Reading undocumented code feels like being an ant walking on a big painting. You can see the strokes of a brush and have an intuition of their direction, but what's missing is an overall idea of how the composition flows. Documentation provides guidance through the bunch of functions and statements that makes software, a bird's eye perspective. It is often the first thing one gets across when approaching a new library or programming language, and it shapes the way a developer thinks about particular piece of code.
At the very first encounter with a new script, details about its source code are unknown. Programming is a play _in medias res_, and documentation acts as narrator. Describing how functions are stitched together, or an algorithm is implemented, it sets the stage for developers to participate. Showing the different steps of a program and how they are connected, it offers entry points for interventions.
For example [Vue.js](https://vuejs.org/guide/essentials/lifecycle.html#lifecycle-diagram), a popular library for building web user interfaces, explains with a diagram the lifecycle of its components: at which moment data are received from a server, at what point an element is rendered on screen, and when it will disappear. What at the beginning feels like magic, gradually appears more clear. Presenting a structure means also presenting a way to reason about it. The reader gains some understanding and agency over the tools they are about to use.
The introduction to a program situates it also within a larger ecosystem: how to install it, and what dependencies it requires to work properly. As Geoff Cox and Winnie Soon elaborate on their decision of a downloadable code editor instead of a web one for their classes, code is more than just a single piece of software. It is also the relations with the configuration of one's own computer and operating system. (Cox and Soon, 2020)
The devil is in the details, and software as well: the translation between human and machine has to be negotiated with all the specifics of a particular programming language or platform. Sometimes for the web, sometimes for a hardware component, sometimes for another operative system. These _specs_ make every piece of code a bit alien and peculiar. Tinkering with code is not just knowing by heart a programming language, but rather having to deal with a lot of different recipes for different occasions, and know how to adapt.
Documentation is not just for beginners: it's a code companion. One never stops reading. Even experienced programmers must refer to docs when first encountering a software, and return to the references when they need a refresher on the syntax of a particular command. They continuously look at code from multiple distances: close to the source code through lines of comment—ignored by the machine, but much appreciated by fellows developers—or from printed books, along with pages of explanations and use cases.
This tentacular surface can reach a programmer in different moment of their life: from the _hello world_ to the _how to uninstall_. This is possible thanks to the multitude of shapes documentation can take: video tutorials and commands cheatsheets, _README_ files and complete guides featuring colored images. Daniele Procida proposes a systematic approach to organize this wealth of formats (Procida, 2017). His framework focuses on the needs of different kinds of readers: by leveraging between practical steps and theoretical knowledge, it charts four main modes of technical writing. Each format comes with its own approach and intentions, and in response to different questions.
This system organizes knowledge around code in a way that tries to meet every user possible. _Tutorials_ offer entry points for the newcomers, while _explanations_ unveal core mechanisms for more navigated readers. _How-to guides_ teach how to get the work done, while _references_ report lists of information ready to be consulted. Different documentations for different readers for the same code.
A text that fails to address who's reading can result inaccessible and frustrating. Although the Diataxis framework doesn't encompass every particular situation, its structure offers good aid to situate documentation within different perspectives. This turns out to be really helpful in the process of writing, as a way to fine tune tones and modulate the nature of shared info.
The same is true for the additional layers of meaning necessary for a process of world building. To enchant software with political aesthetic (Ranciérre through Soon and Cox, 2020) is required to operate at different scales. Within both public and private dimensions, with technical and social frameworks. During a workshop for example, people meet face to face. Here togetherness can glue technicalities as a paratext, questioning reproduction of knowledge and its power dynamics. (See for example _Feminists Federating_, mara karagianni, ooooo, nate wessalowski, vo ezn in Toward a Minor Tech - A peer reviewed Newspaper vol 12, 2023) (Note that the opposite effect is also true, with technicalities as paratext conditioning how people are together)
The potential of documentation to orientate software in the world clashes against some big elephants in the room, and tech culture should stop keep them in captivity. Sure, it would be nice if developers could rely on several kinds of documentation when approaching code. Unfortunately often there is not even one available.
Writing documentation is demanding. It's more delicate than programming, and require a whole set of skills usually not treasured by the dev community. A kind of emotional intelligence and sensitivity far to be found in the competitive wastelands of the IT industry. Here no one wants to write documentation, nor hire someone to do it (Gabriel, 1996). As a result, in a world where software thrives, documentation still seems to be a scarce resource.
It's ok, someone could argue, every question that can be asked on Stack Overflow, will eventually be asked in Stack Overflow (versioning Atwood, 2007). The popular Q&A website for developers is just an example of digital knowledge as a shared effort, together with the endless mailing lists, forums, discord servers and dedicated sources for whatever topic. It's astonishing how online communities can tackle any problem in no time.
But it's not rare for these places to feel unwelcoming, or even hostile. In 2018, Stack Overflow publicly admitted that there was a problem concerning their userbase. The space felt unfriendly for _outsiders_, such as newer coders, women, people of color, and others in marginalized groups (Hanlon, 2018).
For years there have been discussions on the platform about tone. At the question _"Should 'Hi', 'thanks', taglines, and salutations be removed from posts?"_, one of the founders of Stack replied with a [RegEx](https://meta.stackexchange.com/a/93989) to filter _automagically_ what some of the experienced users perceived as noise. This _regular expression_, a way to target specific text patterns in programming, started then to be silently applied to every query sent to the website, trimming out etiquette and leaving just technicalities.
Far from being just an isolated problem, this crudity is deeply embedded in the IT discourse, soaking through technical writings as well. The denigrating expressions of superiority in matters concerning programming that Marino calls _encoded chauvinism_ (Marino, 2020) constitute the main ingredient in the brew of toxic geek masculinity. _Real programmers_ don't use this code editor. _Real programmers_ don't use this programming language. _Real programmers_ don't care about others feelings. Etc.
Ellen Ullman's accounts of the emotional dumbness of her _real programmers_ colleagues give an insight of a problematic behavior, first intercepted and then capitalized by the IT industry. _"In meetings, they behave like children. They tell each other to shut up. The call each other idiots. They throw balled-up paper. One day, a team member screams at his Korean colleague, 'Speak English!' (A moment of silence follow this outburst, at least.)"_ (Ullman, 2017)
Programming means to deal with picky stubborn machines that don't overlook a single typo. It requires a high tolerance for failure. It is frustrating. But to project this frustration onto other users, as in the `Read The Fucking Manual` typical response to a request for help, it's a form of negative solidarity: others should suffer as I did when trying to understand how code works.
Mark Fisher used this image in the context of labor under capitalist realism, where workers are forced in precarity and isolation. Here as in a downward auction, people are driven to bring down each others. (Fisher, 2013). I'm using it with a focus on the emotional component: not just lack of empathy and solidarity, but also reproduction and legitimisation of toxic behaviors in coding communities.
When all the energies are invested in debugging, the quest to find and solve all the problem in a program, and no space is left for introspection, programmers start behaving as machines. This poses serious barriers to the participation of others in the making of software.
Here some examples that go in a different direction.
p5.js is a Javascript library started by the artist Lauren McCarthy as online porting of Processing, itself being a project to promote both software literacy within the visual arts, and visual literacy in software development.
The work of documentation around p5.js offers entry points to the world of programming, taking care of not taking too many things for granted. The amount of care and efforts in their tutorial about [debugging](https://p5js.org/learn/debugging.html), for example, results in a welcoming article with different levels of accessibility. Here the drawings help to visualize complex concepts, the tutorial format is beginner-friendly, and the narration makes for an interesting reading also for who is already familiar with debugging.
One of the most frightening aspects of programming is to be confronted with stack trace errors: when things don't work as espected and red error messages show up. These scarlet letters delivered by code are useful for developers to identify where in the program the error happened, but often they are dense of technicalities and difficult to decipher. Worst case scenario for beginners. The explanations from the p5.js Education Working Group tackle on this nightmare showing not only how to read technical errors, but how to think through them with different debugging methods. From here the stack trace starts to get less alien and scary, less like a wall and more as a starting point to solve the error.
Another reflection on entry points and gatekeeping comes from the english artist and writer James Bridle. Their practice explores the cultural and ecological impacts of digital computation, walking and jamming the thin line between what is showed and what is kept hidden in the technological landscape we live in.
In the artwork `welcome.js` (2016), a tiny Javascript library published open source on GitHub under a permissive MIT license, Bridle injects some greetings in their website (and in all the websites that include the library) to welcome users to the browser inspector.
The artwork is hidden below the surface of the website, printed in the console of the browser inspector, a tool which allows users to see underlying code of the website they are visiting. From here `welcome.js` provides some guidance for newcommers to access, inspect and modify the source code of web pages. A process to open doors and let people in, giving them more agency by demistifying technology.
The script comes as response to a similar, but opposite and violent, message printed by Facebook in the console on their platform to halt users from accessing the page hidden structure. The platform adopted this approach to prevent scams and Self-XSS attacks to their users, that could have been lured into running malicious code in their own browsers by malicious people. However, instead of encouraging their userbase to understand, explore and eventually feel more safe against cyberattacks, the company opted for a full stop, marking a clear line between user and developers.
Who writes code documentation tends to assume a certain kind of reader.
This reader is often thought as similar to who's writing: familiar with the topic, at ease with technicalities, and capable of cut through the precise lingo and esoteric references offered as explanation. Eventually (and in most of the cases) who's reading happen to be someone else. This mismatch transforms entry points into barriers that filter out who can participate to coding.
Historically, technical manuals and software specifications have been addressed to a very specific public of male engineers.
This gendered language comes with an embedded and gendered separation of roles. Consider the excerpt from the GNU `gettext`: _"In this manual, we use he when speaking of the programmer or maintainer, she when speaking of the translator, and they when speaking of the installers or end users of the translated program."_
As Mara Karayanni argues in _Read The Feminist Manual_, published under Psaroskala Zine, the stubborness against gender neutral language in technical writing is but a pretext for refusing to waiver the priviledge of the male programmer.
Writing code documentation is tricky because requires some degree of astral projection. Who's writing is asked to leave their body and describe code from a different perspective. From the point of view of someone else. Unlearn what seems to be obvious and be generous with future readers.
Who is going to read this piece of documentation?
Target to reach vs public to create
Learn to code is like learning another language: not just a new bag of words and a different grammar, but rather a different way to think. It means not just learning what are the things that move and the ones that are fixed, but also how they relate and make meaning together.
Coding means to express ideas with the reduced vocabulary of a programming language.
vvvv is a visual programming language that offers an agile approach to prototyping, adoperated especially in the context of interaction design. It's focused on rapid and iterative work of fine tuning, necessary when dealing with real time inputs, such as sensors, live data, or human interaction.
