Skip to content


This thesis has discussed various definitions and interpretations of the notion of scripting throughout its four chapters, showing a progression of ideas across a series of design projects. The presentation of projects in this order displays both a thematic, and roughly chronological progression in my design research work over the past few years.

Chapter one discussed behavioural scripting within two already-initiated projects, the Godot Machine and Ant Ballet. Using Bateson’s concept of psychological framing (which shares a number of concepts with both Schank and Abelson, and Tomkins’ script theories), the chapter discussed the framing of both of these projects in relation to the field of ‘sci-art’. It concluded that my practice is best defined as work that is informed, but not defined, by an engagement with other disciplines, including science.

Chapter two introduced two projects, Nybble and Scriptych, which employed performative diagrams of computational processes, and were shown in major cultural institutions. Both projects used absurd diagrams as means of understanding, rather than purely explicating, concepts related to computation. Scripts were exploited in each piece as a mode of giving instruction to performers, whilst Scriptych made heavy use of computational scripting in both its creation and performance, which lead to the development of a novel method for interacting with a detailed three-dimensional vector database.

Chapter three focused on computational scripting, using a pair of projects (86400 and 24fps Psycho) as a mode of critiquing specific contemporary technologies. The computationally-generated film 86400 employed a script to interact with the Google Image Search API, creating a database of hundreds of thousands of time-related images, which were compiled into a 24-hour-long film: the work was a diagram seeking to create an immersive critique of the endless onslaught of classification that is pervasive in digital culture, and in turn to highlight its meaninglessness. 24fps Psycho used computational scripting, and a diagrammatic computer interface (Max) to create an ultimately unsuccessful film that was performed live in front of an audience. The chapter discussed the various reasons for this projects’ failures, and speculated that such a project will likely appear computationally naïve as future technologies develop which render the machine learning tasks attempted easy.

Finally, chapter four introduced the creation of the film Network / Intersect via a methodology entitled reflexive scripted design, an original process in which the designer takes full authorial control over a project, designing a set of constraining rules, then producing work by following those rules. Although the process has a major impact on the formal outcome of the project, it is not currently a form-finding exercise in the sense that many parametric design techniques are. Building on the techniques of Italo Calvino and the Oulipo movement, the designer places extreme constraints of their own making onto a project, whilst also rooting it in time and history. Reflexive scripted design has been tested once, on the creation of Network/Intersect; its robustness and use beyond this exercise require further study.

The main argument of the thesis is that the notion of script is one that is both important, and underexplored, in architectural design discourse. The word script is tied to early notions of writing as well as drawing, and is in regular use within theatrical and filmic productions, psychology and computer programming. Within architectural design, however, script is most commonly used in relation to computer scripts, and, by extension, parametric design. This thesis has argued that this narrow understanding of script belies a rich vein of concepts that exhibit potential for new design methodologies, such as reflexive scripted design, which I outlined and tested in chapter 4. This wider understanding of script allows for an architectural discourse that includes conversations about authorship, freedom, and creativity within design.

I have also presented the Camusian ‘absurdb’ as a mode of critique, arguing that the themes contained within the plays of the Theatre of the Absurd – originally written in Europe at the beginning of the Cold War – are in fact relevant to present-day uses of technology. The method used to implement this critique has been through the construction of projects which function as absurd diagrams, or absurd machines. The synthesis of these central themes – scripting, the absurdb, and diagrammatic practice – and the mode in which they have been explored (through active design research) represents a novel enquiry.

The thesis has presented a diverse set of design projects, using different media and methodologies. However, these projects have several points of convergence. Firstly, there are clear thematic links between the projects: each work has displayed a clear trajectory in its development of the notions of script, the absurd, diagram, and computational technology, with each project seeking to pose specific questions which enable an audience to create their own absurdb realisation. Secondly, all the projects employ a kind of ‘script’ as a mode of expressing design intent, from the use of computer scripts in 86400, 24fps Psycho and Scriptych; to performance scripts and instructions in Nybble, Scriptych and Network/Intersect; and the use of a reflexive scripted design process in Network / Intersect. These works developed from my preliminary concerns with the scripting animal behaviour in the Godot Machine and Ant Ballet. The process of translating a script into an artefact, performance, film, or whatever else it may become, is for me the most interesting part of my design practice. It is where the nuances of the mode of scripting are most accentuated, and where the freedoms or restrictions inherent within each kind of script are most concisely articulated. It has been a great privilege to have such a high degree of authorial control over the translation of these projects from script to ‘thing’; many of the characters discussed in the introduction of the thesis (including Ptolemy, Alberti, and even Beckett) now have no control over the way their scripts are translated and implemented.1 This section of the thesis examines some of the peculiarities and differences between the various modes of scripting, and the translation of these scripts into artefacts, films, and performances.

Reflections on Practice#

It is worth noting that none of these projects was developed in isolation from each other: most of the work was carried out in the same studio spaces, using the same computer, with often overlapping production timescales.2 Yet each mode of scripting carried its own set of parameters, requiring different modes of engagement with subject matter, spaces within studios, and even other people. Practicing the various modes of scripting simultaneously often required jumping from one state of mind to another, in order to pursue contrasting logics and practices, for example, the shift from computational scripting to writing an actual film script. Sometimes I would need to find means to transfer from one mode of scripting to another; often this was in the form of a walk, a meal, or exercise. Some environments were far better suited to different modes of scripting: the isolated studio apartment where I lived in Paris was far better for writing computer code than the sociable studio in the Palais de Tokyo, yet the latter was far better for visual inspiration, creative thinking, and sharing ideas. Understanding the requirements of each mode of scripting informed the way in which work was produced. This is of particular importance today, as computational scripting becomes both and more popular, and more accessible, in architectural design.

The computational aspects of the works mentioned in chapters 2 and 3 required the adoption of a completely different creative mindset to other parts of the projects. In the second chapter, I have outlined the way in which I attempted to think through the logic of computational code and create instructions, which would effectively generate the desired performance from a computer. Much of this code-writing made use of external libraries (as described by Ousterhout), which meant that lower-level functions could be achieved with rudimentary knowledge of the underlying principles.3 The machine-learning scripts used to train the word-vectors in Scriptych, for example, were based on a library called Gensim by Radim Řehůřek, which enabled the parsing of large corpora of words to create word-vectors for every word which appeared in a collection of Hollywood film scripts, themselves having been downloaded via automated scripts and processed by Cristian Danescu-Niculescu-Mizil at Cornell University.4 Gensim is open-source, and is maintained by Řehůřek and myriad other volunteers; it was programmed in Python, itself an open-source programming framework with thousands of contributors, and it makes extensive use of Tomas Mikolov et al’s Word2Vec process.5 The interface that enabled dancers’ moves to retrieve these words was built in Max, a visual scripting language, but also made extensive use of libraries which enabled it to communicate with the sensors via Open Sound Control, as well as an instance of Gensim running in Python, and the ‘say’ functions within Mac OS X’s built-in command line interface. I could continue to cite various libraries, snippets of code, dependencies, interfaces, and languages for each project (including several more for Scriptych), but this thesis is about design research, not computer science. In short, using computer scripting languages enables the integration of myriad external libraries which extend the core functionality of a language, making repetitive or difficult tasks easier, or simply speeding up the production process (as described by Ousterhout).6

Although many of the libraries for languages such as Python are open-source, and freely distributed, there is a cost to the ease with which they can be used. This is borne in both ignorance over functionality, whereby the end user does not have to know what the computer is really doing when a certain instance is called, and the risk of creating software that is overly dependent on external libraries. Many libraries are in fact reliant on other, lower-level libraries for their core functionality; a library such as pyMovie is actually a fancy wrapper for ffmpeg, for example. Simultaneously, the high level of dependency that operates within these languages can act as an Achilles' heel: the deprecation of Google's Image Search API during the execution of 86400's image-gathering procedure highlighted how the lack of support for one particular library, API, or even language can have a major impact on the reliability of an artwork. Even the software which ran Scriptych, developed only last year, requires a very specific software and hardware setup, and is at risk of not working simply through routine software updates on the smartphones or laptop it runs on (the computer used to produce the work, and this thesis, has auto-updated its operating system tens of times since then). This is partly due to the low-budget nature of such software development: experimental projects such as this one are a form of electronic bricolage, relying on off-the-shelf hardware and software patched together via untested means. Whilst other artefacts – films, buildings, costumes, books, etc – are all at risk of various forms of degradation, none quite match the fragility of this type of computer-based work of art.7 Attempts to rectify these vulnerabilities and archive software tend to focus on the concept of a virtual machine, a software emulation of specific hardware and operating systems, a necessity which generally prohibits physical archiving and interaction.8 The remedy to such issues might be a step towards more robust, lower-level programming on my behalf– but this takes considerable time and effort to learn, and, as I will explain shortly, such practice comes at a cost.

Writing computer code, in my experience, requires a different mode of engagement to other modes of design practice. Whereas physical design, with pencil and paper, can be completed in a loud studio (perhaps with music playing), in a conversational manner, the writing of code of any complexity often requires its author to engage in intense concentration as they iteratively build layers of complexity into an extremely abstract system. In my experience, this is also intensely antisocial, and my physical studio requirements differ if I am building a model for a film, or writing a programme, which interfaces with an online service. Often I have found myself the only person in any given place working on a code-based project (as was the case in Opera Garnier with Scriptych, or whilst working in Nanji in Seoul on 24fps Psycho). This creates a different atmosphere to the traditional design studio, a place which relies on a high level of interaction between its inhabitants in order to enable the cross-pollination of ideas. Where I have found traditional design logic is best explored through testing it with other people – both as design ideas are formed and often fleshed out in a convivial manner, and asking the opinion of colleagues (or anyone else) – often my code is written in isolation, and the support communities exist purely online, in the form of websites, wikis, and forums.

Much of my non-code design practice contains references to other work (for example, the repeated references to Stanley Kubrick’s films in the Godot Machine and Ant Ballet, or to Beckett in Nybble), or a moment of humour (for example, the posters reminding the viewer they are ‘in Korea’ in Network / Intersect). I try to maintain this ethos whilst writing code, embedding humour or references into the very functionality of the project.9 I am aware that in most cases this will never be seen: the work I produce is generally too embedded in artistic practice to be of interest to many programmers, and that sort of detail is too technical to be of interest to most of my peers in the design, art, or architecture worlds.10

Computer programmers and scripters alike often refer to a state of working they call ‘flow’ – a total immersion in work whereby everything outside of the task in hand becomes unimportant and abstract.11 I have experienced this state whilst working on numerous design projects, but never to the same extent as whilst working with computer scripting. There is something about working on extremely technical issues, which requires a thought process quite different from normal, everyday logic. This causes me to lose myself in my work, which I find immensely enjoyable.12 Similarly, the highly technical nature of computer scripting has a draw for a particular type of person within the field of architecture. As Peg Rawes notes, although there is a near-even ratio of female to male students within US/UK architecture schools, code-based architectural design (such as parametricism) is: ‘more commonly populated by male students who are instructed through intensive workshops which focus on repetition and emulation.’13 Whilst this is problematic within the world of architecture, it is perhaps manifest most explicitly in the remarkably male-dominated workplaces of Silicon Valley. As recent news reports have shown, there is increasing discontent with the aggressive corporate culture perpetuated within Silicon Valley technology companies (in particular Uber).14 Chapter 3 of this thesis discussed one instance where the idealised, quasi-mathematical model of continual movement has been translated into a daily reality for a swathe of Uber’s drivers, who are now the test-subjects in what the company is calling the Perpetual Trip.15 Much of the aggressive nature of the technology industry is arguably perpetuated by the mythology surrounding the Valley’s start-up scene: the image of young (usually male) entrepreneurs frantically building circuit boards, writing code, and using venture capital to create the “next big thing” in a garage is synonymous with Silicon Valley’s corporate technological culture.16

The notion of a mythological origin story is not only found in Silicon Valley technologists. As was mentioned in the introduction, Rawes has argued that the ‘hot house’ methods in which parametricism is taught within universities, small teams, and ‘labs’ has generated a discourse around the style that relinquishes designer agency and equates computation-based design with ‘natural’ processes.17 Rawes claims that latter is reinforced by mythologised connections between code and highly selective slices of genetic theory (such as D’Arcy Wentworth Thompson’s On Growth and Form) drawing parallels between parametricism and organic, nonhuman processes.18 As a result, this ‘uncritical belief in computational emergence’ Rawes argues, ‘reduces diverse organic life to a self-same matter’.19 This argument has parallels with Korzbyski’s map-territory relation (as invoked by Bateson), discussed in chapter 1:20 claiming that an organic system can be explained through a mathematical model (as is the case with many of the morphological arguments in On Growth and Form) is akin to producing a map of a territory; it necessarily requires a process of reduction and exclusion. Using that same map, or parametric model, to then plan territory renders the territory itself subject to the same process of reduction and exclusion. In the case of parametricism, human behaviour and desires are often reduced to a simple agent-based model, which reduces human desire to the status of movement through space.21 Rawes proposes a renewed interest in Spinoza’s notion of geometric and ecological ratios as a means of ensuring human wellbeing and environmental responsibility; I will shortly discuss how I believe Tomkins’ notion of psychological scripting may also offer a means of a more humane mode of architectural practice. Computational scripting does not have to exist in isolation, or as a non-critical discipline within architectural design. The projects presented in this thesis show the ability to use computational scripting to pose critical questions about contemporary technology, as also demonstrated by artists such as James Bridle, Kasia Molga, Heath Bunting, and institutions such as Het Nieuwe Instituut in the Netherlands.

Performative Scripting#

In contrast to the technical nature of computer scripting, the performative scripts I have written in the course of this thesis were created under a wide variety of circumstances. The production of Nybble was the first time I had coordinated and directed dancers; while the making of Network/Intersect was the first time I had worked with actors and film crew who required direction (in the making of the earlier Ant Ballet film, I had been simultaneously the actor and director, technician and experimenter). The process of developing instructions to be given to dancers or actors verbally along with the written scripts themselves was an entirely new concept to me when I began these projects. As the designer of performances, I find it easy to focus on the spectacle of the overall piece, or the technical details, and forget the emotional experience of the performers who will be enacting the work. With Nybble, I was fortunate to be able to bring in my frequent collaborator Abi Palmer to coach and instruct dancers in their movements. During the conception of the piece, I had failed to think about what the dancers would be thinking as they were dancing, and how their thoughts and mental states are translated into the movements the audience sees. Abi developed a language of instruction as a supplement to the technical information I had already provided the dancers with. Her commands asked the performers to imagine they were in some other state, such as being underwater, on fire, made of various substances, to move slow, fast, etc, in order to animate their gestures. This generated an ensemble, which was imminently more engaging to the audience as well as the dancers, who were no longer simply performing rote movements to get into a particular position.

I was later informed by dancers at the Paris Opera that this approach is a common technique adopted by performers to help them memorise routines. Whilst working with choreographer Simon Valastro, and developing a highly technical interface for the dancers to work with, I saw these mnemonic techniques in practice – even for the highly precise gestures required by the software I had created. Movements gained nicknames such as ‘Virgin Mary’ (a praying position), ‘handshake’, ‘mother and child’ or ‘reset’, to render them memorable in sequence, and to imbue them with emotion which the dancer could then express through movement. In the case of Nybble and Scriptych, the dancers’ approach to learning and expression was thus not entirely cohesive with the mode of interpretation required for the software or code; for both pieces I had hoped the dancers would treat the choreography like instructions in a manual, but human scripting proved to be far less binary than its computational equivalent.

The live aspect of these performances further highlighted the difficulty in synthesising human and computational agents. The performances of Scriptych, for example, felt very much like a technical presentation, and tension was heightened by the aforementioned vulnerability of software and hardware interfaces, made quickly and using experimental and even bricolage-style means. The first time the female dancer moved, and the system was tested in front of a live audience, was a tense moment for Simon and I, who were watching the performance from above in a control booth-like position. Both of us had highly technical jobs to do at the time, using the custom interfaces of smartphones and tablets to cue certain parts of the performance (I was monitoring the input from the dancers’ phones whilst Simon was responsible for changing the section of the music, which also controlled the way in which the programme ‘listened’ for movements). This is a position I have often myself during performances of this kind, including 24fps Psycho, Scriptych, Ant Ballet, and the numerous technical demonstrations I have given throughout and beyond this doctoral research. I have heard numerous stories from colleagues who work with similar technology, as well as directly experiencing, the horror of presenting a technical mechanism or software interface or similar to a room full of people, only to have it fail to do what it was meant to do (and in most cases, had been doing only moments before). The tension this feeling invokes was amplified during the performance of Scriptych, as the failure of the software would not solely affect me, but also leave the dancers performing for nine minutes in front of an audience, unguided. This vulnerability also placed strain on the reputation of the choreographer, the curators, and the institutions we were collectively representing. It was, in short, the biggest risk I had every undertaken with a project, but also now feels like one of the most rewarding. Having scrutinised the video recordings of the two performances and the practice sessions, the tension the dancers were carrying in the performance is palpable. Yet I believe this human element actually improves the film, adding tension where the computational script cannot.

The translation of emotions to performance was also present in the making of Network / Intersect. As I had not worked with actors before, I encountered several unanticipated issues in the translation of the written and diagrammed script to actual performance. One shot in particular, where the executive character silently descends a marble staircase, took around fifteen takes. This was partly due to my inexperienced direction: I had a particular style of movement in mind for the actor to use in the descent, which seemed of utmost importance when filming, but found the viscerality of human movement hard to verbalise. Ultimately, the shot used in the film was the product of a series of negotiations and minor tests. The way in which both of the actors embodied their characters was interesting to see; one of the actors would continually play in their role, testing new movements, facial expressions and characteristics, whilst the other preferred to have a more grounded, rationalised explanation of exactly how their momentary performance would fit into the characters’ trajectory. Similarly, it was a revelation to find that the small crew I was working with often only needed to know what their immediate task was, rather than an in-depth knowledge of how a shot fitted into the wider context of the film. This aided the experimental style of the films’ shooting, in which the cast and crew were tasked with small, repetitive, and iterative tasks, much like the Russian propaganda workers they were unwittingly mimicking.

The physical locations in which I wrote scripts varied immensely across the performance pieces. The logic behind Nybble (and the coding of the Microsoft Excel sheet which was ultimately used to sequence the performance) was aided by a series of conversations with collaborators, as well as colleagues. This is a mode of design I am familiar with: conversation and free-floating ideas from myriad sources gradually find their way towards a convergent and cohesive narrative, which can be expressed in some sort of design medium. A similar process was true of the writing of Scriptych: I wrote the palindromic dialogue between the dancers overnight after an intense period of ideas-generation with Simon Valastro close to our first collaborative meeting. With minor changes, this dialogue remained the central part of the performance; most other elements, including most of the computer programming logic, and even the variety of movements available for Simon to create choreography with, came through systematic and iterative developments made after thorough testing of the system. Of all the modes of computer scripting explored in this thesis, this particular one, whereby Simon and I would meet daily for several hours and identify issues, which I would then work on alone through the afternoon and evening to solve, was by far the most satisfying. Consistent and near-immediate feedback on a design proposal, particularly one that interacts with someone whose working methods I find challenging to empathise with (I am by no means a gifted dancer) was highly enjoyable. The process felt like a conversation whereby we each came to understand the others’ perspectives, mediated by a common interest in the development of a performance. Numerous aspects of the performance and interface were a direct result of this closely integrated design-testing process. This is precisely the type of empathetic thought process, whereby one has to shift one’s perspective into the place of another human (rather than machine), which is almost impossible working exclusively in computer script (as opposed to any other design medium). I will be looking to pursue projects where technology requires collaborative and empathetic design and feedback cycles in the future.

The scripting process for Network / Intersect contrasts the scripting experiences of projects such as 86400 and 24fps Psycho. Where both of the more technical projects required dedicated days of computer scripting, largely in isolation from external forces, Network / Intersect was largely conceived and written in the most clichéd way possible: by wandering the streets of Paris with a notebook in hand. The reflexive scripted process was the result of countless late-night walks through the city, as well as conversations with colleagues, friends and strangers alike. The experience of writing a work of fiction – the first I have attempted – was immensely pleasurable, although the synthesis of the myriad ideas which fed into the project took a long time. The process of creating rules in order to construct the script was part in due to the frustration at connecting the concepts I wanted to integrate into the project itself; I forced myself to work within extreme constraints to reduce the number of possibilities that existed. The disconnection between the dialogue and on-screen action was a deliberate choice that was made upon realising that the film required around 65 shots to be filmed in a 3-week span; synchronising dialogue and cutting film to match the project’s rigid time structure would be highly challenging; the film would be far easier to cut if the dialogue could be treated separately. This inadvertently aided the film itself: a dialogue over the top of the film caused the pair of films to appear to be a reflection of the internal states of the characters themselves, channelling the Batesonian notion that ‘human beings operate more easily in a universe where some of their psychological characteristics are externalised.’22

Future research#

This study has examined several key areas related to script, the absurd, diagrams and performance from the perspective of a designer and artist working in performative architectural design and film-making. This is, by virtue of the research methodology, a limiting practice. No one person can reasonably expect to tackle all issues that they might wish to. Whilst this thesis has provided accounts of my own research-based methodology, there are several areas, which I have not yet been able to explore which I believe may yield fruitful research in the future.

The first is the relationship between psychological scripting, affect, and architecture, with the theories of Sillvan Tomkins of particular interest. To what extent can architecture influence the way in which Tomkinsian scripts form, develop and are enacted; and what could this tell us about how to improve design for human inhabitation? Whilst behavioural scripting is a term already in use within parametricist discourse, it generally refers to programming highly reductive ‘agents’ in a virtual environment to simulate their movement (rather than emotions), most commonly whilst in a crowd. A study into Tomkinsian psychological scripting and affect may yield interesting methods for designing better, more humane environments, which are enjoyable to inhabit.23 One possible case study (which I mentioned briefly in the footnotes to the introduction chapter) could be Hogeweyk in the Netherlands, a care community for dementia sufferers (designed by Dementia Village Architects: Frank van Dillen and Michael Bol).24 The community is designed to have the appearance of a generic town, with amenities such as shop, theatre, pub, and public square, with residents carrying out apparently normal lives within the town. It is, in effect, a safe place, offering residents a good deal more autonomy than residents than a traditional care home. One of the significant problems that dementia suffers commonly face is the confrontation with the realisation that they are being looked after, and mistrust that occurs when, for example, staff who they do not recognise have to coerce them into taking medication. Hogeweyk operates to a different principle: staff do not wear uniforms, but rather clothes that imply they are just ordinary citizens in the town; the residents are able to go about their lives in a manner whereby they can continue what might be described as scripted behaviour, buying groceries at the shop, visiting the pub or restaurant, all the while being cared for by the staff (who find themselves playing generic ‘roles’). The carers do not contradict the reality the residents project onto their situation unless explicitly asked; thus, the residents are able to be cared for in a way that minimises conflict, and associated stress and mistrust.25

The second area for future research is the viability of the reflexive scripted design process, both for its current intended media of fictional film and theatre, as well as for other modes of design. So far, I have made one film using this practice, but its robust viability is unproved in any medium beyond this. How might reflextive scripted design be applied to other modes of practice, such as design of the built environment or graphics? Is it only suitable for work that deals with fiction? These, and more questions, are as yet unanswered.

The final area of research that I believe to be interesting is precisely the one I intend to work on next: better defining the relationship between the psychological script, human agency, and various models of freedom. As explained in this thesis’ introduction, human agency, and the extent to which behaviour can be codified and predicted, is one of the central questions of large-scale surveillance systems – themselves often operated by ‘agencies’ of the state. These immense prediction systems, also the product of myriad machine learning computer scripts, rely on inferences about human behaviour based on the ‘logic of resemblance.’26 The systems are constructed on the assumption that with enough data, patterns of threatening or subversive behaviour will reveal themselves through a similarity to previous instances of such behaviour – i.e. that such behaviour is, itself, somehow subject to a script. The relationship between this mode of behavioural analysis with the intention of prediction, as well as the often opaque mechanisms through which it often operates, and the concepts of psychological scripting advocated by Tomkins / Schank and Abelson, is largely unexplored. What’s more, the evolving nature of the predictive systems, and the way in which they codify and predict human behaviour contain potential implications for the three types of freedom – personal, sovereign, and civic freedom – identified by historical sociologist Orlando Patterson.27 This is a subject area which I have hinted at but not explicitly explored, throughout this thesis, one in which I believe the topic of the absurdb may one again prove to be a useful ally.

See Gene D. Phillips and Rodney Hill, The Encyclopedia of Stanley Kubrick, Great Filmmakers (New York, NY: Facts on File, 2002), ix; Emily Blunt, ‘Bluntly Speaking: Interview with Charlie Kaufman’, Blunt Review, accessed 2 April 2017,

Virtual machines are also now a core concept to services such as Docker, Amazon Elastic Web Compute, and Google Compute Engine, all of which now offer ‘cloud-based’ computing services which distribute computation to large-scale distributed computer servers. An interesting history of this form of computing can be found in Tung Hui-Hu’s book A Prehistory of the Cloud. Tung-Hui Hu, A Prehistory of the Cloud, First (Cambridge, MA: MIT Press, 2015).

The notion of fluidity in itself is similar to those expressed by Schumacher in his parametricist manifesto Parametricism: A New Global Style for Architecture and Urban Design. Patrik Schumacher, ‘Parametricism: A New Global Style for Architecture and Urban Design’, in The Digital Turn in Architecture 1992-2012, ed. Mario Carpo, AD Reader. (Chichester: Wiley, 2013), 243.

Today, ‘incubator’ companies such as Y Combinator aim to emulate the conditions that have led to the creation of large-scale technology companies through pairing venture capitalists with hot-house style provision of space and resources. Facebook’s first start-up incubator project in Paris, revealed early this year, extends this myth futher in its name; called the ‘Startup Garage’, it will house 80 desks for a ‘6-month program [^f]or data-driven start-ups’ with direct access to Facebook engineers and resources. Jennifer 8 Lee, ‘Running a Hatchery for Replicant Hackers’, The New York Times, 21 February 2006, Online edition,; Nathaniel Rich, ‘Y Combinator, Silicon Valley’s Start-Up Machine’, The New York Times, 2 May 2013, Online edition,; Polina Marinova, ‘Facebook Is Launching Its First Official Startup Incubator’, Fortune, 17 January 2017, sec. Venture,; Romain Dillet, ‘Facebook to Open Startup Garage at Station F in Paris’, TechCrunch, 17 January 2017,; Facebook Inc, ‘Startup Garage at Station F - Posts’, 6 January 2017,

Douglas Spencer also discusses the influence of On Growth and Form on the work of influential parametric architect Greg Lynn, who uses Thompson’s concept of ‘flexible type’ organisms whose form is purely reactive to its environment as a model for architecture which is equally formed by its surroundings, and as part of Alejandro Zaera-Polo’s arguments about neoliberal market ideals. Douglas Spencer, The Architecture of Neoliberalism: How Contemporary Architecture Became an Instrument of Control and Compliance (Bloomsbury Publishing, 2016), 56,126.

This argument is exemplified in the Jorge Luis Borges short story On Exactitude in Science, in which an empire becomes so obsessed with cartography that it builds a 1:1 map of its territory; the map is useless, and ends up a tattered historical relic. Jorge Luis Borges, ‘On the Exactitude of Science’, in Collected Fictions, trans. Andrew Hurley (New York: Penguin, 1998), 325.

  1. Two examples of people mentioned in this thesis who have been unhappy with the interpretation of their scripts: Stanley Kubrick’s adaptation of Stephen King’s novel The Shining left the author so dissatisfied that he made his own version for television in 1997 (seventeen years after the release of Kubrick’s film), and Charlie Kaufman, who has discussed his dissatisfaction with George Clooney’s direction of his screenplay Confessions of a Dangerous Mind in numerous interviews. Kaufman is notable because he tends to have a close relationship with the directors who translate his writing into films. 

  2. For example, much of the Max and Python computer scripting for 24fps Psycho was carried out in the same studio in Seoul as the post-production of Network/Intersect, often whilst the rendering of Network/Intersect’s high-definition video was taking place. 

  3. John K Ousterhout, ‘Scripting: Higher Level Programming for the 21st Century’, Computer 31, no. 3 (1998): 23–30. 

  4. Gensim: Radim Řehůřek and Petr Sojka, ‘Software Framework for Topic Modelling with Large Corpora’, in Proceedings of the LREC 2010 Workshop on New Challenges for NLP Frameworks (Valletta, Malta: ELRA, 2010), 45–50. Cornell Movie Dialogs Corpus: Cristian Danescu-Niculescu-Mizil and Lillian Lee, ‘Chameleons in Imagined Conversations: A New Approach to Understanding Coordination of Linguistic Style in Dialogs.’, in Proceedings of the Workshop on Cognitive Modeling and Computational Linguistics, ACL 2011, 2011. Available from 

  5. Tomas Mikolov et al., ‘Efficient Estimation of Word Representations in Vector Space’, arXiv:1301.3781 [^Cs], 16 January 2013, 

  6. Ousterhout, ‘Scripting: Higher Level Programming for the 21st Century’. 

  7. Whilst in France, I visited the regional contemporary art collection of Frac des Pays De La Loire which houses some of the French national art collection; the preservation team told me of the difficulty in maintaining computer-based works. 

  8. See, for example the Internet Archive Software Library, or Rhizome’s notable effort to archive and recreate net art from the 1980s to the present day. Internet Archive, ‘Software Library’, 18 June 2014,; ‘Rhizome Net Art Anthology’, Archive, Rhizome Net Art Anthology, (27 October 2016), 

  9. Many of the libraries and communities which exist around the Python language contain in-jokes and references to pop culture: the language itself is named after Monty Python’s Flying Circus; one of the popular libraries to work out time changes is named DeLorean after the time-travelling car in the Back to the Future movie franchise. 

  10. There are numerous online communities where programmers are able to discuss work: Stack Overflow, GitHub, as well as the many forums, wikis, conversation boards and websites which support various platforms and programming languages. 

  11. The term ‘flow’ was coined by psychologist Mihaly Csikszentmihalyi, following research into the creative process with Jacob W. Getzels in the 1960s. Csikszentmihalyi was ‘struck by the fact that when work on a painting was going well, the artist persisted single-mindedly, disregarding hunger, fatigue, and discomfort—yet rapidly lost interest in the artistic creation once it had been completed.’ Such a description could be applied to the state of mind I find myself in whilst programming. Jeanne Nakamura and Mihaly Csikszentmihalyi, ‘The Concept of Flow’, in Flow and the Foundations of Positive Psychology (Springer, 2014), 89. 

  12. Often this necessitates a period of decompression once I am finished (e.g. a walk) in order to return to any level of conviviality. 

  13. Peg Rawes, ‘Spinoza’s Geometric and Ecological Ratios’, in The Politics of Parametricism: Digital Technologies in Architecture, ed. Matthew Poole and Manuel Shvartzberg (London and New York: Bloomsbury Academic, 2015), 216. 

  14. Mike Isaac, ‘Inside Uber’s Aggressive, Unrestrained Workplace Culture’, The New York Times, 22 February 2017,; Farhad Manjoo, ‘Uber Case Could Be a Watershed for Women in Tech’, The New York Times, 1 March 2017, The blog Model View Culture offers regular feminist critique of Silicon Valley culture; see Feminist Technology Collective, ‘Model View Culture | A Magazine about Technology, Culture and Diversity.’, Model View Culture, accessed 3 May 2017, 

  15. Sam Knight, ‘How Uber Conquered London’, The Guardian, 27 April 2016, sec. Technology, 

  16. Stanford University’s website cites the Silicon Valley garage start-up myth as originating in 1939 with the founding of Hewlett-Packard in a: ‘Palo Alto garage. That garage would later be dubbed “the Birthplace of Silicon Valley.”’. Stanford Office of University Communications, ‘Stanford University History Part 3 of 5 - The Rise of Silicon Valley’, 18 August 2008, The “garage” (or “dorm-room”) start-up myth is found in numerous large technology companies’ publicity, from Hewlett-Packard to Google (whose ‘About us’ webpage has a section entitled ‘From the garage to the Googleplex’) to Apple (the garage where the company was reportedly founded was featured as a filming location in the 2015 film Steve Jobs), and Facebook (whose start-up myth was cemented in the Oscar-winning 2010 film The Social Network). Google, Inc., ‘From the Garage to the GooglePlex’, 1 May 2017,; Danny Boyle, Steve Jobs, 2015; David Fincher, The Social Network, 2010. 

  17. Rawes, ‘Spinoza’s Geometric and Ecological Ratios’, 217. 

  18. Rawes, ‘Spinoza’s Geometric and Ecological Ratios’; D’Arcy Wentworth Thompson, On Growth and Form (Cambridge: Cambridge University Press, 1917). 

  19. Rawes, ‘Spinoza’s Geometric and Ecological Ratios’, 218. 

  20. Alfred Korzybski, Science and Sanity: An Introduction to Non-Aristotlian Systems and General Semantics, Fifth edition, second printing (Brooklyn, N.Y., USA: International Non-Aristotelian Library, Institute of General Semantics, 2000), 58; G. Bateson, ‘A Theory of Play and Fantasy’, in Steps to an Ecology of Mind, 13. printing. (Ballantine, 1985), 177–93. 

  21. See, for example, the way in which ‘behavioural scripting’ is discussed in Patrik Schumacher, ‘Advancing Social Functionality Via Agent-Based Parametric Semiology’, Architectural Design 86, no. 2 (2016): 112–13. 

  22. Bateson, ‘A Theory of Play and Fantasy’, 193. 

  23. This might also be aided by the use of Schank and Ablesonian computer-scripted linguistic analysis. 

  24. Dementia Village Advisors, ‘Dementia Village’, 13 June 2014, 

  25. Sally Stewart, ‘Redesigning Domesticity: Creating Homes for the Elderly’, Architectural Design 84, no. 2 (2014): 80–87; Demos et al., ‘The Commission on Residential Care’ (Demos, 2014), 10, 18, 162, 167–69; Jeremy Story Carter, ‘Dementia Village Designed for Dignity’, Radio broadcast, Blueprint for Living (ABC RN, 3 October 2015),; Anne P. Glass, ‘Innovative Seniors Housing and Care Models: What We Can Learn from the Netherlands’, Seniors Housing & Care Journal 22, no. 1 (2014): 74–81. 

  26. Claudia Aradau and Tobias Blanke, ‘Politics of Prediction: Security and the Time/Space of Governmentality in the Age of Big Data’, European Journal of Social Theory, 14 September 2016, doi:10.1177/1368431016667623. 

  27. Orlando Patterson, Freedom: Volume I: Freedom In The Making Of Western Culture, Freedom (Basic Books, 1992); Orlando Patterson, ‘The Ancient and Medieval Origins of Modern Freedom’, in The Problem of Evil: Slavery, Freedom, and the Ambiguities of American Reform, ed. John Stouffer and Steven Mintz (Amherst and Boston: University of Massachusetts Press, 2007), 31–66. 

Back to top