5. Minimal Computing

Definition, philosophy, history

Minimal computing is an emerging conversation in digital humanities about the need for computing solutions in parts of the world where people have only limited access to internet connections and digital technologies. This lack of access, as well as the learning curves that accompany it, exclude huge segments of the world population from participating in digital work. Historically, the cost of purchasing hardware and learning to use complex programming languages has kept digital scholarship disproportionally in largely endowed institutions in the northern hemisphere. Minimal computing emerged in response to the need to decolonize discourse in and about the Global South and its diasporas by making technology more widely accessible, independent and easy to use. Alex Gil notes that many academics who are called to take part in institutional initiatives that seek to promote inclusion and diversity are also engaged in the economically challenged spheres and hemispheres where fundamentals of computing are scarce (“Design”).

The material inequalities that minimal computing engages with make it a critical movement that asks for balance between gains and costs in areas that include social justice issues and de-manufacturing and reuse, and the effects of e-waste. Ownership of records and files and control over knowledge production are tremendously important and even more so in places where records and documents are subject to censorship or seizure by authorities and political groups. Minimal computing thus seeks ways to use computers under significant constraints of hardware, software, education, network capacity, global relationships of power and knowledge production, and other economic, infrastructural and material conditions, including issues of aesthetics, culture and the environment.

Minimal computing seeks solutions that can live on physical devices rather than proprietary clouds, and includes the maintenance, refurbishing, and use of machines which are not high performance as well as machines built specifically for high performance in specific tasks but with minimal additional needs. One of the key building blocks used in minimal computing is a static site. Unlike a dynamic site, which only comes into existence when the reader summons it up by with a click of their mouse, a static site already exists in the files on the computer and can be generated without a connection to the internet. Gil’s description of the difference between a static site and a dynamic site is helpful in understanding how important this difference is:

A static site differs from a dynamic one in the way the documents are produced, not in whether they are interactive or not on the front-end, a common misconception. A static site is produced long before the reader asks for one of its documents on the browser. A dynamic site is produced on the fly, at the moment the reader asks for a document on the browser. In most cases, a static site implies the absence of a database. This has enormous implications for the production of literary documents that have hopes of becoming part of a record. Having your documents recreated every single time you need them based on an algorithm is probably not a good bet for the sake of longevity, not to mention the cost of using electricity to re-create the document every time, when once or twice probably suffices.¹

Just by browsing through its elegant front end, one would never know that archipelagos | a journal of Caribbean digital praxis - is housed on a static site. archipelagos is a platform for open-ended and multiform engagement through the creation of archives, databases, cartographic applications and other digital projects among the people of the Caribbean and its diasporas (archipelagos)¹ . It is fully open access and charges no author fees or copyright privileges. The archipelagos website and PDFs are generated from markdown files using Jekyll and ConTeXt², respectively, resulting in a lightweight mobile-friendly website that works well with mobile phones, bandwidth differentials, and Caribbean data costs. archipelagos’ philosophy, described on its “About Us” page, perfectly exemplifies the radical potential of minimal computing for digital humanities work: “[its] cutting edge comes […] not from a misguided sense of techno-utopianism—or worse, techno-determinism—but as a new small axe, hacking away still at that master’s house” (archipelagos).

An example of how minimal computing work can preserve historical information, strengthen cultural ties and enrich the geographic imagination is the Sudan Memory Project, which used a variety of minimal computing methods to digitize and conserve written and photographic material pertaining to 19 major ethnic groups and including over 100 languages and dialects that was being held in Sudanese cultural institutions, where decades of war had put much of the cultural heritage and practices recorded in Sudanese archives under threat. This project, led in 2018-19 by Professor Marylin Deegan of Kings College in London, U.K., set up a professional cataloguing and digitization service in Khartoum as well as on roving facilities in other regions. A team of researchers and archivists digitized a large volume of at-risk content, including 100,000 maps, more than 10 million fragile photographs and AV materials documenting disappearing cultures and customs, and staff and volunteers in local cultural institutions were trained with the view to building digitisation and conservation capacity in Sudan in the future (“Conserving Sudanese Cultural Heritage”).

This kind of digital conservation work can also be done with Wax, an exciting minimal computing project for producing digital exhibitions focused on longevity, low costs, and flexibility. Wax produces beautifully rendered, high-quality image collections and scholarly exhibits by means of Ruby gems for processing image data and associated metadata which live as digital exhibitions on a static site built with the Wax Jekyll theme. Thanks to the way it tiles images, Wax’s files are flat and weigh very little, making it a truly portable offline alternative to Omeka³. The Wax project was led by Marii Nyröp in 2018 and is currently maintained by Marii Nyröp and Alex Gil at Columbia University Libraries. It uses open source libraries and frameworks including Jekyll, IIIF, OpenSeaDragon, Rake, and ElasticLunr. Wax sites use flat HTML, CSS, and JavaScript files that do not need to communicate in a complex way back to a server. This makes them cheap, safe, and generally easier to maintain than some of the heavier software for displaying graphics that many of us may be familiar with. The skills needed to create Wax sites are largely transferable for use in other digital projects, such as making standardized image files and metadata records and builds around them, handling canonical information, scholarly content, and site styling.

Wax has a relatively high but general-purpose learning curve which includes using the command line, git, Jekyll, creating and normalizing data files, using file-naming conventions and best practices, and editing HTML, CSS, and some JavaScript (Wax). Wax can serve well as a pedagogical tool for teaching these skills but may be too ambitious for an introductory digital humanities course. However, wax is a good workflow for building collections of graphical resources for use in digital humanities projects and could become a library for students in prison to use.

No Connect is another minimal computing project that could be fun to use in a prison classroom. It is a template for Jekyll static site generator that works on a local machine or a USB key, without an html server, based on the standard Jekyll theme and developed by Alex Gil and Dennis Tenen at Columbia’s Group for Experimental Methods in the Humanities in 2015. It taps into the affordances of sneakernets⁴ and has as one of its goals to “carve out space for digital work to continue under adverse circumstances” such as those caused by an unfriendly government or copyright regimes (Gil and Tenen). Committing text and code to memory on USB sticks may prove too volatile to do in prisons given their strict prohibitions on recording information and their definitions of contraband. Maybe, however, if the instructor is held accountable for all student use of devices, it could be allowed.

USB memory sticks and a case of laptops were taken into Rikers Island correctional facility by Columbia’s Experimental Methods Group in 2015, for what is probably the most exciting minimal computing project done in prison in the U.S.. Rikers Story Bot is a coding workshop and an algorithmic storytelling project that started as an initiative from the Center for Justice at Columbia University, with the idea of running an “Intro to Python” workshop for the young people incarcerated at Rikers and for Columbia University students interested in digital literacy (Rikers Bot). Students built a twitter bot with Python and narrated their story in tweets, establishing a chain of communication between communities on the outside and the jail. As of May 3, 2020, there are 529 Rikers Bot tweets. The most recent is from 2018. Not much was needed to carry out this powerful project, and a tremendous amount – for the participants, the public including the academic communities and other communities, and the historical record - was gained.

In a discussion about minimal computing something must be said about the Rasberry Pi (Pi), a series of small single-board computers developed in the U.K. in 2009 for teaching basic computer science in schools and around the world. In 2014, James Smithies and a team of developers launched a digital humanities project consisting in building jamessmithies.org, a virtual research environment (VRE)⁵ on a Rasberry Pi. Smithies’s goal was “to use a minimal computer and free software to create a prototype environment that is under my control, without the constraints and expectations of a large project, and (crucially) capable of ongoing extension and elaboration by a lone scholar” (“Full Stack DH”). The choice to build a VRE on a Rasberry Pi is interesting because Pis have very limited computational power, so if something works on a Pi it would arguably work on almost any device. Smithies’s VRE at jamessmithies.org is an API built with Django, a Python-based web framework designed for newspaper websites. While jamessmithies.org looks like a website, it has a blog that incorporates a Twitter Feed as well as a complete downloadable Ubuntu operating system, among other resources.

Smithies’s VRE experiment is also interesting from a minimal computing perspective because Rasberry Pis are cheap. A closer look, however, shows that if you want to use a Pi as you would a laptop, it is not really that cheap. The Pi basic computer or mother board costs between $35 [1 GB RAM] and $55 [4 GB RAM] (Rasberry Pi). To turn that into a laptop, you can buy a desktop kit for about $120 from Rasberry Pi (“Products”). In addition to this you still need a monitor and keyboard, so the price is about the same as buying a cheap laptop and installing Ubuntu, which is free and open source (Smithies installed Ubuntu on a Rasberry Pi to replace Raspbian, the closed source operating system that comes with the Pi). And all of this is a lot of work, which, as noted by Gil in 2015 in “The User, the Learner and the Machines we Make,” comes with a steep learning curve and thus one of the constraints that minimal computing, as opposed to minimalist computing, aims to avoid (Gil, “The User”).

It is interesting to get involved in the material affordances of the hardware we use.There is a lot to learn from people in the global south who are by necessity much more resourceful and experienced in negotiating material affordances than most people in the U.S.. However, my goal is to build things with computer programs rather than build computers, and if I am operating in a place where there is no hardware at all and do not necessarily have advanced technical skills, the Pi is not a good choice. I am much better off using a cheap, refurbished or donated laptop, installing Ubuntu, and collaborating via a LAN. However, an interesting takeaway from Smithies’s minimal approach is that because he hosted his VRE on a a tiny computer with only 2GB RAM for three years⁶, he showed how much can be done if files and programs are kept as lightweight as possible so that they can be accessed or run on any device.

Minimal computing does not aim to replace complex systems, but rather to work in tandem with them. The work that digital humanists develop using complex systems – for example, Manifold Scholarship⁷, which allows independent writers to publish beautifully curated works, or Omeka, notably used in projects like Preserve the Baltimore Uprising 2015⁸, which digitally preserves crowdsourced content that documents the events surrounding the death of Freddie Gray - is excellent use made of powerful tools. It is of the utmost importance that minimal computing solutions be relied on alongside complex systems so as to enrich the academy with more practitioners who do not have access to powerful computers or reliable broadband.

Because digital humanities has broken away from the academy in many ways at the same time that it is increasingly becoming an integral branch of most scholarly fields, it welcomes non-traditional practitioners, and its collaborative yet dispersed approach to project development can afford a degree of privacy that might be reassuring to people just out of prison and starting to look for new footholds in an unfamiliar world. Digital humanities inherited a lot from communities of hackers where outcasts and otherwise marginalized people found relatively safe places to be at home. It is therefore a real way to bring in and retain non-traditional students in our universities.⁹ Because minimal computing’s search for solutions revolves around the central question of what is needed and explores how to meet these needs at minimal cost, it is exciting to approach teaching digital skills in prison, where needs and constraints are very specific, in this way.

Minimal computing and pedagogy

A classroom limited by the significant constraints particular to prisons is an interesting place to develop technical skills, because everything has to be built from scratch. Building from the ground up is the best way to acquire solid digital skills. An artisanal approach gives learners tools to engage in sustainable practices with a lot more freedom and versatility than if they were to learn digital humanities methods of inquiry with “soft” tools such as Omeka, TEI, Voyant, Scalar, Tableau and Carto, which we cannot use anyway because they mostly require a connection to the internet, which prisons do not have. Furthermore, because one of the main aims of teaching digital humanities in prison is to empower students by giving them autonomy and control over what they produce, it is preferable to stay away from ready-made interfaces we do not own. After leaving prison, students who completed an introduction to digital humanities course will be well prepared to learn to use any number of interfaces they might need to get a job, all the while conserving an understanding of the larger political and socio-economic context such interfaces exist in. Gil observes that by learning methods of digital humanities inquiry with the above-mentioned soft tools, “students learn very little about computation, systems, networks, and by extension the social implications of the modes of production that they enable. Even more importantly,” writes Gil, “use of these tools hides the vectors of control, governance and ownership over our cultural artifacts” (“Design”).

Solutions for working under prison constraints

Hardware

If the introduction to digital humanities I envision were offered to, say, fifteen incarcerated students, it could be done with seventeen laptop computers, one per student, one for the instructor and one which would function as the local server. These would all be connected through a Local Area Network (LAN).¹⁰ The LAN connection would provide students with access to the course syllabus, readings and tutorials, and allow them to share work, edit and annotate code and texts, engage in conversations on a blog, build and play games, and do a lot more. The LAN connection setup is simple and inexpensive, and, especially if it is hardwired, allows users to share large files without a connection to the internet.

In the setup I imagine in prison, the LAN could be established in two ways. One of these would be to connect the server laptop to a router and a network switch (the network switch would expand the number of ports on the router). The other way would be to hardwire the LAN by connecting the server laptop only to the network switch, and the network switch to the other laptops. In this case no router would be needed. From a minimal computing perspective, it is preferable to use only the switch, as we are interested in avoiding the use of more hardware than we need. If, however, some functionality only the router can provide is needed, a router should obviously also be used. If possible, each student should be given a USB to practice a lot of version control (Git). These USB sticks would serve the same function as GitHub in a very rudimentary way, allowing students to push and pull changes to shared repositories.

Software and operating system

In the Intro to DH course model I developed as part of this project course readings and tutorials for learning technical skills are made available to students on a static site built with Ed., a Jekyll theme developed by Alex Gil. This pedagogical use might be a new use of Ed., but it is adapting beautifully. At one point I considered using Hugo, another static site generator, because Ed. does not work very well with Windows as of yet (at that point I was considering using PCs).

I chose to work with Linux rather than Mac OS or Windows because Linux is free and open source. Furthermore, it bundles Ed. beautifully.¹¹ A big concern when making the choice to go with Linux was that for purposes of using their newly acquired technical skills to gain employment after leaving prison, it would be good for students to be familiar with an operating system more widely used. Macs were out on account of their price. After some deliberation I bought a cheap Dell laptop, wiped out its Windoows operating system and installed Ubuntu (Linux), to see how that felt. It felt great and will be perfect for students who are learning to code.

Next: 6. P2CP: a Case Study and Possible Proposal