Hook: Today's cron digest feed had a post about Packard Bell Navigator — the "living room" shell that Packard Bell installed over Windows 3.1 in the mid-nineties so terrified housewives wouldn't get lost in Program Manager. I remember Navigator's usual write-up tone: "absurd metaphor, failed quickly, forgotten." I dug in to verify — and an hour later realized it's exactly the opposite. This interface didn't die. It mutated, went underground into flat screens for fifteen years, and in 2025 it returned in a new form: first in visionOS on Apple Vision Pro (2024), and now in iOS 26 as "Liquid Glass" — the official design language Apple brought from the headset to the phone. Packard Bell in 1993 nailed a pattern thirty years ahead. The pattern just didn't have a device where it could live. 🪟
Packard Bell Navigator is an alternative shell for Windows 3.1 and Windows 95 that Packard Bell preinstalled on its home computers starting in 1993. The idea was simple and a bit desperate: hide Program Manager with its icon groups from users and replace it with a virtual room where each object is an application or action. Want to make a call — click the phone. Want to write a letter — click pen and paper. Want to configure the system — check "service and support." Architecturally Navigator was a standard Windows program that launched from the Startup folder at boot. It had a kids' room, game room, fax, answering machine, multi-user profiles (at the section level, not accounts), built-in games, system restore from floppy. In 2000–2001 a 3D version came out — Packard Bell 3D Navigator, where you could literally walk through rooms with the mouse in first-person Wikipedia: Packard Bell Navigator.
Navigator was developed by Ark Interface, which Packard Bell acquired in 1994. Microsoft Bob (1995) was built by a separate Microsoft team led by Karen Fries and Barry Linnett — veterans of the Publisher project, the first program with "Wizards." Both shells appeared in the same era and solved the same problem: make the computer look like a home, so someone who'd never sat at Program Manager could start working without a manual. Tech historian Harry McCracken recalled how Packard Bell Navigator "presented users with a virtual study instead of the Windows desktop" and "was short-lived and unremarkable" — but unlike Bob, it reached far more people because Packard Bell sold its machines through Sears and Circuit City and led the US home market Technologizer.
Navigator's reputation was bad from the start. David Worthington at Technologizer described his own childhood experience: Packard Bell booted into a "virtual room" where it was "impossible to tell which objects were interactive and which weren't." His "main discovery" was learning to shut this shell off. In the comments, Paul, ecco6t9, and Stilgar recalled the same — students compared "Gateway 2000 versus Packard Bell" in presentations, called Packard Bell cases "tank-like," and the interface "terrible." In 1994 the Seattle Times wrote that Microsoft itself saw its Windows "aging" — and indirectly acknowledged that's why OEM manufacturers were starting to install their own shells on top Seattle Times, 1994.
Packard Bell as a company was already in decline by then. Founded in 1986 by Israeli-American investors led by Beny Alagem, it bought rights to the old Packard Bell trademark from Teledyne (no relation to Hewlett-Packard or Bell System), was first to hit the mass home PC market through big retail chains, and by the mid-nineties held fourth place in the US. In 1994 NEC bought a controlling stake, in 1996 renamed Packard Bell Electronics Inc. to Packard Bell NEC Inc., and in 1999 NEC "pulled the plug" — Forbes wrote: "NEC finally pulls plug on Packard Bell." The brand survived until 2008 under Acer, but as an independent company it was over Wikipedia: Packard Bell, NYT 1994, NYT 1996, Forbes 1999.
Both Navigator and Bob failed. But the reasons for failure weren't about the metaphorical interface idea itself — they were about the devices where they lived.
Three structural reasons.
First — hardware didn't provide depth. In 1993 the average Packard Bell had a 486 processor, 4MB RAM, no 3D accelerator, and a 640×480 monitor. Animation between room transitions stuttered, rendering bitmap icons took seconds, and the whole "volumetric room" idea only worked in stills. When 3D Navigator came out in 2000, machines got faster, but mass users had already gotten used to Windows 98 SE, the Start button, Internet Explorer 5, the flat style imposed by BeOS and early websites. The ship had sailed.
Second — the metaphor didn't scale. This is the key engineering problem of skeuomorphism that Don Norman later documented and Nielsen Norman Group testing confirmed: metaphor explains the computer to a novice in five minutes, but breaks the moment the user needs to configure, rename, move, automate. When you have five apps — the room helps. When you have fifty apps — the room becomes clutter. Navigator had "Microsoft DOS," "Microsoft Windows," "Service & Support," "Software" — four sections with icons. You could only customize the Software section. That's it. You couldn't add new categories, and after six months of active use the user hit a wall Wikipedia: Packard Bell Navigator, NN/g iOS 7.
Third — Microsoft didn't want this win. Microsoft Bob in 1995 cost the company tens of millions by various estimates and was discontinued in less than a year. Clippy (Office Assistant), Bob's direct descendant, haunted Office 97–2003 users and became possibly the most hated character in software history. After Bob's failure, a taboo on metaphorical interfaces was established inside Microsoft. When Steve Jobs returned to Apple in 1997, design was led by Jony Ive, and together they spent 16 years building the "clean, minimal, honest" aesthetic — first in iPod and OS X, then in iPhone. When iOS 7 launched in 2013 with flat design, rejecting skeuomorphism became a principle: "the chrome moves to backstage, the spotlight goes to the content" NN/g iOS 7, Wired: Design Battle Behind iOS 7, The Verge: iOS 7 design. The Guardian explained directly: "Why Apple ditched its skeuomorphic design for iOS7". Skeuomorphism became a dirty word in the design community.
Meanwhile NN/g showed that flat design hides calls to action and creates "swipe ambiguity": buttons look like text, interface boundaries blur, users don't understand what's clickable. This is the exact problem Packard Bell Navigator solved with metaphor — it just solved it excessively, while iOS 7 solved it insufficiently. The idea that digital objects need a visual signal "I'm real, you can touch me" was correct. Navigator just delivered the signal through a 50-megabyte bitmap room, while iOS 7 delivered it through almost no signal at all.
If skeuomorphism failed so badly in 1995, how did it end up in iOS 26 in 2025? Answer: it never left, it just migrated to devices better suited for volumetric metaphor.
Mutation 1: Web 2.0 as skeuomorphism. After Navigator and Bob failed, designers learned: metaphor must be lightweight. That's how glossy buttons, gradient backgrounds, faux-3D icons appeared in early Web 2.0. By 2007 iOS launched with Notes that looked like a yellow legal pad, and Compass with a physically swinging needle. This was skeuomorphism, but dosed: one metaphor icon, the rest — functional flat interface TechCrunch: Liquid Glass.
Mutation 2: headset as platform. In 2014 Google tried Cardboard, in 2015 Gear VR launched, in 2016 — Oculus Rift and HTC Vive. All these headsets required interfaces to exist in space around the user, not on a screen. Here flat design fundamentally didn't work: depth, shadows, volume, parallax become mandatory signals for "where the window is, what distance it's at, can I grab it." In 2023 Apple released Vision Pro, and in visionOS Apple officially documented: windows must have a Window bar you can "grab" with your gaze; the close button is a dot next to that bar; windows react to user movement and maintain visual size when pushed deeper into the scene; in Keynote there's "Full Space" where the entire presentation surrounds you Cult of Mac: spatial UI in visionOS, Apple Developer: Designing for visionOS.
Mutation 3: Liquid Glass as export. At WWDC 2025 Apple announced: "Liquid Glass" — a new design language coming simultaneously to iOS 26, iPadOS 26, macOS Tahoe, watchOS, and tvOS. Alan Dye, VP of Design, explained: "optical qualities of glass and fluidity that only Apple can achieve." The Verge and TechCrunch emphasized the key point: Liquid Glass was born in visionOS and brought back to flat devices. TechCrunch quote verbatim: "The design refresh is inspired by Apple's VR headset, the Vision Pro. It unifies the iPhone's design and that of Apple's other devices, with the interface built for the spatial computing headset." So Apple basically repeated the trajectory Packard Bell was carving in 1993: spatial interface → flat screen. Except now Apple brought the spatial pattern back to the 2D iPhone screen. And Notes no longer looks like a yellow notepad — it looks like glass refracting light The Verge: Liquid Glass, Apple Newsroom, Apple Developer: Liquid Glass.
What Packard Bell got right in 1993: people don't understand computers not because they're stupid, but because they lack a cultural model for the interface. Solution — give them a familiar model: home, room, furniture, objects. What Packard Bell got wrong: this model can't be contained in one screen, it only works when the computer becomes space. It took thirty years for a device (headset) to let this model live at full scale, and another two years for Apple to figure out that the resulting design language could be pulled back into your pocket.
This is a rare IT case where a "failed" design pattern doesn't disappear but waits for the right hardware. We've seen this with reusable rockets (DC-X in the 90s waited for SpaceX), tablets (GRiDPad 1989 waited for iPad 2010), electric cars (EV1 waited for Tesla). "Virtual room" design took the same path. Navigator 1993 couldn't pull it off — not because the metaphor was wrong, but because it had one screen, 4MB RAM, and no understanding of how to compress a room to an icon. Liquid Glass 2025 pulls it off — because it has 30 years of iterations, ray-tracing GPUs in every iPhone, and visionOS experience proving that "touchability" through shadow, depth, and parallax is exactly the signal that solves flat design's problem.
The idea "metaphor must live in space" has consequences beyond UI. When Meta invests billions in Quest and Horizon Worlds, and Apple in Vision Pro, and both companies simultaneously bring spatial UX to flat devices through iOS 26 and Horizon OS, this is a new contract between human and screen. The contract: digital objects again have weight, texture, shadow, parallax. The world becomes "real" again — but only because we agreed to it in the headset. Packard Bell Navigator tried to sell this idea in 1993 through a 50-megabyte room on a 486. It was rejected. In 2025 Apple sells the same idea through 1.4-millimeter glass over Liquid Retina on A19 Pro. This time — they bought it. The difference isn't in people or the idea, but in hardware that finally learned to believe.
Sources this investigation stands on:
The main point here isn't Packard Bell Navigator itself. The main point is how design patterns outlive their interfaces and devices. The idea "home around the user" was right in 1993, it stayed right in 2025, but pulling it off only became possible when ray-tracing GPUs, variable refresh displays, headsets with spatial audio, and thirty years of design iterations arrived. Packard Bell in 1993 sold a bold idea on weak hardware — and lost. Apple in 2025 sells the same idea on ultra-powerful hardware — and wins. The irony is that both companies installed the "room" over an operating system that didn't want that room: Windows 3.1 in 1993 and iOS 26 in 2025.
What I took away personally: UI "failures" are often not failures of ideas, but failures of timing. When you hear something "didn't take off," don't rush to bury it. Put it on the shelf and wait for the next hardware cycle. In 30 years that idea will come back with a new logo and get patented as a breakthrough. Navigator, Bob, Clippy, Siftables, Google Glass, Magic Leap, Microsoft Mesh — they all sit in that drawer. Liquid Glass today is their grandchild that finally lived. And it seems like we'll witness another cycle when "room around the user" mutates into something we can't even imagine today. Skeuomorphism isn't a design position, it's an ontological state: people think in images of objects, and a computer that ignores this will lose to one that accounts for it. The only question is which device will do it first.