of the overhead projector. But then again, even in this rather simple network of bulbs, lenses, and repairmen, a paradox of recursivity keeps knocking on the door: while mapping the bulb, it starts forking into new networks (for example that of heat, metal-thread, wolfram, and halogen molecules), which again fork into new networks and so forth and so forth. And this is just an overhead projector: imagine how it would be to try to “follow the action” of, say, a broken laptop. Is this ever-expanding network not a contradiction within any attempt to “follow the action” and to map out in order to reveal?
6 This phrase is a quote from Bruno Latour and Steven Woolgar’s Laboratory Life: The
A rather different approach is to be found in Michel Foucault’s es-say “The Subject and Power.” Foucault proposes a way of analyzing power relations through the “antagonism of strategies” as opposed to analyzing power from what he calls “the point of view of its in-ternal rationality” (Foucault 1982, p. 780). In order to understand power mechanisms, Foucault examines relations which are esca-ping or resisting power rather than the relations constituting power mechanisms themselves.
When Latour and Star turn to the hidden and discarded, it is for the purpose of revealing, or in Star’s words, “surfacing” otherwise lost aspects, such as, for example, “invisible work” within informa-tion systems (Star 1999, p. 385) or, as in Latour’s work, the network of actors and objects making up technical apparatuses. In contrast, Foucault works with the idea of a double movement in which the one constitutes the other, and through which we can understand the co-dependency of the visible and invisible, mutually constituting each other as possibilities for knowing. Whereas Latour and Star re-present all encompassing researchers, out to reveal and map how our realities are constructed, Foucault on the contrary points to interrela-tions, contexts, and multi-directional and transversal power structu-res. To understand something about society’s dark sides such as mental illness (i.e., historically speaking, insanity), Focuault ([1961] 2009) therefore importantly highlighted the need to investigate how society understood and defined sanity.
At first, COBOL as a programming language with a history of being detested, relegating it to illicitness, makes it akin to aberrant pheno-mena such as spam and viruses within information architectures. But then again, if COBOL is running seventy percent of the entire global worlds’ business transactions (Micro Focus 2017), or is behind nine-ty-five percent of all ATM swipes (Hartman 2017), it is not really an outsider or the other. With this contradiction in mind, COBOL can perhaps best be understood as similar to spam, as a marginal phe-nomenon that turns out to be absolutely central to communication,
since it makes up more than fifty percent of all emails.7 This makes
7 “Global spam volume as percentage of total e-mail traffic from January 2014 to September 2019, by month,” Statista, accessed Feb 15, 2020, https://www.statista.com/statistics/420391/spam-email-traffic-share/. See also Kristoffer Gansing’s discussion of spam as simultaneously marginal and central to online communication in “Spamculture: The Informational Politics of Functional Trash” in Christensen, Jansson and Christensen (2011).
COBOL a marginal phenomenon that can be studied in a way similar to how the studies of spam and viruses by Finn Brunton (2013), and Sampson and Parikka (2009), or Parikka (2007), which, in line with Foucault, examine the anomalous in order to understand hegemonic discourses in communication systems. From this perspective, I see COBOL as absolutely central to, not communication flows in this case, but the flows of global business.
COBOL’s ambiguous state of being simultaneously marginal and central further manifests multidirectional domains of power such as Foucault studied. As a later development, intersectional theory today provides means of critically interrogating such interlocking or multi-directional power structures by addressing how power is constituted through relationships rather than universal categories. The approach has its roots in black feminist activism in the US in the 1970s, whe-re practitioners and writers addwhe-ressed the needs for understanding complex power structures in relation to race, gender, sexuality, and
class and how they were mutually constituting each other.8 At the
beginning of the 1990s, Kimberlé Crenshaw (1991) applied the term “intersectionality” in her analysis of the marginalization taking pla-ce through interlocking means of oppression in favor of hegemonic
mono-categories of power.9 Crenshaw’s interrogations were situated
within legal studies. She emphasized the focus on three main doma-ins of power: the structural, the political, and the representational domain (Crenshaw 1991, p. 1244). A slightly different approach was later proposed by Patricia Hill Collins, who foregrounds the ways in which the intersecting oppressions are organized in, what she terms, the matrix of domination (Collins 2000, p. 18). Collins proposes an analysis of power across, not three, but four “domains of power”
8 The Combahee River Collective was a group of Black feminist lesbians engaging in activism, writings, and discussions in the 1970s. Without actually using the term “intersectional” their text “A Black Feminist Statement” advocated for a multi layered approach to oppression (1978, p. 275), and demonstrated how traditional forms of mono-categorization lead to marginalization. They were ex-periencing how their positions as black and lesbian made it difficult to identify with existing activist groups, i.e. the Black nationalist movement, attended stereotypical and sexist gender structures (pp. 276–277), whereas white feminist movements neglected the specific conditions for black women (p. 278). Accordingly, the Combahee River Collective acknowledged the needs for working transversally with not only one, but “a whole range of oppressions” (p. 276).
9 Crenshaw’s article contained a thorough analysis of how Black women are being marginalized and discriminated against, not only by the hegemonic powers, but also through social work, since the womens’ “intersecting” relations of oppression (they were Black and women) were neglected by the activist groups who were concerned with either women or Black people. In this way the women who were the most oppressed were actually neglected (Crenshaw 1991).
(Collins and Bilge 2016, p. 27). Throughout my research process, I have been strongly influenced by Collins’ take on intersectional theo-ry and used her four domains as a guiding principle in my analysis of how COBOL is situated within a multidirectional matrix of different domains of power. These are the interpersonal, the disciplinary, the structural, and the cultural domain (Collins and Bilge 2016, p. 27). Collins encourages a wide application of intersectional theory in the context of critical social theory, rather than restricting it to research only related to identity politics (2019). A case in point is Collins and Sirma Bilge’s interrogation of the social unrest related to the 2014 FIFA World Cup tournament in Brazil (2016, pp.5–13). Through their analysis of how power played out within the four different do-mains, the authors reveal how a mass media spectacle like global football serves political ends.
An examination of how power structures are “intertwined and mu-tually constructing” and thus amplify each other (Collins 2016, p. 7), may also help to avoid jumping to quick conclusions, as when Western scholars exoticize the maintenance and repair work that ta-kes place in developing countries. This is part of my discussion of so-called “broken world thinking” (see the Maintenance chapter). Here, intersectionality can, importantly, highlight the researcher’s own position within the research, which I will examine further in the following part.
The Problem of Authority
In Bruno Latour’s example of Pandora’s Box, it is implied that the two different sides of the Janus face, the old and the young resear-cher, are one and the same: If the “young” scientist has to manifest himself and succeed, he has to become “blackboxed” like the older researcher. But is this the only way to imagine research? That suc-ceeding in science equals blackboxing? Is there no way for science to embrace “solid” fluidity rather than just solidifying (truth versus non-truth)? Somehow, the statement from Latour seems to repro-duce the same authority as the authority which he is criticizing. It is the all-encompassing uncovering without paying attention to the researcher’s own position in the uncovering process. As the problem
