How Solitaire Changed Windows (And Vice Versa)

Before Windows, playing a card game on a computer was an uncommon thing to do. There were card games on mainframes, on university timesharing systems, and on a handful of early personal computers. They existed as curiosities, written by hobbyists, distributed on tape or 5.25-inch floppy, and played by whoever happened to know they were there. Most office workers who used computers never touched one.

After Windows, card games on computers were everywhere. By the mid-1990s, Solitaire was installed on more machines than any commercial piece of software in history, simply because it shipped in the base install of every Windows PC. A receptionist, a lawyer, a factory supervisor, and a twelve year old could all describe the green felt background and the bouncing cards at the end of a winning game. The card game had become part of the operating system’s personality, and in a less obvious way the operating system had become part of the card game’s personality too.

This piece is a history of that exchange. We look at the people who built Windows Solitaire and FreeCell, the design decisions that made them stick, the cultural meanings the games acquired once they were in the water supply of the American office, and the long slow rearrangement that has taken solitaire from a Windows accessory into a browser habit, a mobile app, a subscription product, and back again.

The first computer card games lived on mainframes. IBM timesharing systems had text-based Blackjack and Poker clones by the early 1970s. PLATO, the University of Illinois teaching system that ran from the 1960s through the 1980s, picked up a broad library of card games during its long lifespan — and that library is where our story really begins. PLATO was one of the first networked computer environments that took graphics seriously, and the combination of graphics, a large user base, and a culture of open authorship made it a fertile place for hobbyist games.

Paul Alfille, a medical student at the University of Illinois, wrote a PLATO implementation of FreeCell in 1978. The program was unusual in two ways. First, it ran in graphical rather than text mode, which meant that the tableau looked like an actual patience layout rather than a terminal transcript. Second, Alfille chose a rule set that produced solvable deals almost every time, distinct from the Klondike-style games whose random shuffles leave the player blocked on a large fraction of hands. The design choice was deliberate: Alfille wanted a game that rewarded thinking rather than luck. PLATO FreeCell was used by students and staff across the Illinois system for years. It did not become culturally famous in its own right, but it seeded two generations of programmers who would later carry the rules to other platforms.

On the early personal computers of the 1980s — Apple II, TRS-80, Commodore 64, early IBM PCs — card games existed but were scattered. Hobbyist BASIC listings in computer magazines, shareware disks, and commercial cartridges each carried a handful. There was no canonical card game on any of those platforms. Solitaire was one option among many, and most users did not keep any card game installed. That pattern held through the rest of the decade. The operating systems of the 1980s came with utilities, not pastimes.

Wes Cherry was a Microsoft intern in 1988 when he wrote the version of Klondike that would ship with Windows 3.0 in 1990. The program was a side project, written outside of his assigned work, and it survived to production because the Windows team needed a simple, visually attractive application to teach users how the new mouse interface worked. Dragging a card from one column to another was exactly the sort of demonstration the Windows team was looking for: it required fine motor control, it exercised the click-and-drag metaphor, and it was forgiving of error because a misdrop just returned the card to its previous home.

Cherry’s own account, offered in interviews over the years, places the game’s origin in personal boredom and the availability of the Windows SDK. He has described writing it as a way to learn the graphics API, with no expectation that it would ship broadly. The decision to include it in the Windows 3.0 retail package was made by product managers who recognized it as a gentle on-ramp for a user interface model that was, for most of the PC-using public, new. That pragmatic choice — ship the card game as a mouse tutorial — is the single most consequential decision in the cultural history of solitaire.

The code itself was small: a few thousand lines of C, driving a handful of bitmaps, a simple deck shuffle, and the rules of Klondike draw-three with a single redeal. The implementation was faithful to common paper Klondike rules, with the conventions that were already familiar from the Hoyle guides of the 1970s and 1980s. It did not introduce variants. It did not customize the deck. It simply ran the card game that most American households already played on their kitchen tables, rendered in the same crisp style that Windows 3.0 applied to everything else.

The game shipped without credit on the splash screen — Microsoft did not advertise Cherry as the author, and for years the program’s authorship was a minor piece of Windows folklore rather than common knowledge. Cherry himself left Microsoft in the early 1990s to pursue unrelated work. His name now appears in most histories of Windows, but the credit came late and came from the press rather than from the employer.

The card faces that shipped with Windows Solitaire were drawn by Susan Kare, whose design work had already made her famous in graphical computing. Kare had built the original Macintosh icons at Apple in the early 1980s, including the system font, the command key, and the trash can. By the late 1980s she was taking contract commissions, and Microsoft hired her to produce artwork for several Windows 3.0 applications. Her hand is visible in the Card file, the icons of the Accessories group, and the full fifty-two card deck that rendered in Solitaire.

The deck Kare produced for Windows was pixel-dense by the standards of the time. Each face card was hand-drawn at a specific resolution, reusing the silhouettes of paper Bicycle cards but rendered in the flat, readable style that Kare had perfected at Apple. The suits were crisp. The pips lined up. The back-of-card patterns, chosen from a menu, gave players a small measure of customization: checkered, dots, robot, castle, island, and several others. Most players do not remember the designer’s name, but they remember the cards.

What made the card art culturally sticky was its combination of fidelity and distinction. The face cards were clearly playing cards — a player who had never touched a mouse could still identify the King of Spades — but they were also clearly Kare’s, in a way that shaped a generation’s mental image of what a digital card looked like. Later versions of Windows kept the deck, occasionally modernizing it but always preserving the basic visual vocabulary. When the Microsoft Solitaire Collection refreshed the art in 2012, player complaints were immediate: the cards no longer looked like Solitaire. That complaint is really about losing Kare’s visual grammar.

Jim Horne was the Microsoft engineer who ported FreeCell from its PLATO origins to Windows. Horne had played Alfille’s original FreeCell during his own years at university and remembered it well enough to reconstruct the rules and layout on a Windows target. The port appeared first in the Microsoft Entertainment Pack and then as part of the Win32s subsystem bundled with Windows for Workgroups 3.11. Later, FreeCell shipped in the base install of Windows 95 and stayed in the base install through Windows 7.

Horne’s most important technical contribution was the deal-numbering system. Rather than shuffling a deck freshly on each game, he seeded a linear congruential pseudorandom generator with the deal number the user selected, then drew cards from that generator to build the tableau. The result was a deterministic mapping between deal numbers and deal layouts: deal #1 on one machine produced the same cards in the same columns as deal #1 on any other. The choice to expose deal numbers in the user interface — with a dialog that accepted numbers from 1 to 32,000 — turned FreeCell into a shared puzzle across the entire Windows installed base.

That shared-puzzle dimension is what distinguishes FreeCell from Solitaire in the cultural record. Where Solitaire games were one-offs, pulled from a private shuffle that never mattered to anyone else, FreeCell deals could be discussed by number. Players called each other to say “try deal 617, it’s a monster.” Office workers compared wins on specific deals the way kitchen-table chess players compare positions. Magazines ran columns on famous deal numbers. The deal-numbering system created the first piece of common card-game folklore to emerge natively from the digital era.

The choice to cap the deal range at 32,000 was arbitrary, a byproduct of the 16-bit integer arithmetic in the original shuffle. Later ports extended the range (often to 1 million or beyond), but the original cap shaped expectations. Players knew the Microsoft set was finite and catalogable, which in turn created space for the Internet FreeCell Project and the decades of solver work that grew around it. The conceit that a card game could have a complete, shared, exhaustible library of legitimate positions is Horne’s invention.

Horne has discussed the port in interviews as a project undertaken partly out of affection for Alfille’s original and partly to fill the Windows Entertainment Pack lineup. Like Cherry, he has been generous about sharing credit and has consistently described the work as low-stakes at the time. That understatement is probably accurate; neither engineer appears to have expected the programs to become the most-installed games in computing history.

A useful footnote: the 32,000 deal cap has occasionally been presented as arbitrary folklore, but Horne has explained in interviews that it simply reflected the range of a signed 16-bit integer used by the deal-entry dialog. Changing the cap would have required updating the shuffle code, the input parsing, and the stored-game format in the Windows Registry. At the time, nobody at Microsoft considered 32,000 deals insufficient. The number was large enough that no player would exhaust it, which turned out to be half right: the player base as a whole did exhaust it, through coordinated community effort, a decade after release.

Once Windows Solitaire was on every corporate desktop, the discourse about it changed. Journalists calculated hypothetical productivity costs. Consultancies floated figures in the billions. Internal IT departments weighed whether to delete the executable from corporate images. The estimates were mostly hand-waved. Research on the actual workplace impact of Solitaire was limited, methodologically shaky, and often motivated by editorial angles rather than by rigorous measurement. But the phenomenon was real enough: workers were playing, and their managers were noticing.

A counter-thesis emerged at the same time. Some researchers and business writers argued that Solitaire was not a productivity leak but a productivity tool — that a few minutes spent moving cards reset attention, reduced frustration, and made longer focus sessions possible. That claim was also hard to measure with precision, but it lined up with broader research on micro-breaks and cognitive recovery. The more generous version of this argument pointed out that Solitaire had quietly taught millions of people to use a mouse, which made the next wave of graphical applications easier to adopt. On that view the game was infrastructure disguised as pastime.

IT departments responded in predictable ways. Some hid or removed Solitaire from the standard image. Some blocked the executable at a group-policy level. Some shrugged. The countermeasures rarely worked for long, partly because new versions of the operating system reintroduced the games and partly because employees who wanted to play simply copied the executable from home. By the late 1990s, most corporate policies had settled on tolerance punctuated by occasional crackdowns.

The cultural effect was larger than the economic one. Windows Solitaire produced a shared ritual — the clicked deck, the bouncing cards, the small dopamine hit of a finished foundation — that ran across industries and job titles. The game gave office workers a common language for a specific kind of micro-idleness. That language outlived Windows itself.

Once Windows Solitaire became universal, it became a shorthand. Films and television used a shot of Solitaire on a monitor to communicate “this office is dead inside” or “this employee is no longer engaged with their work.” The image was unambiguous in a way that few other computing artifacts are. A spreadsheet might mean work; an email might mean anything; the green felt tableau meant one specific thing. The visual vocabulary was so settled that even viewers who had never played Solitaire could read it.

The game also became internet material. Screenshots of the bouncing-cards victory animation circulated on forums through the 1990s and 2000s. Memes about being caught playing Solitaire by a passing boss predate most modern meme conventions. The cards, the pattern backgrounds, the timer in the bottom corner — each became a visual shorthand in adjacent corners of internet culture. When Microsoft retired the classic art in 2012, the reaction included a non-trivial thread of nostalgia: the game belonged to the people who grew up with it, not to the company that shipped it.

Solitaire appeared in crime dramas (a detective killing time), in sitcoms (a coworker hiding a window), in comedic news segments (a senator caught playing during hearings), and in serious journalism about the slow texture of office life. The game’s ubiquity made it useful symbolically in a way that few pieces of software have matched since. The modern equivalents — TikTok in a meeting, Slack during a call — do cultural work that Solitaire did first.

The Microsoft Entertainment Pack, released in several volumes starting in 1990, bundled Solitaire and FreeCell with a lineup of other small games: Minesweeper, TicTactics, Cruel, Golf, Tetris, Taipei (the Mahjong game), and a handful more. The Pack was a retail product, sold on floppy disks through computer stores, and it is the first place that FreeCell reached a broad consumer audience. Klondike Solitaire shipped with Windows 3.0, but FreeCell was an Entertainment Pack original before it migrated into the base install.

The Entertainment Pack is also where several of Microsoft’s other durable game franchises got their start. Minesweeper became the second most iconic Windows game, and its original artwork also came from Kare’s contract work. Taipei introduced a generation of American players to Mahjong tile-matching. The Pack was, in effect, Microsoft’s casual-games label before casual games had a label, and it trained the company in the economics of small, habit-forming software long before mobile app stores existed.

The Entertainment Pack line faded in the late 1990s as its key titles migrated into the base Windows install. By Windows 95 the Entertainment Pack had effectively become the set of games you already had, and the retail boxes stopped appearing on shelves. What remained was the pattern: Microsoft owned the casual-games relationship with the Windows user, and that ownership was taken for granted across the next twenty years.

Windows 3.0 shipped Klondike Solitaire. Windows 3.1 polished the visuals and kept the core game. Windows 95 added FreeCell to the base install, along with Hearts (networked) and the rest of the Entertainment Pack lineup folded inward. Windows 98 and Windows ME were incremental: minor visual refreshes, the same game, the same deals. Windows XP brought Spider Solitaire into the default set, which widened the franchise and introduced millions of users to a card game they had not previously played.

Windows Vista and Windows 7 refreshed the visual design (the famous Aero-era card backs, the slightly glossier table surface, a redesigned animations palette). The core games did not change in any meaningful way, but the visuals carried the era’s overall aesthetic. Windows 7’s Solitaire is, for many long-time players, the canonical version — the one they remember when they remember the game.

Windows 8 was the break. Microsoft removed the classic games from the base install and replaced them with the Microsoft Solitaire Collection, an ad-supported full-screen app distributed through the Windows Store. The change was strongly unpopular. Players noticed immediately that the game they had been playing for twenty years had been swapped for a product with banners, video pre-rolls, and a subscription offer. The core rules remained, but the experience had moved from “part of the operating system” to “a separate app with a different philosophy.”

Windows 10 and Windows 11 continued the Collection model. The classic Solitaire and FreeCell binaries still ran if copied over from older machines, and a modest cottage industry grew up around repackaging the Windows 7 executables for later systems. The Microsoft Solitaire Collection itself remained a Windows Store fixture, now bundled across Xbox Live achievements and connected to Microsoft accounts. The arc from “game you found in Accessories” to “app you signed into” took about twenty-five years.

The Microsoft Solitaire Collection, launched in 2012 on Windows 8, bundled Klondike, FreeCell, Spider, Pyramid, and TriPeaks into a single app with an achievements system, daily challenges, and an in-app store. The app was free to download and showed advertisements between games and as banners during play. A subscription product (Premium) removed the ads for a monthly or annual fee. The Collection was controversial at launch and has remained a mixed product in public reception.

The controversy had two layers. First, the Collection replaced rather than supplemented the classic games, which meant users who upgraded Windows found their favorite free product swapped for an ad-supported one. Second, the advertising intensity was notably higher than players were used to, with 30-second video ads common between games. Microsoft adjusted the ad mix over the years, and the Premium subscription has remained an option, but the perception that the Collection is a monetization product rather than an included utility has stuck.

Despite the reception, the Collection is probably the single most-played piece of software Microsoft ships. The install base is enormous, the player retention is strong, and the daily-challenge mechanism gives the app a reliable return-visit rhythm. As a business line, it is successful. As a cultural artifact, it is the thing that displaced something beloved.

The reaction to the Collection created space for a browser-based and mobile-app solitaire ecosystem to grow around Windows. Open-source projects like PySolFC preserved the classic rules and deal numberings. A generation of independent websites reimplemented Klondike and FreeCell with modern conveniences (unlimited undo, hint systems, statistics tracking) while keeping the visual grammar players remembered. Mobile apps took the same ideas to iOS and Android, where a single good Klondike app could earn a subscription revenue stream rivalling the Collection itself.

That ecosystem is what we participate in at SolitaireStack.com. The continuity from Alfille to Cherry to Horne to the open web is direct: we ship the same rules, we render the same cards, we honor the same deal numberings, and we write for players who either grew up with Windows Solitaire or inherited the habit from someone who did. The card game has outlived the platform that carried it, which is the final move in the story this page tells.

The mobile era added a second wave. iOS App Store and Google Play Store gave solitaire developers a direct channel to users who had moved on from desktop computing entirely. For a substantial share of players under forty, the first place they ever saw FreeCell was a phone, not a Windows machine, and the first Klondike they touched was a mobile app served with banner ads similar to those that now populate the Microsoft Solitaire Collection. That convergence — ad-supported mobile apps on one side, ad-supported desktop collection on the other — tells us something about the market’s current resting state: solitaire is free to access, but it is usually monetized somewhere in the loop, and the platforms that deliver it compete on the quality of their monetization as much as on the quality of their implementation.

Our hope is that the history here helps readers see what they are doing when they open a browser tab to play Solitaire: they are participating in a decades-long line that runs from hand-dealt patience games on nineteenth-century parlor tables, through the PLATO terminal rooms of the 1970s, into the office computers of the 1990s, and out into whatever comes next. The game is older than the computer, older than the operating system, older than Microsoft. Windows did not invent it. Windows made it universal.

Our primary sources for this piece are interviews with Wes Cherry and Jim Horne conducted by technology journalists over the years (we paraphrase rather than quote), Microsoft’s own release materials for Windows 3.0 through Windows 11, Brian Dear’s PLATO historical archive and the documented record of Paul Alfille’s 1978 FreeCell implementation, and the academic and trade-press literature on workplace computing and casual games. Susan Kare’s card art is documented in her own published design catalogue and in Microsoft’s release notes from the Windows 3.0 era.

Productivity-loss estimates cited in general culture over the years have rarely been rigorously methodologically supported, and we have not reproduced specific dollar figures here because the sourcing for them is usually weak. The qualitative claim that Solitaire became a universal office phenomenon is supported by a long track of contemporary news coverage, internal IT policy debates at large employers, and later academic retrospectives.

Claims about the origin of specific variants (especially Spider Solitaire, whose pre-Windows pedigree is disputed in the card-game literature) are hedged in the text. Where we have had to choose between competing accounts, we have leaned on the sources closest to primary documentation.