By 2018, I had become comfortable enough with Linux, Windows and macOS that choosing between them no longer seemed worth another round of preparation. I could take the setup in front of me and begin, leaving its wallpaper and most of its preferences as they were. There would still be things to fix and occasions to change something, but I no longer felt that I had to arrange the whole computer around myself before using it.
That ease becomes more interesting to me alongside the things I am still learning to handle. I began juggling with balls, round and well balanced, small enough to fit my hands, then moved on to clubs, which took more effort, and later to rings and hoops. Five balls are now comfortable enough for sustained, relaxed juggling; I can also juggle five rings and manage short runs with five clubs. Poi have been a much longer acquaintance. Learning the basics took more than five years, and juggling four or five of them still comes with considerable difficulty.
Years of practice and a decision to leave the computer alone can look like opposing attitudes toward effort. They make sense together through what I want to spend my time doing. With the props, learning to handle the objects is part of the activity I have chosen; with the computer, arranging the environment can delay the work I wanted it for. Playing Computer, Playing Music approaches a related question through juggling and electronic music, asking how intention and bodily effort participate in an outcome mediated by a tool. Here, being able to continue comfortably matters more to me than how closely the object expresses my preferences.
I use generic for things that can be reproduced at scale at a cost that makes them widely available. Many people can obtain substantially the same browser, editor or manufactured ball and put it to different uses. This resemblance gives them something in common before they begin, while leaving open what each person will be able to do with it. A factory can produce a thousand similar balls without reproducing the ability to juggle five of them, and my present comfort says as much about my relationship with the form as it does about the form itself.
Even the apparently simple description “hand-sized” depends on whose hand will receive the object. An ideal sphere offers a useful geometric tolerance for an ordinary toss catch, since rotation leaves its surface the same and there is no designated handle that must arrive first. An actual ball also has weight, filling, seams and friction, all of which enter the encounter. Its recognizable shape is a beginning from which people with different hands and histories have to find their way.
With clubs and rings, some of what shape leaves for the performer becomes easier to see. The Cirque du Soleil and National Circus School training manual teaches control of a club's rotation so it can be caught by the handle, while its ring lessons address rotation and wobble. A ring can look as simple as a ball and still present a plane whose tilt matters to the next catch. The shared impression of roundness leaves these differences in handling unresolved. Juggling training manual
The same component can also enter a substantially different arrangement. Flowtoys documents converting its coral props between balls and poi, retaining a common part while changing the assembly with a leash and handle. Here, familiarity can survive in the component while considerable learning remains in the combination. That possibility helps me describe my experience with poi, although the time it took me belongs to my own learning history and offers someone else no timetable. Coral balls and poi guide
Juggling is the clearest representation I know of an activity in which customization is expensive and commitment to a default becomes bodily. A menu can change an editor's typeface without changing the computer; a different club has to be designed, made, bought and carried, then learned in flight. The juggler supplies much of the adaptation. With a matched set already in the bag, altering timing and grip is usually closer than commissioning another mould or building another object.
David Cain's history of the juggling club begins before that bag contained anything standard. Late nineteenth-century performers used exercise clubs or made their own. Edward Van Wyck's hollow clubs drew requests from other jugglers, and in 1895 he left performance for a retail business run from a Cincinnati shop. The subsequent record runs through wood, wicker, cork, fibreglass, foam, plastic and composite construction, with particular performers and makers altering weight, resilience, decoration and balance. A niche practice was already a technology culture, although much of its development lived inside individual acts and workshops. The Development of the Juggling Club, part 1
The market widened while remaining personal. An IJA profile from 1984 described roughly a dozen prop companies and called the trade a small business shaped by the people who ran it. Dave Finnigan's inexpensive Juggle Bug club had nevertheless reached stores across the United States and distributors in eleven countries. A later account says the popularity of juggling rose sharply during the 1980s, drawing Brian Dube, Todd Smith and Renegade into competition over quality and price. Workshop, catalogue and mass retail continued to overlap. Juggle Bug profile, 1984 Club history, 1990
Jay Gilligan describes what that overlap does to technique. After nineteen years with the same kind of plastic ring, he had asked little about its diameter, thickness, weight or material; colour was the practical variation he and his friends discussed. In his account, the familiar ring traced back to a circle drawn around a cookie tin in a Taiwanese factory in 1976. Repetition had turned a convenient factory template into the normal ring against which later sizes were named. The Evolution of Juggling
Smaller and larger rings then produced combinations whose timing depended on their unequal sizes. Gilligan's Manipulation Research Laboratory and work with Renegade extended the question into triangles, linked rings and clubs with open bodies. The inherited prop supported existing tricks and quietly described the field in which tricks were sought. A new object changed that field, while demanding another round of practice and production.
Materials made for larger trades repeatedly enter this small one. Mass-produced tennis balls can be used directly or cut and filled for practice; the tennis industry has already supplied the hollow rubber shell, felt, consistency and distribution. Cain records a 1962 practice club made by inserting a fourteen-inch aluminium pipe into a rubber ball, and current prop builders combine standard aluminium tube with machined connectors and 3D-printed heads. Plastics support another range of bodies, knobs, handles and shells. Each neighbouring industry lowers one threshold while leaving a juggler or small maker to decide what the borrowed part should become. Tennis-ball manufacture Filled tennis balls in IJA teaching Cain on aluminium-pipe clubs A current tube-and-print construction
The present size of the juggling-prop market remains uncounted in the public sources I found. Its variety is visible in what makers currently offer. Henrys manufactures injection-moulded parts and blow-moulded club bodies in Karlsruhe; Renegade assembles clubs to order from selectable bodies, handles, caps, knobs and parts made by several manufacturers; Flowtoys builds interchangeable tubes, connectors and light units into poi, staffs, clubs and other arrangements. Standard components and unusual experiments occupy the same shelves. Henrys company history Renegade custom and hybrid clubs Flowtoys modular system
Electronics make some props hybrid instruments. Current LED clubs and balls can hold programmable modes, motion sensing and wireless connections, with an application coordinating a show. An Adafruit project goes further toward the workbench: its files describe printable shells, lights, batteries and microcontrollers for wireless juggling balls. At this edge the physical prop has firmware and a controller, yet impact resistance, weight and the price of making another set remain stubbornly material. Flowtoys LED juggling props Adafruit wireless juggling balls
Photography moves the same negotiation into a system of replaceable parts. Film stock, lens, filter, flash and development process let a photographer change the instrument and the image at several points. Leica sold a screw-mount camera with three interchangeable lenses in 1930, and Nikon's F mount joined successive bodies to a growing lens collection from 1959 onward. Compatibility turns a purchase into a path through later choices, while every lens still has to be made, carried and learned. Leica camera history Nikon F-mount history
Photography has also carried a fixed-lens mass market alongside those systems. Kodak's inexpensive Instamatic simplified loading through a film cartridge and sold more than fifty million cameras between 1963 and 1970. The Ricoh GR line has kept a compact body and fixed wide-angle lens since its film version appeared in 1996, even as firmware and controls developed around it. These cameras concentrate decisions in the manufacturer so the photographer can carry less and begin sooner. The history keeps returning to the trade between an open mount and a settled optical package. Kodak Instamatic history Ricoh GR history
One branch of electronic music places the choices directly on the surface. An integrated instrument settles most of its signal path inside one enclosure; in a modular synthesizer, oscillators, filters, amplifiers, envelopes and sequencers can be selected separately and connected with patch cables. Doepfer began its A-100 system in 1995; the company now describes the Eurorack format that followed as an ecosystem of several hundred suppliers and several thousand modules. Its mechanical and electrical conventions allow modules from different workshops to share a case, power supply and control signals. Doepfer A-100 history and compatibility notes A-100 technical notes
That modularity reaches down to electronic components. Doepfer publishes introductions to resistors, amplifiers and basic circuits as well as modifications and low-cost case construction; open-hardware makers publish schematics, circuit-board files and bills of materials. Someone can buy a complete voice, solder a kit, alter a circuit or begin on a breadboard. The freedom is substantial, and so are the cost, electrical knowledge and attention involved. For a modular-synth player, arranging the environment may itself be the music-making chosen for the evening. Doepfer A-100 DIY Greenface Labs open synthesizer designs
For my purposes, computing shortens the material distance between a default and a variation more radically than the other examples here. A program can be copied and changed without manufacturing another physical instance, and one computer can host tools for writing, drawing, recording, simulation and control. Free software makes this possibility explicit in the rights to run, study, modify and redistribute code; extensible programs also offer narrower interfaces for settings and plug-ins. The access is uneven because skill, time, hardware, licences and hosted services still decide which changes a person can actually make. Within those limits, software gives do-it-yourself work an extraordinary range and a very low material threshold. GNU free-software definition Visual Studio Code extension capabilities
Software that feels ordinary to me has undergone a similar encounter with practice. Some of its ease comes from decisions made by its designers, and some comes from knowing my way around the environment well enough to give its peculiarities little attention. From the comfortable position of saying that it works, I cannot neatly divide the credit. Accepting what is already there may be something I have learned to do.
Paul Dourish's account of appropriation gives a useful name to the adaptation that happens through use. People incorporate technology into their practices, sometimes repurposing it without explicitly changing its configuration. An editor can keep its shipped settings while becoming part of someone's particular way of writing, organizing material or building a project. The visible setup records very little of this relationship, which makes an unchanged desktop a poor measure of how personally someone uses a computer. The Appropriation of Interactive Technologies
This also makes me less willing to separate appearance cleanly from practical use. Donald Norman's discussion of interface conventions describes the learned expectations carried by familiar cursors and scrolling behavior. A recognizable arrangement can help someone act because they already understand what it suggests. What looks generic on the screen may preserve knowledge acquired elsewhere, allowing a person to continue with less preparation. Affordance, conventions, and design
Another person's visual adjustments may address difficulties I barely notice. Text spacing can affect access to reading, and the W3C's guidance describes preserving content and functionality when readers change it to suit their needs. From my position, the change might seem cosmetic; for the person making it, the page may become usable. Comfort, preference and function meet in such adjustments, and enjoyment supplies a reason of its own. Someone can spend an evening choosing a theme without having to account for it as a future gain in productivity. W3C text-spacing guidance
My initial thought was that generic software contains a funny contradiction: as it becomes better and more configurable, there seems to be less practical reason to configure it. What I can defend from my own experience is more specific. A default that fits my work leaves fewer problems I want to solve by changing it, while a better configuration tool may make a worthwhile adjustment easier. Both improvements can happen together. Whether I use the new possibility depends on the difference it would make to what I am doing and the effort of obtaining that difference.
Eric von Hippel describes a closely related arrangement in Democratizing Innovation, where a default serves the user while a toolkit remains unused until there is sufficient incentive to employ it. The possibility can stay in the background for a long time without disappearing from the relationship. I may leave the settings alone partly because changing them remains possible if my needs change, or if I encounter something I can no longer comfortably work around. Democratizing Innovation, pages 162–163
Consider that same untouched configuration on two workplace computers. One person accepts it because it suits the work and remains free to alter it; another is prohibited from making changes. Their desktops could look identical, even though the next obstacle would meet them with different options. Central administration can preserve personal settings too, as Citrix's Profile Management documentation describes, so the question extends beyond whether the machines were deployed from the same setup. Someone still decides which differences the organization will support. Citrix Profile Management
A restriction might remove an awkward configuration from the help desk's responsibilities while leaving the person who needed it to spend more time working around its absence. That added effort belongs beside the administrative saving when we call the standard setup cheaper. Even formal permission to change a setting may offer little help to someone without instructions, assistance or a way to recover from a mistake. A person who never intends to modify source code could still benefit from being able to ask somebody else to do it.
This is what I mean by freedom as maintained optionality. Keeping a useful change within reach involves people preserving working interfaces, usable exports and recovery procedures, as well as deciding which older arrangements can continue to be supported. Second Order Freedom extends this concern to the alternatives that collective choices leave for someone arriving later, whose needs may differ from those of the people choosing today. The cost depends on the possibilities being preserved and who needs them. Counting menu entries tells us little about whether a person can make the one change that matters, and supporting every past configuration indefinitely could itself prevent useful work.
Boris Fontanov's Holding Topology offers a way to think about continuation through the physical act of release. An object can leave direct bodily retention while remaining part of an intended sequence, with a later catch or other encounter still available. The model abstracts across props by recording this relation, leaving much of their specific handling outside its description. My differing ease with balls, clubs and poi belongs among those omitted details, where similar patterns of holding and releasing can demand very different timing, coordination and endurance. Holding Topology
Release itself can require sustained attention to what happens next. The connection I find useful for software concerns being able to leave direct intervention while retaining a workable continuation. I can use a default until an obstacle appears, then make an adjustment, seek help or move the work elsewhere. Leaving the configuration alone becomes a more comfortable choice when those routes remain within reach, even if I go a long time without using them.
Generated software adds another place for this question to arise. A shared system could produce a different tool for each person, preserving generic production while varying the result. The affordability of that arrangement would depend on what happens after generation, through use, correction and eventual replacement. A tool made specifically for me might fit the immediate task while still leaving me dependent on somebody else's service for the next change.
It would also contain decisions I had left unspecified. If I asked for an application to organize something, I might leave its ordering, labels or storage behavior open, requiring the system to settle those choices or ask me. Whatever it supplied would become another set of defaults, now reached through an interpretation of my request. I would still want to know what I could revise, how I could retain my material and whether I could continue using the result if the original service became unavailable.
There is also value in making a familiar kind of tool again when the second making is used as an inquiry. An in-house application turns a broad requirement into permissions, imports, error messages and recovery paths that colleagues must pass through to finish actual work. Each obstruction forces a choice about what the system does and an account of what another person needs to repeat it. Use gives documentation concrete facts to name: the accepted file, the permitted role, the observed failure and the way back. Those facts remain bounded by that organization and workload.
Generic technology can grow from this movement through a particular application. Rails' origin account says the framework was extracted from Basecamp after the application came first; Rails 5.2 later introduced Active Storage and Redis Cache Store as further extractions from production use. In each case, the particular system supplied a place where the common component had already been exercised. Re-doing familiar work inside a new application can therefore produce reusable technology together with an account of the situations that shaped it. The story behind Rails Rails 5.2
Implementation can correct the document as well as the code. During WCAG 2.0's Candidate Recommendation phase, implementation experience led the working group to clarify guideline language and add supplemental information when implementers misunderstood it. The report also drew on varied sites, including sites created without close help from the working group and public “in the wild” examples. An in-house application supplies a first ground of observed use; other implementations reveal how far its documentation travels. WCAG 2.0 Implementation Report
Continuing development introduces the further demand of preserving what already works. In its March 2026 evaluation of coding agents, the EvoClaw preprint followed dependent development milestones and found that agents could keep implementing features while regressions accumulated, making later work more difficult. This evidence is bounded by the systems and benchmark studied, leaving the broader economics of personal software open. It nevertheless gives a concrete reason to distinguish a working result from an arrangement someone can sustain. EvoClaw
That distinction is already familiar to me from a short run with clubs. Success has happened, while comfortable continuation remains another achievement. The account of generated commit messages in When Reading Costs More Than Writing gives the cost of assessment a concrete setting: reading a supplied draft can take longer than composing a replacement. With a generated tool, I may likewise save effort on making a change while still having to assess the result, explain corrections and learn what moved. Those demands may become easier too, but a tool that repeatedly adjusts itself to fit the immediate request could also ask me to relearn it. I would have to follow the whole activity to see where my attention was going.
My willingness to accept a default belongs to this larger experience of handling things. I can leave much of a computer's configuration as I find it because I can already use it for what I want to do, with the possibility of intervening when that stops being true. With poi, I am still giving time to the encounter itself, learning what I can sustain with the objects I have.