An artifact is a human-designed thing. When I use this word, I can refer to an mobile phone interface or a post office or a church or a law or a car. Artifacts have an intended purpose, even if it’s decorative or the result of a process that was the creator’s actual motivation. Artifacts may also have an incidental purpose dictated by the user and unanticipated by the creator. Usability refers to how seamlessly an artifact allows a user to interact with it, towards the user’s intended goal. Some artifacts are designed to be usable by more than one different type of user, or multiple user groups trying to accomplish different goals. And many artifacts are not designed with general usability in mind at all.

This post is mostly about how things that are patched together can become corrosive and dangerous.


Introducing Kludge

The definition of kludge:

The dictionary tells us that a kludge is “an ill-assorted collection of parts assembled to fulfill a particular purpose…a clumsy but temporarily effective solution to a particular fault or problem.” The term comes out of the world of computer programming, where a kludge is an inelegant patch put in place to be backward compatible with the rest of a system. When you add up enough kludges, you get a very complicated program, one that is hard to understand and subject to crashes.

Kludge can happen largely for two reasons: because an environment is difficult or because the agents responsible don’t give a damn.

Difficult Environment:

  • Insufficient resources (funds, appropriate building material)
  • Insufficient vision (ex. a small application that works fine but doesn’t scale well)
  • Wrong vision (artifact is built for one thing, used for another)
  • Multiple competing interests creating social/development friction.

Bad Builders:

  • Incompetence (perhaps the builders are not equipped to provide an adequate solution)
  • Insufficient knowledge (builders don’t know what their users want)
  • The cost of artifact failure isn’t relevant to the builder (ex. merely fulfilling a contract to the letter. without respect to success of the product the contract is related to; or perhaps the artifact is the only solution in town and so fail rate is almost unimportant to keep market share)
  • The artifact is a quick fix and not meant to be a permanent solution (ex. spending too much effort creating the artifact hurts the builder)
  • Feature Creep: Artifact must be all things to all people. Stupidly large, unfocused scope.
  • Exception Creep: Other interests are creating pressure to create loopholes, backdoors, or other exception-creating elements.



On bureaucracy and paperwork:

The hope of some of the French revolutionaries was that paperwork would rationalize the state, that it would depersonalize power and destroy the corrupt networks of aristocratic influence. Kafka quotes from a 1791 French administrative directory that advised that “letters of recommendation will be perfectly useless” in petitioning the government and “might even become dangerous, in that they can foster the belief that one is soliciting a favor or a grace that one does not have the right to obtain through justice.” As Kafka puts it, “A world of privilege was becoming a world of rights; the personal state was becoming the personnel state.”

Paperwork was also to be the means for allowing all of a nation’s people to scrutinize government activities, an intention enshrined in Article 15 of the Declaration of the Rights of Man and Citizen: “Society has the right to ask all public agents to give an accounting of their administration.” This mirrors the contemporary enthusiasm for open government and transparency among some activists and is the apparent raison d’etre for WikiLeaks. The idea was taken to astounding (and absurd) lengths by the Jacobins, who mandated that “all relations between all public functionaries can no longer take place except in writing.”

Funnily, bureaucracy is now popularly understood to be antithetical to both the “impersonal” argument and the “clarity” argument: government bureaucracy can be influenced by powerful interests to protect them through obscure, messy, unreadable proclamation. Forces with the capital to research and innovate in legislation can be rewarded for understanding and drilling into obscure legal topics. There is interest in broadly demystifying some of these bizarre systems but there is a great deal more influence and capital being put into very deep, narrow dives into kludge to protect their influence. TurboTax quietly lobbies heavily against simplifying the tax code and removing public need for their product. Car dealerships are turning to state governments to fight Tesla Motors, which is not using dealerships; sharing-economy startups are finding themselves fighting against legal assaults from the traditional businesses (hotels, for instance) who would lose out if people could rent rooms/houses directly to each other; the list goes on much, much further. Bureaucratic “transparency” does not necessarily help if we can’t even tell what we’re looking at.

To be clear, bureaucracy is just another organizational structure with pros and cons. Singular, authoritarian structures and ad-hoc, technocratic structures also have downfalls that five minutes with a pen and paper could predict. And overwhelmingly, most policy complexity in the United States is not the result of “bad builders” but of “difficult environments”.

A superficial analysis would predict that this proliferation of veto points [in the American system] would lead to inaction, generating a systematic libertarian bias. But in practice, every veto point functions more like a toll booth, with the toll-taker able to extract a price in exchange for his or her willingness to allow legislation to keep moving. Most obviously, the toll-taker gets to gobble up porkbarrel projects for an individual district in exchange for letting legislation move onto the next step. This increases the cost of legislation, but as John Ellwood and Eric Patashnik argue, it might be a reasonable price to pay for greasing the wheels of a very complicated legislative machine. […]

The multiplication of veto points, therefore, does not necessarily stop legislation from happening, but it does considerably raise its cost and, more importantly, its complexity.


Kludge seems to develop fairly easily these days. I decided to write about it because both what Kludge is and how it’s made seems to be a point of contention and skepticism in our cultural outlook- and I don’t just mean national politics. Kludge is a smattering of half-obscured facts that build conspiracy theories. Kludge is the result of processes designed from a fear of losing autonomy that, ironically, rob us of autonomy. It’s a bunch of quick fixes that allow broken systems to proliferate. I think it’s a great word for a big problem, a problem that will only get bigger with time.

Instead of optimism, the specter of paperwork permits cynicism to flourish. Privilege may temporarily disappear into the meticulous procedures of paperwork, which become recognizable rituals of impartiality, even if no one is satisfied with their actual performance. Anyone who has ever visited the DMV has taken part in this grand democratization of frustration. In a state where the DMV is the model institution, everyone is equal in that they are equally miserable. But paperwork also opens new avenues for the exercise of influence that are just as opaque as any earlier systems abused by elites. As documentation proliferates, so too do auditors auditing the clerks, and auditors auditing those auditors, and on and on to theoretical infinity. This network of data and overtaxed inspectors and processors has the effect of creating a miasma of competing claims for legitimacy, as well as ample opportunity for doling out preferential treatment, circumventing the law, subverting authority, serving oneself. Information becomes obfuscation, particularly under the pressures of “surveillance and acceleration,” which Kafka isolates as the contradictory demands of state power. The state needs to know more to function fairly, but with more information comes more urgency to process it all, yielding even more information to process and sending fairness further over the horizon.


Kludge will become more and more problematic as we start trying to patch together ways of dealing with the rapidly rising deluge of data spilling all the time, from everywhere at once. Before the 21st century, people often starved for information. As the 21st century continues, though, we will all drown in it. I imagine that a lot of problems will be solved in odd patchwork quilts and buried away to be exploited or to cause unanticipated problems not terribly far down the line.