Apple Inc's Liquid Glass is a design language introduced this year which is computationally inefficient, in some ways a regression in terms of usability, and looks extremely cool. It’s fun to play around with, as you’ll find yourself slowly scrolling the content under a liquid glass element to watch the refracted image warp and stretch. This major departure from the dominant design language of the decade prior has been met with divided reactions. As the #1 Liquid Glass fan, subtlety does not make my list of concerns. My opinion diverges from the critics as I find this a welcome change from the generic, sterile, flat design that plagued user interfaces for the 12 years prior (the Corporate Memphis/flat UI era—dark days for software design, finally come to an end). While that's all I'll say about liquid glass for now, the rest of this post remains laser focused on the user interfaces where aesthetics take precedence over functionality and how design trends and technological advancement follow each other, approaching the topic through a focused retrospective look at one of the least functional and most stylized UI elements from two of Apple’s design languages ago, now about 20 years old.

Cover Flow

Cover flow was conceived in December 2004 by designer Andrew Coulter Enright, and independently implemented by developer Jonathan del Strother, who sold it to Apple shortly after. In Enright's blog post introducing the idea, he opines what was lost in the then-recent transition to digital music:

“As much as I love metadata, I just don't get as much satisfaction scrolling through 10,000 songs as I did flipping through 100 cracked CDs (let alone a couple hundred dusty LPs).” Album art was either missing entirely, or tacked onto a sidebar as an afterthought in the music library interfaces of the day. Music—of all the other things you can access on the computer—is something where speed and density are not the most important priorities, and a wall of text is a particularly ineffective way to present it.

The original mockup, seen above, was directly inspired by a three-dimensional multi-person video conferencing interface in iChat, a FaceTime predecessor. Cover flow was a highly skeuomorphic UI, representing a “faster, more artistically complex and more emotionally potent way to experience our collections.” I somewhat disagree with that first point later in this post, but the rest holds true.

Cover flow is now absent from any iteration of Apple's music player software from the past decade. In the years since the heyday of this UI element, major changes in the music software space include streaming becoming the dominant music format. Every music streaming app, of course, includes images of cover art throughout the browsing experience. They are undoubtedly visually attractive and designed to capture and hold attention. Apple Music’s main page features rows of horizontally scrolling rows of album covers, slightly overlapping the boundary to indicate the presence of content past the edge (a technique popularized by Windows Phone 7). A mix of albums, mixes, and playlists, both familiar and novel, is presented to you in an order determined by an algorithm that predicts what will be most relevant at the time you load the app. The library view features a two-dimensional grid of album covers, and the now playing screen features the album cover prominently, taking most of the width of the screen with playback controls below it. Other apps such as Spotify feature largely the same elements. These are mature, >1 decade-old apps where success is measured in subscription conversions and retention. Whether or not that incentive leads to good design may be of scope for this post.

Enright was tuned in to shifts in the software ecosystem in late 2004: in his words, people love an "artistically complex and more emotionally potent way to experience our collections." "Jukebox" software existing up to then, such as early iterations of iTunes, presented music as a list of tracks with a few columns such as title, artist, year, etc. with visual appeal rivaling nothing except perhaps a particularly low-effort spreadsheet. People used it in spite of its design because digital music was novel. The market for it had yet to be commodified. The technology itself was rapidly evolving, with record labels scrambling to grapple with a world where music distribution was divorced from physical distribution of media: sharing mp3's could land you implicated in one of the highest-profile court cases of the decade. Digital music was cool and edgy—the software used to play it back didn't need to justify itself with any particular look or feel. Not to say that nobody cared: theming was the realm of hobbyists—the Winamp skins of the era leaned into an immature and edgy vibe, completely absent any requirement to take themselves seriously.

(author's note: I was a small child at the times being referenced in much of this post, and am gleaning second-hand accounts from the web. Accuracy cannot be guaranteed.)

The Maturing Desktop, the Emerging Mobile Device

By 2004, the iPod, three years into its run, was well established as the crowd favorite portable music player, and iTunes was both the dominant digital music store, and the desktop app with a monopoly on interfacing with the iPod.

Apple in the early '00s was leaning into their Aqua design language, featuring bubbly, animated UI decorations throughout OS X. Microsoft was soon to introduce Aero to Windows, the well-loved design language rich with transparency and naturalistic flair. These were digital interfaces built to be shown off for their own sake. Tech companies were flexing their creative talent by incorporating allusions to nature and the physical world in software, differentiating it from the drab software aesthetics of the late 20th century when personal computing was rapidly evolving and its use cases were novel enough not to require much thought given to look and feel. Computers makers didn't need to justify the use case of the desktop, they needed the software to become effortlessly cool and beautiful—qualities of enlightenment and luxury. This is a natural evolution of technology: first it justifies its use case; after that's done, it needs to look and feel nice in order to have staying power.

The iPod could be seen as an evolution of the earlier mp3 players, which were all ugly and couldn't decide what shape to take. By contrast, Tony Fadell's original iPod looked and felt nice. The software that ran on it was absolutely barebones, with menus comprised of lists of text, and a now playing screen that showed the playback status. It had a very low resolution monochrome LCD after all—there's only so much you can do with that hardware. That didn't stop it from being a wild success though, as this device had nothing to prove in terms of software design.

Cover Flow was emblematic of the superfluous visual detailing in digital interfaces. It doesn’t add anything that a simple 2D scrolling interface can’t do, from a functional perspective. Cover flow shows you the edges of your album covers, which overlap each other, sitting parallel at about a 45 degree angle to the plane of view, angles mirrored on either side of the center, with only the center position being straight-on and visible in its entirety.

(Many third-party implementations get the projection wrong: they use a three-dimensional perspective where the ends of the list are further from the camera than the center. The projection to achieve the original effect is a combination of an orthographic camera, and perspective distortion applied to each item. In other words, there is no perspective height difference when moving from left to right in the list (only the center item is enlarged as it "moves forward" when selected (neither an orthographic nor perspective camera achieves the effect)), but the front edge of a tilted list item is taller than the rear edge. It doesn't actually work in 3D space and is not easy to vibe code or to implement in a game engine. Many implementations also get the tilt angle wrong: a naïve approach is to tilt the items continuously as they move from left to right. The correct approach is for the list items to remain at a fixed angle until they reach a certain proximity of the center position, where the tilt animation starts. There is also a small dead zone surrounding the center position where the center item slides without tilting.)

The visual metaphor vaguely invokes a stack of records, or CDs on a shelf, where you don’t need to inspect the full surface of an item to identify it; just a little of the edge will suffice. The key to the imaginary model here is that there are two important axes to the collection in physical space: the axis on which items are stacked, and the plane of the surface of each item, where you must stick your finger in and perform a little tilt between axes, or slide an item parallel to the surface, to reveal the contents of the item. This is how one might browse a shelf of CDs with just a little wiggle room. Cover flow takes this familiarity and expands it into something more whimsical.

On a digital interface, of course, there’s no such space restriction as with a limited-length list of media on a physical shelf, and no required spatial translation to access the contained information. You can present everything from any perspective, and scroll an arbitrary distance in any direction. Cover flow forces the metaphor for purely aesthetic purposes. Its biggest (and only) strength was always that it looked cool. Its designer knew this, Apple knew this, and they featured it prominently in promotional materials for mobile devices.

Cover flow was also featured in OS X's Finder, presented as one of the four main view options. Its use cases in this context were somewhat niche compared to the more versatile grid and column views, but there can be found on the internet accounts from designers and photographers who lamented the feature's later removal. Its appeal was the extremely large thumbnail size for browsing galleries, with a central item surrounded by items in its vicinity in the sort order, in a linear layout designed primarily for visual browsing where metadata is relegated to a small footer. The gallery view that replaced it places the thumbnails in a tiny row and only a single item in the scaled-up viewer above the thumbnail row.

The early iPhones featured cover flow as the landscape mode interface within the music app. In 2007, a large part of the pitch for the iPhone was its widescreen multi-touch display, which was a departure from the earlier smartphone format of a small display situated above a hardware keyboard. Although it wasn't devised for this purpose, cover flow was the perfect flashy UI toy to show off the display width and multimodality of the iPhone.

That being said, there’s nothing you can access through cover flow that you can’t also access through some other list interface. It is distinctly slow and one-dimensional. Compared to the "normal" list view in the iPhone's (or any other) music app, cover flow has low information density, and introduces visual complexity that does not aid in quickly reaching a position in a list. It fills a lot of space with the image thumbnail for each item, and does not facilitate skipping quickly to an index—it can be indexed, as in the Finder implementation, but the width of the items and the animations introduced make it much slower—both computationally and perceptively—compared to a vertically-arranged list of items represented by text labels. A purely vibes-based UI element needs to justify itself by fitting the vibe of the intended experience of the product as a whole—thus the feature’s existence on Apple platforms was relatively short-lived. It sold iPhones at a time when digital music players were familiar, primitive cell phones were familiar, but the concept of a device with a large touchscreen that could be used either vertically or on its side was something in need of a pitch.

Having served its purpose, cover flow was deemed superfluous and swept up as clutter, amidst the flattening of UI that took place in the early to mid ‘10s. A landscape grid view replaced it in the iPhone music app, but that lasted less than two years before a unique landscape-mode music browser of any kind was dropped in 2015. The concept of tilting your phone sideways for a different experience is ultimately not something that really caught on—in the music app or anywhere else, except for games and videos that only work sideways.

The last version of cover flow by that name was axed upon release of Mac OS Mojave in September 2018. The cover flow in Finder, for the last few years of its existence, was a flattened version of the interface, showing large thumbnails slightly overlapping each other in a row, but missing the angles and reflections.

As an aside, many sources cite a 2008 patent lawsuit from now-defunct patent troll Mirror Worlds as the reason for its discontinuation. I have fact-checked the patents referenced in the suit and found that they do not actually contain anything resembling cover flow. Mirror worlds was awarded damages, Apple appealed and got the judgement overturned, and the Supreme Court declined to hear an appeal from Mirror Worlds. It is clear in the context of Apple’s design direction in the '10s that this legacy interface element was on its way out regardless of what the suits preoccupied themselves with.

The three-dimensional, animated quality of cover flow is largely missing from any modern music app. There exists a version of it as an option for KDE Plasma's task switcher. This is cool, and it's well done, but my opinion is that it is not well-suited to that use case, because a task switcher needs to be extremely fast and the thumbnails all glanceable. It's not something that you browse at a leisurely pace for fun, as you might with a music collection or photo gallery. It is my opinion that cover flow is excellent for those use cases.

The Present and Future

Nearly 21 years have elapsed since Enright's blog post reminiscing on the feeling of flipping through CDs that was lost in the digital music revolution, and I am feeling much the same way. Note: I do not think that streaming music is bad for you, or that new apps are ugly, or that software is getting worse, on the contrary, I think much of the newest software design as of late September 2025 is beautiful compared to that of the decade prior. I have decided to opine cover flow specifically. I miss the software metaphor for flipping through CDs. The good news is that it isn't going anywhere, as over the past few months I have brought it back (don't say "cover flow" though, it's just called the carousel) in a new music app called mtoc, on my preferred platform (Linux desktop) with a ground-up reimplementation in Qt 6, making a sincere attempt to stay faithful to Apple's original rendition. You can check it out here. Future posts on this blog will detail the project, its design process, and implementation.

Further Reading

Navimation: Exploring Time, Space & Motion in the Design of Screen-based Interfaces (International Journal of Design, 2010)