Remember when it was easy to distinguish between app and document icons? Apps had irregularly-shaped icons, and docs had tall rectangular icons.
In order to reflect on that, one must also reflect on the function of icons to serve both a paradigmatic and readily intelligible purpose.
Application icons (long before Apple steered folks toward calling them “apps”) were, for at least the first 25–30 years, literal or readily relatable/intelligible signifiers for what each application delivered in function. Stickies was represented by classic, canary-yellow post-it notes; Photoshop was represented by a digitally stylized eye; and Safari was represented by a compass (itself riffing from when Netscape Navigator, Firefox’s antecedent, was represented by a boat steering wheel).
Document files associated with applications adopted variations on the document base icon being a portrait-oriented page of paper, paired with readily identifiable graphic elements one associated with the application icon. This worked well for Word documents (the blue “W”) and Illustrator files, as these still relied on simple representations of paper.
Frankly, “this just worked” — not solely because of a plain, A-B relationship between application and associated file, but also because these often nodded to their functional origins in an analogue world. This was a kind of pre-skeuromorphism. QuarkXPress was especially good at doing this back during the QXP3 days: application and document icons hinted at pasted layouts from the time when book print, newpapers, and posters were set up on pasteboards.
Some of what changed, as ties to those analogue origins became less relevant, were efforts to develop applications with no direct relationship to analogue antecedents. VLC comes to mind: only an insider within the motion picture industry might make the literal connection between the use of traffic cones (when a production team is filming on set in a public space) and a movie; for everyone else, that cone just became the “VLC icon” — especially so for anyone who’s come of age since VLC became a media player.
Since then, obviously, plenty of applications have emerged for which there’s no way to reach back to an analogue origin, so abstractions became commonplace.
This is also when applications — once standalone binaries/products — came to be known as “apps” and were held in equal stead with “apps” which were little more than springboards for a company’s own web service/portal (without relying on a browser to get there, and also to circumvent any customizations one used on their browser). Basically, this included every social media platform on the planet and any service reliant on front-end springboard tech like electron.
So yah, while this complicated how application — or, “app” — icons were designed, they
did manage, for a time (over the last dozen years), to distinguish themselves with a literal icon, shaped however the “app” developer envisioned it. Twitter was a bird; FB was an F (shocker); and Flickr was two circles (one blue, one magenta).
And by and large, this approach still worked as it had since the Xerox PARC, Lisa, and OS 1 era: files reliant or associated with an application took on a “page”-esque shape to denote it was a document dependent on that application. Even so, social media applications — “apps” — don’t generally have associated file icons, so that relationship began to recede, especially as many former applications became cloud and subscription services, in which former “files” no longer have a home on one’s own device.
tl;dr: blah blah blah from a former turnkey art director who got burnt out long ago and switched careers.
Now we have square app icons (with rounded corners) and slightly-taller doc icons. Once Apple started doing it, other developers rushed to copy for seemingly zero actual benefit. In fact, of the 28 apps currently in my Dock, only 2 don't have squared-off icons.
The square — or, rather, the square, as softened by Bézier curves — was, I’m guessing, a way to denote that an “app” could be pressed, literally, like a virtual button on a glass UI.
Which, fine, might work when navigating within the kludgy constraints of digits mashing on glass and expecting a haptic response. This can work on an iPhone, iPad, or Android device. But for former icons which came to be before this move to virtual buttons (and forcing a uniformity on the “button” as an interface), it meant shrinking them to fit inside the one-form-for-all button, i.e., that Bézier square.
The biggest problem with shoehorning all GUIs into this paradigm is the all-digital (that is,
finger digits) nature of their function doesn’t translate
at all to the legacy, cursor-based UI of a laptop or desktop GUI display — at the very least, for those which lack a touch screen. It simply doesn’t work effectively. Moreover, it forces the eyes and the brain to do more work in parsing the bounds of an icon which, as you noted, are now all the same dimensions. This brings up accessibility issues, especially as eyes age.
Frankly, it’s not hard at all to draw a direct evolutionary lineage between the (useful) favicon on web sites — which weren’t really a thing before around 2002, or maybe 2003 — and the uniform, Bézier square icons used across glass UIs and, counterintuitively, on cursor-based UIs. “Neatness” doesn’t, however, translate to “more user-friendly”. UI neatness for its own sake is form without par function.
And this is where, I conjecture, Ive’s muckraking in the human interface design broke application-document relationships: good industrial design principles translate poorly to HID principles. It further constrained application/app designers from producing visually intelligible icons — now reduced to fitting something inside that square which somehow manages to both be legible and not confused with other “app” icons.
(I’ll even be gripey and crusty enough to argue how the MBP’s Touch Bar was another Ive special beset by fundamental questions of form leapfrogging function for sake of superficial aesthetics, in Ive’s eye. Even as Ive was backing away in 2014 and Touch Bar didn’t arrive to production until 2016, it has all the hallmarks of Ive trying to shoehorn industrial design values into human interface design.)
unloved child of tl;dr: This is all a train-wreck, and it exhausts me just to have to think about it (again).