This text is a translation from MAUDET, Nolwenn. Design tactique. Tèque, 2023/1 N° 3, p.40-69. DOI : 10.3917/tequ.003.0040. URL.

Highlightings have been added by Doriane, from a reading point of view of the Declarations research.

Tactical Design

Digital technology has reshuffled the cards of design, as it has seriously challenged the possibility of designing alternatives. It is generally easy to switch makeup or car brands but it is much harder to leave many of the applications we are now using. Of course, designers can still propose new social networks or video platforms that would work differently from YouTube for example. But as these monopolistic platforms grow, the possibility of seeing one’s product chosen by users falls steadily. If we take Instagram as an example, it would be technically very easy to create an alternative, there are actually quite a few already, but what counts is not the interface or the proposed functionalities, it is that making people move to a new platform or making them leave a social network is extremely difficult¹.

If we cannot compete on equal terms with platforms that exist in the digital world, then trickery and tactic become all the more tempting modes of action, pushing some designers to enter resistance. Rather than struggling to come up with alternatives whose chances of success are minimal, we can at least try to counter existing systems. All the more so as designers are generally far removed from the decision-making circles that can have a direct effect on these platforms, in the technology companies themselves or in the regulatory bodies.

It is these disparate practices, which I have named tactical design, that I would like to talk about. For several years now, I have been seeing a handful of projects that no longer seek to fill a need, a gap or propose new ways of doing things but, instead, initiatives that seek to directly and explicitly counter existing platforms, services or applications. In fact, you may already be using them on a daily basis. The best-known example is certainly the ad blocker, an accessory that has become indispensable for many when browsing online, and whose sole purpose is to counter the display of ads. And even when they are not designed by designers per se, they still intersect with a host of design issues.

Recent forms of design protests

Protest practices are not new to design, or even reserved for the digital realm, and I would like to place tactical design briefly in this long history. If we don’t want to go too far back, critical design is perhaps one of the main models of contemporary protest design. The term, found in the 1999 book Hertzian Tales, was proposed by English designers Anthony Dunne and Fiona Raby to describe their own practice. In the context of critical design, while objects are intended to challenge the existing, they are not built in direct opposition to technologies or systems. Rather, they seek to question the obvious and open up new possibilities. Rather than solving problems, the idea is to identify contemporary issues and make them visible. This is also reflected in the choice to distribute their work through gallery exhibitions rather than in shops. More recently, in direct filiation with critical design but also drawing on the humanities, design researcher Carl DiSalvo developed the idea of adversarial design in his eponymous book published in 2012. It shares with critical design a desire to question what already exists, but assumes a more explicit political commitment. Antagonistic design also assumes dissensus rather than consensus, and intends to show that there is no one true way of seeing the world. It asserts cleavage, in contrast to many design approaches. Among the projects DiSalvo cites as examples, many are forms of visualization. If the desire to confront the world is present, rather than a direct resistance to digital objects or products, the projects presented by DiSalvo often seek to reveal, make visible and read existing power structures.

Tactic and Design

The “tactical design” label I would like to propose here shares a lot with these approaches, in particular an explicitly political and often antagonistic approach, but it departs from them in its claim to actually solve the problems it identifies, rather than making them visible or sparking debate.

The term tactic is taken from the terminology proposed by French philosopher Michel De Certeau², and is often used to refer to the appropriation of everyday objects by those we call users. DiSalvo himself uses the term to extend the definition of who designs and who participates in his work³. Tactics, as opposed to strategy, is a practice that unfolds in relation to what already exists, and can only exist in relation to it.“Tactics can only exist in relation to the other. So it has to play with the terrain imposed on it, as organized by the law of a foreign force. It does not have the means to remain within itself, at a distance, in a position of withdrawal, anticipation and self-gathering [...]”⁴. The term also takes on a strong temporal dimension, here too in contrast with the term strategy. A tactic is “a skill in seizing the occasion, the propitious moment, the opportunity”⁵.

Beyond design, the term tactical has recently been used a lot in the context of urban planning. For several years now, we have been talking about tactical urbanism. You may have heard the word in connection with the implementation of coronapistes, the temporary bicycle paths that were hastily created after the confinement of 2020. But in general, and to put it very quickly, it refers to the practices of reappropriating a given space. Tactics employed include, for example, reassigning parking spaces to street furniture or planted areas. A potentially very diverse set of practices, but generally thought of as being initiated by a local community and deployed on that same local scale, although many urban planners also employ tactical urbanism to test things or bypass tedious processes. As with digital technology, the fact that public space is a scarce commodity, and that it is almost impossible to extend it in cities, leads to an antagonistic approach to oppose its current uses. While the term may seem appropriate to reuse when talking about a range of existing practices in digital design, there is an interesting irony in mobilizing it to talk about designers’ practices when, in de Certeau’s work, it is actually used to talk about the practices of those who are faced with designs.

Creeping in the opponent

The first example of tactical design I want to introduce is a project called Turkopticon⁶. Before explaining how it works, I need to briefly introduce the platform it is up against: Amazon Mechanical Turk, a website and service operated by Amazon as a meeting place between, on the one hand, employers who have a large volume of micro-tasks to get done, and, on the other, workers who perform these tasks in exchange for money. Amazon legally defines workers as contractors, i.e. people who are not subject to the same labor laws as employees. This allows Amazon to prevent these workers from asking for minimum wage⁷.

Going back to Turkopticon, it is a project initially developed in 2008 by Lily Irani, a researcher in human-computer interaction (HCI) at the University of San Diego in California, and Six Silberman, a computer scientist. Rather than fighting directly against this new form of work, recognizing that it may offer benefits to workers that they are unable to understand from their perspective, they set out to identify its limits. According to their initial survey of platform workers, one of the problems was the huge gap between the tools given to employers and those given to workers⁸. For example, it is impossible for the latter to file a complaint if an employer feels that a task has not been carried out properly, but they also lack tools to fight the low wages on offer. Turkopticon’s primary objective was to make these problems visible by proposing a tool that directly addresses them, showing that it would be very easy for Amazon itself to respond to this problem.

In concrete terms, Turkopticon lets workers rate and comment on the different employers on the platform. It also offers them a discussion space, so they can get in touch with each other and build solidarity. The tool was developed as a browser extension, more precisely a “user script”, which runs directly on the Amazon Mechanical Turk web page and adds functionality to it. Far beyond its initial academic and activist scope, the tool was adopted by the Turker community and is still used and maintained today.

Turkopticon can be described as a form of tactical design because of two main aspects. Turkopticon’s first tactic is to insinuate itself directly into the ennemy’s realm. Turkopticon is not a third-party service built alongside Mechanical Turk; it has invited itself in and taken its place directly on the site itself, which is possible thanks to open Web standards – more on that later. Rather than recreating a completely separate site, Turkopticon takes the existing interface and enhances it. On the page of the official site that lists the jobs on offer, a small icon appears next to the employer’s name, allowing users to hover over notes that have been given to this employer by other workers on the platform. This small drop-down panel also lets workers rate or comment on this employer in situ. Insinuating yourself in this way means that you don’t have to redevelop an entire infrastructure, and that you can make use of both the interface and the data on the Mechanical Turk platform. Information is visible when and where it is most relevant to workers, making it all the easier to adopt such a tool, as there is no need to use another site or even switch page.

The second tactic employed is the idea of re-employing a type of functionality widely used in this kind of platform, namely the idea of rating. The principle of permanent rating, which has been widely developed with digital services, has often been analyzed as a tyranny. Indeed, rating is one of the main ways of monitoring and policing click workers. Perhaps you have already heard of those Uber drivers who were expelled from the platform because they had lost a star. This system, then, is generally directed towards workers. Here, the trick was to take this principle and apply it to rate employers instead. The turkers have also been able to choose their own rating parameters: today, they evaluate above all the speed of payment and the fairness of the proposed rate in relation to the job. They also make visible the average rate offered per job for each employer. Irani refers to this reappropriation as tactical quantification: “making employers accountable, while at the same time integrating themselves into the rhythms of Amazon Mechanical Turk. « Tactical quantification is a use of numbers not because they are more accurate, rational, or optimizable, but because they are partial, fast, and cheap – a way of making do in highly constrained circumstances.”⁹ While Irani herself doesn’t use the term tactical design in her work, she does claim a “design from below” as opposed to a design from above: “techniques from below show the circumventions and counter-practices of those who are subordinate to or subject to governance from above”¹⁰. She also places the tool in the filiation of tactical media art, which since the mid-1990s has sought to subvert the media through temporary interventions.

Turning the opponent’s strength against himself

The second example I would like to present is Consent-O-Matic¹¹, again a browser extension. It was developed by Midas Nouwens, an HCI researcher at the University of Aarhus, Denmark, with the help of Janus Bager Kristensen, Rolf Bagge and Clemens Nylandsted Klokmose. Consent-O-Matic is an extension designed to counter cookie banners. Here is how its creators present their project: “Cookie pop-ups are designed to be confusing and make you ’agree’ to be tracked. This add-on automatically answers consent pop-ups for you, so you can’t be manipulated. Set your preferences once, and let the technology do the rest! We are privacy researchers that got tired of seeing how companies violate the EU’s General Data Protection Regulation (GDPR). Because the organisations that enforce the GDPR do not have enough resources, we built this add-on to help them out. We looked at 680 pop-ups and combined their data processing purposes into 5 categories that you can toggle on or off.”¹²

In concrete terms, when we arrive on a page, Consent-O-Matic, if it recognizes the cookie banner in use, will fill it in for us, according to our preferences. The extension makes the cookie banner disappear by filling it in quickly enough to ensure that it doesn’t appear at all, or hardly. Its benefits are twofold: it allows you to systematically and automatically refuse cookies, and at the same time it makes the annoying cookie banners that hinder access to so many sites disappear.
To successfully counter these cookie banners, designers have seized a loophole and turned it to their advantage. Here, It is the centralization and quasi-monopolistic situation of the players to be countered that paradoxically makes them so fragile. If every website had its own homemade cookie banner, it would be almost impossible to counter them automatically. However, most sites have preferred to use generic banners offered by a handful of private players that emerged following the implementation of the RGPD. Although precise figures are hard to come by, in 2020 in the UK, almost 60% of the 10,000 most-visited websites used a cookie banner offered by 5 companies¹³. Paradoxically, this hegemony makes them more fragile, as a single, well-directed attack will be able to reach a very large proportion of websites. By focusing on these few dominant players, and hitting the right spot just once, it is possible to reach a large part of the Web. In terms of tactics, Consent-O-Matic’s particularity lies in the fact that it uses the adversary’s strength to turn it against itself, an approach that differs from Turkopticon, which seeks to improve the platform on which it insinuates itself. Here, the ambition is more explicitly antagonistic.

Nouwens and his colleagues call their approach “adversarial interoperability”¹⁴ after a term proposed by Cory Doctorow¹⁵. Interoperability is the ability of a system to communicate and function with other, sometimes technically different systems, and adversarial interoperability is the idea of interfacing with other systems without their permission or support. According to Doctorow, it is fundamental to the history of computing, and has enabled its vitality.

Diverting an opponent’s strength to a competitor’s advantage

An example of a relatively similar tactic is the Amazon Killer extension¹⁶ (yes, another browser extension, more on that later) created by Elliot Lepers, who defines himself as a designer-hacker working “to invent new forms of political practice”¹⁷. The idea behind the extension is also extremely simple: you go to Amazon and search for the book of your choice. Once on the book’s page, instead of the traditional button allowing you to buy on Amazon, another button appears, offering you to buy the book in a bookshop, and redirecting you on click to the book’s page on the Place des librairies site to do so. The extension is perfectly integrated into the interface, taking up Amazon’s graphic identity to such an extent that you quickly forget it’s not the original interface.

Here, the tactic is to use Amazon’s strength, i.e. the very high quality of its product search system (books in this case), to serve another competitor: the Place des librairies website, which lets you search for books available in bookshops around you, and whose search system is often less efficient. The extension also lets us redirect our reflexes, as we often end up on Amazon’s book page via links we have encountered on the Web, and then buy on the spot for the sake of convenience and speed. This extension is not the only one using this tactic of appropriating and hijacking the adversary’s strength, as can be seen, for example, with Peertubify¹⁸, a plugin that indicates via a banner at the bottom of the screen on the page of a YouTube video when it is also available on Peertube, a decentralized video viewing platform. Here, too, the tactic is to take advantage of the quality and comprehensiveness of YouTube’s search engine, hijacking it to serve another competing player.

Cleaning up the existing

A final tactic, perhaps the most widely used on a day-to-day basis, is that of cleaning, i.e. removing or hiding interface elements. The most emblematic and widespread example of this tactic is certainly the ad blocker. But while ad blockers are, as their name suggests, primarily focused on ads, others have taken the desire to clean up certain digital services a step further by cleaning up their interfaces. Often, the core service provided by digital platforms is perfectly satisfactory, but it is the various accessory functionalities that are particularly undesirable. Often added after the fact, as the platform is updated, they almost always help to attract attention and increase the time spent on the platform – when they are not outright dedicated to it. This is why video recommendations on YouTube or recommended posts exist on Instagram, to name just two of the best-known examples. Against these practices, a large number of extensions have flourished in recent years, such as Unhook¹⁹, an extension whose creators I’ve not been able to identify, but whose aim is to “remove all distractions” on YouTube to refocus the interface on watching the video. In practice, Unhook removes all recommendations, suggestions and even comments, leaving only the interface elements related to the video you’re watching. It also prevents videos from launching automatically, so you don’t get caught up in watching other videos.

Another example is the Demetricator²⁰ created by American artist Ben Grosser. For him, the fact that numerical indicators are omnipresent in the interfaces of digital services is leading to a transformation of our relationship with people and content. In response, he has developed an extension available for Instagram, Twitter and Facebook: the Demetricator, which makes all the numbers present on the interface disappear, i.e., in the case of Twitter, the number of likes, retweets and followers. While the tool’s primary ambition, according to its author, is to provoke questions about the effects of interfaces: “how does an interface that highlights a friend counter change our perception of friendship?”, it is nonetheless functional and usable on a daily basis.

The tactic employed by these various extensions is very simple: they hide certain parts of the interface, taking advantage of the possibility of overriding the style rules that govern the display and composition of web pages. Using extensions, it is relatively easy to make entire parts of a site’s interface disappear without any further ado, and without affecting the rest of the site or even the way it functions. However, this facility only concerns the display of pages, and it is often more complex to fight functionalities that are not directly visible or identified in the interface, such as algorithmic curation.

The browser, a tactical design enabler

While reading this article, you may have noticed that the projects I have presented all have one thing in common: they all use the browser as a starting point and technical support for their tactics, which is one of the main limitations of tactical design at a time when many uses have shifted to smartphone applications. The browser is a very special piece of software in terms of the powerful tools it makes available to a digital resistance. These tools are based on the decentralized and open nature of Internet protocols (TCP/IP) and the Web, which “shifts the capacity for innovation from the center to the periphery, and gives a great deal of power to the user”²¹. What other infrastructure gives end-users so much leeway in displaying the content they distribute? When we compare the browser with other software we use, the difference is striking: at best, the latter allows us to adjust a very limited number of parameters. This power exists because the browser allows very fine-grained control over the data it receives, thanks in particular to the extensions that can interact and transform web pages. This capacity for control, widely exploited by designers, is far from being the result of chance; it is inherited from the client/server paradigm and the technical constitution of the Web, which from the outset decoupled content from its formatting.

This decoupling can be seen very clearly in the history of the CSS language used to format the content of a web page, which is itself structured via the HTML language²². Indeed, in the early days of the web, CSS did not exist; graphic choices for content styling were determined directly by the browser, in other words, by the reader. The invention of CSS was an attempt to reconcile the tension between author and user in defining the styles to be applied to a web page: “the author of HTML documents has no influence over the presentation. Indeed, if conflicts arise the user should have the last word, but one should also allow the author to attach style hints.”²³. In the history of media, the Web thus represents a radical transfer of power to the end user. It is this power that has been relatively maintained (even if a large proportion of Web users are unaware of it) and that still lets us today benefit from an infrastructure – the browser – that lets designers create tools of resistance against hegemonic digital platforms.

The singularity and power of the browser is all the more obvious when you realize that outside this environment, tactical design practices are far less widespread, particularly on smartphones, where it is much more complicated to interfere with mobile applications²⁴. The application ecosystem is very different from the freedom afforded by the openness and standardization of the Web. Technically, each mobile operating system (iOS and Android, to name the main two players) has its own development language, even if the use of Web languages is possible and is becoming increasingly important. Above all, applications are tightly controlled by the hegemonic App Store and Google Play Store in the West. Applications go through a highly restrictive validation system, making it much more difficult to implement tactics discreetly.

Modeste Design

Taking a step back from the various approaches implemented by tactical design, one of the dimensions I would like to emphasize is its modesty, which must be set against its very narrow room for maneuver. Tactical design can only be modest in its ambition, because it only tackles problems that are always very circumscribed and specific, depending directly on the scale of the often highly localized solution it can provide. It is also directly linked to and dependent on the platform against which it operates, with no real claim to its own autonomy. As the aim is generally less to improve than to “make less worse”, tactics, like the tricks referred to by philosopher Georges Vignaux, are developed in a minimal-effort way²⁵.

This modesty is reflected in a generally distanced relationship with aesthetics, another common thread running through many of the projects I presented earlier. The cleaning tactics are particularly interesting, because their aesthetic ambition lies precisely in a work of subtraction that is often as discreet as possible: one must succeed in subtracting interface elements without disfiguring what remains of them. As for the other tactics, they most often seek to integrate and blend into the interface, to the point of making us forget that they exist and that the interface has been tampered with. To this end, they generally adopt the interface’s graphic identity, reusing CSS styles already in use on the page, always seeking a minimal-effort approach, in stark contrast to the traditionally accepted idea of a designer creating singular forms.

This obvious lack of interest in the visual appearance of tactical design productions seems to me to be very much at odds with some of the artistic practices that nevertheless share a common desire with tactical design, namely to seek to concretely counter existing oppressive systems. Perhaps the best-known of these artistic practices are make-up and masks designed to counter facial recognition. Despite similar ambitions, these works tend to be visible and photogenic, because they need to exist within an artistic regime shaped by the economy of attention²⁶. Is it because tactical design projects have a different place within this economy of attention that they remain so visually modest?

Individualized solutions

While aesthetics may separate them, a same criticism can be made regarding tactical design and artistic practices of technology resistance. It is well summed up by Os Keyes, in an article for reallifemag.com on anti-facial-recognition make-up, “[...] the popularity of ideas like it that imagine defeating the machines solely with personal interventions — serve mostly as a distraction from the kind of organized action that could better resist facial recognition.”²⁷

Indeed, all the projects I have introduced have to be installed individually, placing the responsibility for their protection on the people themselves. Knowing how to protect oneself behind ad blockers makes us forget that many other people don’t do it, not least because they don’t know how. Worse still, it sometimes even makes those who have protected themselves forget that it would probably be more beneficial to fight this nuisance on a systemic level. This problem is linked to another paradox raised by tactical design: the effectiveness of tactics often depends on their invisibility. The more the proposed tactics are used and shared, the more they can attract the attention of the system they seek to combat. Once alerted, it can usually easily modify its interface, making the tactic immediately ineffective.

This can lead to a tug-of-war, or rather a race to the bottom, like ad blockers that have to be constantly updated to adapt. Should you make your tool known, so that it reaches and is useful to as many people as possible, at the risk of being spotted and easily countered? On the contrary, should you keep your projects as discreet as possible, at the risk of seeing their reach severely diminished and reserved for a small community of insiders?

Another paradox raised by tactical design is the role it can play in legitimizing or accepting digital technologies. Don’t we reinforce dominant systems if we help make them more liveable and therefore more acceptable? In particular, they may seem out of step with other alternatives such as boycotting or even stopping using these oppressive systems. Are tactics just band-aids on wooden legs, where a much more systemic resistance is needed? Despite its success, the creators of Turkopticon recognize that its very existence perpetuates and legitimizes Amazon Mechanical Turk by helping to protect its workers. Amazon’s platform benefits indirectly from various initiatives by private players, such as Turkopticon, who add functional enhancements free of charge. Ideally, Irani would have liked Amazon to change its system to include more protections for workers. But this didn’t happen, and on the contrary, Turkopticon has become a piece of code that workers rely on in their daily lives, a tool that they themselves need to maintain.

Who can afford tactical design?

Another issue for tactical design is the sustainability of the tools developed within this framework. Although Turkopticon was born as a tactic, an opportunistic and circumscribed response, it has gradually become normalized, used on a daily basis by users who have come to depend on it. So what happens to tactics, which are by definition short-term practices? How do you switch from a regime of immediate, bricolage type of response to a tool maintained over the long term? Having to react and update your tactics every time the platform you are trying to counter is updated can be time-consuming and unrewarding. This question is closely related to that of maintenance in the context of open-source software. However, Turkopticon seems to have successfully negotiated this shift: the tool is now managed by a collective of turkers, the fruit of long-term work on the part of its authors. Turkopticon’s real victory is perhaps less the tool itself than the fact that it has been able to participate, however modestly, in the organization of turkers and their own forms of solidarity.

The question of the sustainability of an essentially non-commercial practice raises the question of which profiles can afford to engage in this approach. Beyond tactical design, this question arises for all critical designer practices, understood by designer Afonso de Matos as “This outlook perceives design as more than a market service: it aims to show that designers can be agents of powerful social and political change, beyond the boundaries of the commercial client-commission setting”²⁸. In this article for the newsletter Other Worlds, he asks who can afford critical design, and lashes out the lack of critical reflection on the socio-economic conditions that enable these projects to exist. The profile of designers practicing the unpaid activity of tactical design is very different from that of the average designer working in an agency or as a freelancer, and only a minority of designers can afford these practices: we note the strong presence of research-related profiles and designers claiming an activist practice.

In this sense, tactical design is still a sideline for these professionals. But where critical design has built up its own “closed market”²⁹, tactical design seems to have escaped it for the time being, mainly because it remains focused on projects that are above all functional, at the service of users. The question remains, however, of how to perpetuate and share these tactics, if it is even possible. One answer certainly lies in the greater inclusion of activist practices. Indeed, designers are far from being the only ones to develop extensions in an attempt to combat platforms. Many digital user advocacy organizations are doing the same. Researchers Karen Louise Smith and Elysia Guzik, who have sought to understand the place of browser extensions in these organizations’ work, show that they are a means of implementing a particular relationship to activism, and that they can become a useful tool in the arsenal of campaigns run by these organizations³⁰. By becoming more explicitly involved in activist approaches, tactical design could perhaps overcome the dichotomy between localized individual response and necessarily organized systemic activism.

From style-tools to folkloric fabulations (Added notes by Doriane)

At the start of Declarations research project, I stumble on this text in Tèque magazine. The first chapter of the research had for goal to develop different tools and protocols for exploring CSS poetic materiality and cultural dynamics. The idea was to approach CSS not as a practical technic that shapes (act) on a content but as the studied object of the research (style as content itself). And in order to do that, we needed tools that are not CSS but that can look at, inspect, interact, modfiy with the corpus of stylesheets existing already on the web. This text beautifully portrayed browser extensions as a new zone of agency to study CSS. They allow us to investigate the experience of the web, without being clearly on one side of the user-designer spectrum.

We started by making extension that collect and analyse web-design/stylesheets, to look into how the web is declared (by people and machines). Then by remixing it in a live manner, interjecting ourself in the middle of this design ecosystem to often simplified as an emetor-receptor dialogue. Showing the open nature of the web as a resilient force, even in a context where platforms authority are becoming the norms that limit the plurality of shapes & voices. By looking at CSS, Declarations takes this specific angle of web-design, not as modest/invisible tactical opportunities, but with all its various stories of web aesthetics-interactions-cultures. It's worth mentionning a bit of history. User-customisations used to be a definitive elements of spaces on the web - think of geocities, or early web handmade pages. Individuals and collectives where not only expressing themselves through "content" but through "forms" to take back this binary distinction of the W3C. Years later, blogging platforms such as myspace, skyblog, tumblr or PHP-powered forums where still strongly rooted in web-expression being linked to multiplicity of shapes, and providing various tools for CSS themes and customization, it is the time of CSS zen-garden. Now, in the era of instagram, twitter and tiktok platforms are homogenised and customisation are blocked off by corporations in place of their smooth brand identity. It's no longuer your "corner of the cyberspace" that you decorate like your room, but a service you use: an app. Even if you want to make your own website, design systems has been solved for you by experts: use squarespace or wix. They relie on similar hegemonist framework that inevitably produces seamless normative interfaces sold as design optimisation. The buying proposition of Figma by Adobe for the unthinkable sum of 20 billion dollars is only a hint at how owning the means of webstandards is a source of power. Technical decisions lead to cultural impact. We can only imagine how the choices of what/how we the practice CSS standard could have redefined our attention to various visual elements, its impacts on our lives expanding far outside of the screens space. But browser extensions reminds us that whenever we acces a page, it is being downloaded locally on our machine in a human readable way and through declarative language we can bend the fluid theater of what is materialising on our screen.

For matter of precision, it's worth mentionning that Declarations approach has of tactical design that "it can only exist in relation to the other. So it has to play with the terrain imposed on it, as organized by the law of a foreign force." But in it praxis it maybe shares more with critical or antagonist design. As an artistic research, Declarations doesn't pretend to counter specific actions, or to turn extensions into usable tools for the people, but as a mean to explore and retell. It plays with the idea of "adversarial interoperability" as a poetic and narrative strategy.

One of the first extension developed was Scratch. Scratch is named both after the practice of DJ scratching, and scratching the surface to reach the code of the webpage. Scratch allows to append any etherpad as a CSS stylesheet on every website at once. It is quite similar to Stylus, but with the added principale of live collaboration. By redesigning the web through an etherpad, changes can be viewed and updated live by all the extension users, creating a small collective altered perception of the web. Scrap, an extension and public server that document the can collect and curate the CSS on indivual object on website: buttons, searchbar, typography, ... Like a bag to collect objects during a walk? can we cut CSS from website like a scrapbook as we browse? can we collect and curate a library of designed elemen where the styles (CSS) are the content (What we read), while still taking with us the ambiguous textual-visual materiality of CSS? Words, an extension that inform you on the CSS words used on every website. How to make visible the power dynamics behind the CSS declarations that each websites uses, by situate/contextualizing the property that declare its design. What selector are used the most, what can the classname reveal, can it tell about the context in which is was written: now or in the past, by one amateur or by a corporation; when where the property used on a page implemented in the standard? and by who? Those interjecting style-tools doesn't work with CSS per say with javascript that act on the CSS, to study it.

Later, during the one week Declarations residency at Constant in April 2024, the wonder for web-extensions propagate to a group trans-disciplinary artists. weather uses a weather station API (Application Programing Interface) and connect it to your daily browsing experience. If there is wind outside the wind also blows on the element of any webpage. The weather extension is a poetic way of showing how websites are continuously changing, how they were never meant to be a fixed version of themselves. Designing for the web means embracing change and connectivity. focus act as both a perfomative process and a poem. This came out of a experimentation of trying to convert any websites to an interactive fictions, right after a discussions on web accessibility and the accessibility tree. The idea of telling a website one element at a time by rendering everything else invisible, is like the chronological process of a screen-reader, but become also a performative telling, a new interaction with an interface. The poem goes as follow. The idea that such few lines of codes can change the entierty of the web made it feel like holding declarative magic power.

/* make everything invisible */
*{
    opacity: 0;
    transition: 1s opacity;
}
/* only show what has focus */
*:focus,
*:has(:focus){
    opcaity: 1;
}
/* on any webpage, you can press TAB 
to focus the next element.
we're telling a webpage one element at a time.*/

While the tool approach persist, it also becames something else than what we could qualify as tools. There is as many possibility to reinterpret the web as there are sensibilities to interface; and maybe we can use this liminal space of agency to tell new stories about the web, stories that can't be tell only from the making of a website, or from using one. Declarations, inspired by those tiny hyper-situated poetical extension, took this idea of Web Fabulations as a new thing. Explored at the Hacker and Designer summercamp in Jully 2024, and then followed by a serie of workshops in art schools and collectives. From a thinking of inspired by declarative languages, I would like to simply list some of those Websites fabulations as folkoric tales of the web instead of practical tools. Not something to scale or improve to a large amount of user, but more as doors to stories.

• 🌐 distant: website shrink relatively to server distance
• 🌸 mindfull: bluring if moving too fast, flowers grows as time passes
• 🎲 unordered: unordered list properly unordered (mess, dancing, randomised)
• 🪞 fractured: every link divide the page in two (previous, current) recursively, showing every timeline at once
• 😼 cat: a cat tries to reach your cursor, if it does it crash
• 🦷 biting: on each page one link is trapped, clicking it make the website bite you (and close the tab)
• 💭 remembered: websites disapears as they are visited but you get the contact of someone that was there
• ⏳ aging: websites rots, or ages, according to domain name registration date
• 🍪 bloated: when eating too much cookies your should wait to digest them

or some localised

• instagram but alt text (AI written with bits of human written hashtag)
• toys vers kfine
• ebay as hype designer bw shops

If Declarations is a research of how declarative languages shape the web, an important lessons was learned here. Tactical browser extension often are tactical to one website, but the exercice here was slightly different: we are trying to be tactical to every websites at once. One of the propose exercice was to sit in a circle and ask, each at a turn, "what if the web was ... (liquid, dark, intimate, noisy, ...)". When introduce this seems like an impossible task: redesigning the entierty of the web? However, even when written by complex intricate frameworks, the web conserve a semantical aspect. As the example of the Focus extension shows, with the craft of precise selectors give CSS an unexpected declarative power that can manifest cross-websites impacts. As web-designers we are formed to think about a one-to-one design-project relation, but declarativeness breaks through that unicity and could be seen as a tactical choice as a technology to address the vast ever-changing landscape that is the web. In her video essay Why Is CSS So Weird? (2019), Miriam Suzanne presents how Declarativeness on the web uses the flexibility of natural language to design for "an inherently problematic medium" and "a multitude of unknown canvases". (more on that)

The text Tactical design also phrases an important friction. From the experiment done with the Scratch extension, when 12 people are constantly updating the styles of the whole web on an etherpad, it quickly becomes chaotic and overwhelming. When we redesign the web through a shared extension, we realize that the challenge is more about ourselves as a group of individuals than about programming technicalities. On a political level, to fork from the mainstream perception of the web, and to create our own alternative bubble can be seen like a co-housing situation: it is for sure not scalable, and talk about our usage of the web as group. To handle our common space we need some structure, some social protocols, we need to learn about each others sensibilities. More importantly, even if we manage to build a solid, resilient, alternative perception of the web, it will be a lived fiction, as it isn't a reality of most people but only for the specific subset of user participating in the experiment. In the same sense that using addblock make us forget that adds exists, because we don't see them anymore, the more we live outside of society the less we feel the need to fight against its wrongs. In that sense, pushing the open standard, culture and regulation to change in a specific direction is something else entierly. The openess of the web can also be used to shift the responsability of making it a nice place away from the tech giant corporation and to individuals.

I am immensely thankfull to Nolwenn for allowing this re-publication of the article, and the effort she made in translating her own text. I am happy to have the opportunity to interlace this artistic research with its reference in such a circular way, letting the text shapes our practices and our practices shapes the text.