Fig 1. The self-operating napkin (September 1931)

\’gold,berg\ adjective.
A comically involved, complicated invention, laboriously contrived to perform a simple operation.

Webster’s New World Dictionary 

In the early 1900s, the cartoonist Rube Goldberg drew a series of mechanical chain-reaction illustrations in a bi-weekly series called “The Inventions of Professor Lucifer G. Butts,” which consolidated as a referent of complex means to execute daily life simple tasks in “the most ridiculously inefficient way possible.” Goldberg never built a physical machine; instead, these machines were satiric artistic representations that prompted thinking and critique about the technological developments at that moment.

Rube Goldberg machines have four main features:

  • They have serial connections between one step and the next since they represent chain-reaction mechanisms.
  • They use daily life elements and aim to execute trivial tasks in daily life, like using a napkin (Fig. 1).
  • They illustrate or materialize tensions between the complex and the simple
  • They are usually ephemeral, in the sense that once a step, or the whole machine is “completed,” re-executing it implies assembling it again partially or totally (think, for example, of the M step of Fig. 1; once the rope is cut, a new one has to replace it to execute that step again).

As think-prompting machines, instead of (just) extended domino shows, Rube Goldberg machines provoke questions such as what is simple and what is complex? How does simplicity transform with time? How really simple is what we perceive as simple?

This work tackles these questions through the creation of a platform to assemble Rube Goldberg machines connected through the Internet.


Technical Design

Fig 2. shows the proposed architecture. This architecture views the Machine as a complete structure that links modules through the internet. Each Module is a physical, traditional Rube Goldberg Machine with internal chain reactions. The representation of the steps in the figure is a nod to Spiegelman’s words: “Rube Goldberg is the guy who figured out how to get from point A to point B using all the letters in the alphabet.”. This architecture allows connecting different machines to assemble a Meta-machine.

Architecture diagram

Fig 2. Architecture of the Rube Goldberg (Meta-)Machine connected through the internet. 

The technical design includes the representation of both Machines and Meta-Machines as double-linked lists, where each node contains a link to the next and other to the previous one and a series of data that describe the module or the machine. Additionally, I modeled a Participant as the creator of a Module. Fig. 3 presents the data structures.

Data Structures

Fig 3. Data structures for the Meta-Machine, the Machine, the Module and the Participant. 

I used a json format to represent these data structures (Fig. 4)

JSON sample

Fig 4. JSON sample of the representation of machines and modules. 

One important property within the structures is state that allows to model Machines and Modules as finite-state machines. Fig. 5 A and B shows the representation of the states and transitions for a (Meta-)Machine and a Module, respectively.

Finite-state representation for machines and modules
Fig 5. Finite-state representation, including states and transitions, for the (Meta-)Machine, and the Module. The Meta-Machine and the Machine have the same states representation. 
The architecture is service oriented, providing services for the Meta-machines, the Machines, the Modules, and the Participants. The functions were implemented over Firebase, using Cloud Functions

Narrative Design

The proposer defines a common topic for the entire machine, called a campaign that could be related to a historical event, a scientific concept, a book, or a character, among as many as they may imagine. The narrative flow follows a “cadaver exquisito” (exquisite corpse) strategy, so each participant only knows how the previous module ends, and has to inform how its own finishes.
Campaigns
Fig 6. Examples of possible topics or campaigns. 

Prototype

Rube Goldberg machines’ outputs are the automation of simple operations. Among the classical ones, we have to turn on/off a light, water a plant, pop a birthday balloon, and light a candle. Following Goldberg’s seminal idea of these machines as satiric representations of automation and new technologies, I proposed scrolling reels as the “simple”, daily operation that may prompt thinking and critique (Fig. 7). Discussing with the friends who participated in this prototype (Acknowledgements), we wondered whether it was actually a simple operation. Is scrolling reels only related to the gesture of the finger? How much energy is (really) consumed in the act of scrolling for hours?

Screenshot scrolling reels

Fig 7. Scrolling reels. 

We decided to make a prototype to tinker with this idea, setting these constraints:

  • About the means: We could use physical (example) means and connect to the Internet through embedded platforms such as Arduino, Raspberry, or NodeMCU. We could also develop a digital machine using tools like Minecraft (example), Construct, or Unity3D. Finally, we could combine digital and physical parts to build a hybrid or tangible machine (example).
  • About the narrative: From the question of the actual simplicity of scrolling reels, we arrived at “Energy” as the narrative guideline. What is energy? What forms does it take? How is it transformed? How is it transported? How do we perceive energy? Why (whether) is “energy” a problematic concept? How can we play with energy? So, we thought and tinkered with these and more questions revolving around energy and created a quick prototype given the time restrictions we had (one week of extra time in addition to our jobs and personal life).
Vid. 1 Prototype of a Rube Goldberg Machine through the internet for scrolling reels. 

Final Reflections and open questions

Predictability 

The machines Goldberg conceived often included unpredictable behavior, such as that of animals (e.g., Fig. 1 step E). Similarly, material constructions of Goldberg machines involve high degree of unpredictability. In our prototype, Gabriel’s dog playing with the ball or the changing weather affecting the light in Carlos’ house introduced difficulties in operating the machine. In the Rube Goldberg Machine connected through the Internet, local modules tend to be the most vulnerable to unpredictability, whereas the connection between modules (via the Internet) is the most stable component. However, the initial design of the Internet component in this proposal was focused on ensuring the creation of the modules and the communication between them during operation rather than supporting testing, making trial-and-error difficult. Over the course of creating the prototype, it became clear that Goldberg Machines, being more for prompting thinking and tinkering than for achieving functional precision, require a prioritization of the testing phase. In future iterations, providing functions that allow participants to simulate triggers from preceding modules and confirm that subsequent modules receive the signal could streamline testing.

Registration

Modernity has certain obsessions that Goldberg machines help to criticize. The most obvious is the obsession with control and automation, which demands continuous and efficient operation. Goldberg machines, in contrast, are inefficient and temporary, often designed for single use. Another modern obsession is with recording, documenting, and generating inscriptions of everything. How do we keep a record of an Internet-connected machine? And is it even necessary? One of the greatest difficulties we faced during prototyping was recording the machine’s operation in real time. We used a Zoom call, where the host had to manually switch between cameras as each module finished –a challenging task to perform efficiently. I also included a simple visual interface on a public website that tracked each module, displaying acronyms for the participant and a color-coded states (blue: done, yellow: operating, green: get prepared; see Vid. 1 top section). Future work could automate the camera switching process within this dedicated site, using a custom development for video meeting software based on platforms like Jitsi.

Between centralism and distribution
How distributive is a machine connected through the internet, where each module operates independently but connected to one or two more modules? The premise of Internet-connected Goldberg Machine was simple: local modules, developed independently by tinkers from anywhere, only needed an internet connection to receive a trigger to start and to send a message once completed. Regarding communication and coordination, when subscribing to the machine, each participant was supposed to only know the final step of the previous participant and to inform which would be their last step. However, questions still unsolved arose. What could happen if many participants subscribe at the same time? How do we define the initial and final modules? How much time should be provided to a participant to communicate their final step? Is it necessary to have a central manager for the machines? Who controls what?
Invisibility 
The fastest and the most invisible steps were the transitions between one module and the next one, through the Internet. For the sake of the machine function, the API “made invisible” this step as much as possible becoming paradoxical in two modes. First, because even when it seems timeless, the Internet component is the most complex at the infrastructure scale. Second, because the transparency of this connection in the chain-reaction goes in opposite direction to the illustration and materialization of classical Goldberg machines that “remind us of a time when we could see how the machines around us worked.”

Acknowledgments

Giovanny Piedrahita, Jaime Herrán, Carlos Hernández, Gabriel Ávila, Juan Carlos Ponce.

References

[1] https://www.rubegoldberg.org/all-about-rube/a-cultural-icon/  

[2] https://www.tcj.com/rube-goldberg-butts-in/  

[3] Inside the surprisingly dark world of Rube Goldberg machines – The Verge 

[4] D. Olsen and M. J. Nelson, “The Narrative Logic of Rube Goldberg Machines,” in Interactive Storytelling, N. Nunes, I. Oakley, and V. Nisi, Eds., Cham: Springer International Publishing, 2017, pp. 104–116. doi: 10.1007/978-3-319-71027-3_9.

Leave a Reply

Your email address will not be published. Required fields are marked *