The Pixar animated short Lifted was released in front of Ratatouille in 2006:

I’ve always been amused by this imaginary interface, which is so clearly not how any sort of computer would work.

Or so I thought. These are photos I took in Melbourne in 2024 of CSIRAC, Australia’s first digital computer from about 1949:

This is a “console” of the computer, used to tactically probe or input specific memory addresses (in binary), and to control functions like stopping and starting the program. Any proper programming and eventually inputting data would happen using gentler I/O devices like typewriter keyboards, paper tape, and magnetic storage.

Physical consoles like this one were last seen in the 1970s on hobbyist home computers such as the Altair 8800, and the Console app on your Mac diligently spitting out logs is its spiritual and virtual successor. But even if a CSIRAC console feels hostile today, 75 years ago it was quite the opposite:

And [CSIRAC] helped there too. It could display all its working registers and the last 16 instructions executed. It could be given an address at which to stop (a “breakpoint”), and be stepped by one instruction at a time. It even had lights to show the computer’s internal states. This was a user-friendly computer.

CSIRAC stood for Commonwealth Scientific and Industrial Research Automatic Computer, a typical naming scheme of the era. We also got ENIAC (Electronic Numerical Integrator and Computer) in 1945, BINAC (Binary Automatic Computer) in 1949, EDVAC (Electronic Discrete Variable Automatic Computer) in 1946, ILLIAC (Illinois Automatic Computer) in 1952, and then SEAC, SWAC, ORDVAC, TREAC, AVIDAC, FLAC, WEIZAC, BIZMAC, RAMAC, and UNIVAC.

The story goes that the name of 1952’s MANIAC (Mathematical Analyzer Numerical Integrator and Automatic Computer) was chosen to highlight and put a stop to the goofy naming practice. Did it work? I am not sure. Not only two more MANIACs were produced, but we also got 1953’s JOHNNIAC (nicknamed “pneumoniac” since it needed a lot of air conditioning), and SILLIAC (Sydney ILLIAC) in 1956. The last computer I can find using that naming scheme was TIFRAC, operating in India between 1960 and 1965.

CSIRAC had real work to do, but today it is known chiefly for being the first computer to play music in real time. The quality is… I’ll let you judge, with links below pointing to short MP3s preserved by Paul Doornbusch and subsequently Internet Archive:

• Auld Lang Syne
• Chopin’s March
• In Cellar Cool
• (I particularly enjoyed an alt recording of In Cellar Cool where CSIRAC itself appears in a background as a constant humming presence.)

Do you miss your PC speaker yet?

Engineers working on other room-sized computers of that era did similar things; whether this was solely one of the first attempts to humanize the big scary machines, or a distraction from the computers’s typically military uses is left as an exercise for the listener.

Today, one of the 1960s machines still plays music, headlining a fascinating annual tradition – every December, the PDP-1 restoration crew at the Computer History Museum in California invites visitors to sing carols with the computer older than most of them.

The last photo takes us back to where we started. Neither CSIRAC nor PDP-1 might be user-friendly by today’s standards but damn, wouldn’t you want some of your computer’s interface to feel this way?

A fun 24-minute video from Technology Connections about designed sounds in real life: elevator dings, airplane chimes, railway crossing dings, and so on.

While I am sympathetic to the notion that sound pollution is a thing we need to be concerned with, the choice between silence and sound pollution is a false choice. There’s a lot of those happening these days, probably because we’re so stuck in binary thinking. But as airplanes show us, we can design sounds which aren’t obtrusive, but which are helpful. And when you get yourself out of binary thinking, you can do things like make your most obnoxious apps be silent while your important ones make themselves known, and in ways which are meaningful to you and pleasant to everyone else.

It is an interesting parallel to the post about syntax highlighting from a while back, and one of the posts about cartography design I shared recently; they all explore how you can create a richer space capable of conveying more information without overwhelming people, by being intentional about the design.

The year is 1981. Your IBM PC is equipped with a tragic speaker that sounds awful for anything except occasional beeps. (Those beeps sound awful, too.)

You can’t afford a sound card and besides, sound cards for your PC have not been invented yet. You can’t even afford a floppy drive, so you’re one of the rare people who actually uses an audio cassette player as a storage device – a technique usually reserved for more primitive machines that have half the bits your new PC does.

But there’s a silver lining. Your cassette player has a little relay that controls its motor. You can engage and disengage the relay at will.

So, someone figured out that toggling the relay kind of sounds like a metronome. Like percussion. It’s a hack, but in the sonic landscape inhabited solely by your sorry speaker, it’s a breath of fresh air (scroll to 7:26 if you don’t land there automatically):

The year is 2026. Your computer itself is the size of an audio cassette, fits in your pocket, has better storage, graphics, sound, pretty much everything compared to a 1981 PC. It even has a special haptic motor. Except, that motor can only be controlled by native apps, and there is no official API to do it from a browser.

But there’s a silver lining. Tapping any checkbox on a site generates a haptic pulse. And that apparently works even if the checkbox is hidden and if the computer is doing the tapping.

So, someone figured out a way to use that to build a library that gives websites powers to provide haptic feedback. It’s a hack, but damn if it’s not one someone took to its logical conclusion.

I love these kinds of hacks, and I wonder what’s going to happen to this one. Will it fly under a radar, or will some websites start abusing it? If so, will Safari clamp it down, or will it actually give people a proper API for haptics?

I have recently stumbled upon two websites that try to do something interesting and inspiring when it comes to showing scale.

John Wallace’s Tangible Media Connection’s initial appearance might not feel very well-crafted, but jump to any page (for example this one) and it’s astonishing how great the photos of the objects are.

They’re great not just on their own (it’s really hard to photograph metals and plastics!), but also consistent with each other when it comes to angle, style, and – most importantly – scale. I am not sure if I have ever seen on online museum do this before. It’s very well worth checking out.

The other example is Neal Agarwal’s recent Size of Life. The whole website is delightful, with subtle music and sound effects, great handling of keyboard navigation and swiping, and so on. And the way it resizes objects and uses transitions to always keep you oriented is something a lot of other interfaces, even for productivity apps, could learn from.

Of course, now I wonder what the first website would feel like with the user interface of the second.

I found this weirdly delightful: There are a few new ringtones in iOS 26, but they’re not new new ringtones – they’re sort of “riffs,” or maybe remixes of a default Reflection ringtone.

If you don’t have an iPhone, here’s a short video where you can hear them. I’m guessing Apple sees the default ringtone as sort of an audio brand, and wants to invest in it more.

The only thing I don’t like are those names: It feels each one is following a different naming scheme.

(Side note: I am 180° from Gruber’s take on new Apple TV sonic logo. The previous one was better – recognizable and interesting. The new one is bland, milquetoast even. It instantly reminded me of the Windows 95 startup sound.)