Matt Birchler recently did a nice YouTube video praising the open source tile manager AeroSpace. There is a lot to love about AeroSpace right from the get-go. While I definitely wouldn’t call it Mac-assed, since AeroSpace is for very advanced users and developers who are comfortable with text configuration files, my sense is that AeroSpace is made by people who care deeply about the Mac. Part of this is how Nikita Bobko and others supporting the project have deeply considered how AeroSpace works. Unlike Stage Manager, AeroSpace has a clearly defined lexicon which is deployed across copious documentation. Like Stage Manager tries to ease window management with more automatic windowing, AeroSpace tries to ease tile management with more automatic tiling. Users don’t need to manually arrange windows into a grid with AeroSpace. It just happens. From there users can configure and tweak how windows are automatically tiled to their heart’s content.

The idea of composing sets of apps into custom workspaces is particularly appealing to me. I find super apps (apps that are effectively multiple apps in one) mildly off-putting. Most IDEs are super apps for everything related to software development. They contain version control managers, terminals, text editors, and so on. While many IDEs do all of these things reasonably well, their super app paradigm is effectively a usability barrier to using other apps that might otherwise be better. Instead of using Visual Studio Code, for example, I could imagine a world where I have a coding workspace consisting of BBEdit, Terminal, and Git Tower. The added benefit of this sort of multi-app workspace is that I could still use the individual apps a la carte or mix in other apps as needed.

While I’m sure many people started using tile managers to build custom workspaces, I suspect many more turned to them as a way to address desktop clutter. I’ve written a couple of times already about the modern day problem of desktop clutter. Thanks to an abundance of resources (memory, processor speed, etc…), people can open more and more stuff (apps, windows, tabs) without slowing their computer. Because their computer never slows, there is no intrinsic mechanism that forces users to ever close anything and so their more and more stuff stays open until they get overwhelmed. Tile managers reduce desktop clutter by discouraging or preventing overlapping windows, which physically constrains the number of windows that can be visibly open at a given moment.

Maximally constrained user interfaces are impossible to clutter. Lots of people were drawn to the original iPad specifically because it really was just a big iPhone and they loved their iPhone in part because it was too simple to mess up. I get it. I prefer small hotel rooms when traveling solo because larger ones just give me more space to lose track of my stuff, but small hotel rooms are not without trade-offs. The tiny room at the boutique hotel I stay at when visiting my employer’s New York office isn’t really suitable for anything beyond sleeping and hygiene. Even working on a laptop for any extended period of time would be a challenge. A hotel room that is too small to do work is great until you want to do work. An iPad that doesn’t feel like a computer is also great until you want to do computer-y things with it.

Desktop tiling managers are definitely not maximally constrained like the original iPad nor are they even anywhere near as constrained as previous versions of iPadOS with split view and slide over¹, but they are, by their very nature, constrained. Beyond physically constraining the number of windows visible at a given time, tiling managers also constrain the shapes of those windows. I wrote² about this when reviewing Stage Manager on macOS Ventura:

Personally, I’ve found gridded UIs to be lacking ever since I first saw Windows 1.0. When using grids, apps’ sizes. and more importantly their aspect ratios, are dictated by other windows that are on screen. Say you want to work across a spreadsheet, word processor, and a browser. Not only do all of these apps benefit from a large amount of screen real estate, both the word processor and browser need a minimum amount of size just to be entirely visible. In a gridded UI, some or all apps would have to compromise on space and you would have to scroll constantly within each app to see the entirety of what’s being worked on.

People who use tile managers undoubtedly have strategies for mitigating this inelegance. Tiles don’t have to be even subdivisions of a given display so you can, for example, adjust the width of a word processing tile to that of its document. AeroSpace in particular seems to offer lots of tools for technical users to hone their tiled workspaces. That said, the very nature of tiling according to the size of the display limits what adjustments are even possible.

Part of me feels bad that I used AeroSpace as the jumping off point to argue against tile managers. Its makers clearly have put a praiseworthy of thought and care into how it works, but it was seeing such a well considered tile manager that solidified my thinking. AeroSpace is the most appealing tile manager I’ve seen on the Mac and while I’m certain there are plenty of workflows where AeroSpace shines, being physically constrained by an always on tile manager that dictated the number and shape of open windows would feel like a gastric band to me. Rather than wholly automatic systems like AeroSpace or Stage Manager, the best solution to desktop clutter for me remains to regularly close the stuff I open, that’s only just a little automatic.

In olden days computers had just two emotions. They either happily worked as expected or were too sad to boot. Computers today have a range of emotions, but tragically have no way to express them. That’s why our scientists developed MacMoji using the latest in sticker technology, so your favorite computers can finally convey exactly how they feel.

As a parent working a full time job, I regularly seek out creative outlets that I can manage in my limited spare time. MacMoji started out as one such outlet. The idea of combining more modern emoji with the classic startup icon was too fun to resist. I could gradually illustrate one or two, share them on the Relay Member Discord, and iterate as needed based on feedback. At some point, a Relay member suggested I turn these illustrations into an Apple Messages sticker pack. The idea was such a no brainer that I did just that…eventually. You can now buy the Sticker pack for just $0.99 over at the App Store. You’ll find the F.A.Q. over here, which addresses why these stickers aren’t available in the EU. My thanks to the Relay member community for their feedback and encouragement in creating these stickers.

There are two arguments some use to try and diminish Steve Jobs contribution to the Macintosh, and by extension all of desktop computing. The first and by far most common is to say that Jobs merely copied what he saw at Xerox Parc. While there is absolutely no doubt both the Macintosh and NeXT grew out of what Steve saw, he even said as much, the system at Parc was akin to an automobile before the Model T. It was unrefined, complicated, and not user friendly. This is why Microsoft copied mostly from the Macintosh (and later NeXTStep) rather than anything from Xerox to make Windows.

The second and more obscure argument is that Jobs merely took over the Macintosh project from Jef Raskin, the suggestion being that Raskin invented the computer and that Jobs swooped in to take credit at the last second. What that argument omits is that Raskin’s vision for the Macintosh was very different than what shipped. How different? Raskin didn’t want a mouse. Here’s Andy Hertzfeld over at the venerable Folklore.org:

He was dead set against the mouse as well, preferring dedicated meta-keys to do the pointing. He became increasingly alienated from the team, eventually leaving entirely in the summer of 1981, when we were still just getting started, and the final product [utilized] very few of the ideas in the Book of Macintosh.

We know this is true not just because of Hertzfeld’s own account, but also because Raskin did eventually get to release his computer in 1987, the Canon Cat. Sure enough, it indeed didn’t use mouse and instead relied on what were called “leap keys”. Cameron Kaiser recently went into detail of how the Cat worked.

Getting around with the Cat requires knowing which keys do what, though once you’ve learned that, they never change. To enter text, just type. There are no cursor keys and no mouse; all motion is by leaping—that is, holding down either LEAP key and typing something to search for. Single taps of either LEAP key “creep” you forward or back by a single character.

Special control sequences are executed by holding down USE FRONT and pressing one of the keys marked with a blue function (like we did for the setup menu). The most important of these is USE FRONT-HELP (the N key), which explains errors when the Cat “beeps” (here, flashes its screen), or if you release the N key but keep USE FRONT down, you can press another key to find out what it does.

Needless to say, the Cat wasn’t the huge success Raskin hoped it would be.

The default of any physical space is clutter, in that keeping things tidy requires persistent concerted effort. People who succeed at sustained tidiness rely on systems, habits, and routines to reduce that effort. Disposing a single delivery box, for example, is much easier when a single process is defined for all delivery boxes. Even if the physical effort of breaking down and moving the box is largely the same, the mental effort is reduced to nothing because the decision of what to do with the box has already been made. In that sense, reducing cognitive effort ties directly to reducing physical clutter, which in turn reduces cognitive clutter.

Digital spaces are no different than physical ones. Their default is also clutter. Just look at most people’s photo and music libraries. The difference is that digital clutter is much easier to ignore. You can try to ignore the delivery boxes stacking up around the foyer, but their growing hindrance to day-to-day tasks is obvious. Digital clutter doesn’t take up physical space so most of it can remain out of site and out of mind. You only deal with a cluttered music library on the occasion you make a playlist. There is however digital clutter that does hinder people’s day to day — their desktops. Windows (and tabs) can very easily stack up like empty boxes in the foyer to the point where they constantly get in the way. I wrote about this when reviewing Stage Manager in macOS Ventura.

Windowed interfaces, like those found in macOS and Microsoft Windows have historically been manual. The user opens, arranges, closes, minimizes and hides windows in whatever manner that suits their needs. When Mac OS and Windows came of age in the 80s and 90s, computers were only powerful enough to do a few things at once. These limited resources meant a given task typically involved launching a few apps, manually managing their small number of windows, then closing everything before starting the next task… I find managing a small number of windows more satisfying than burdensome. Users on today’s computers can easily amass dozens of windows from a variety of apps. Furthermore, these apps and windows persist, even between reboots. There is no intrinsic impetus that forces users to quit latent apps or close latent windows. Manual windowed interfaces became cognitively burdensome when faced with unlimited persistent windows found in modern desktop computers. While some still find them delightful, more and more people find desktop computers harder and more annoying.

Stage Manager on macOS tries to solve the problem by automating which windows are visible at a given moment. Even though my review of Stage Manager was on the positive side, it was ultimately too finicky for me. I love the concept of sets, just not enough to manually maintain them. It’s the same problem I have with Spaces. Lots of people use Stage Manager and Spaces as tools to organize and streamline their workspaces, but for me, these sorts of virtual desktops simply become mechanisms to have more windows. They facilitate clutter by hiding it rather than reduce it.

As it turns out, the best solution to window clutter for me is not some extra layer of window management. It’s less windows. I even said as much in that very quote from a review I wrote three years ago.

I find managing a small number of windows more satisfying than burdensome.

And yet it wasn’t until this summer that I actually changed my habits, so what took so long?

As a middle aged man who works a full time job and is actively involved with parenting… well, let’s just say I am less adept at identifying when and how I should change my habits. After all, a lot of my habits at this point are exactly the kind that help me minimize effort. Beyond that though, the only option built into macOS for quickly quitting out of apps is to log off with “Reopen windows when logging back in” unchecked, which doesn’t quite work the way I want it to. There are a handful of apps I always want running and don’t want to have to re-open whenever I resume using the computer. These apps could be added to login items, but I also dislike windowed apps launching automatically. They can be slow, demand extra attention through various prompts, and steal focus. Yuck. What I really wanted was to quit out of all but a handful of apps before locking the screen so that I could start instantly and with a clean slate the next time I use the Mac.

Once again, AppleScript to the rescue¹. Using AppleScript, I could set a whitelist of apps to keep open, and then quit out of everything else². Shortcuts then let me chain this script with other actions to confirm my intentions before locking the screen. Finally, I was able to add the shortcut to my Stream Deck so now at the end of my work day, I push the “Off Duty” button. Even when I have to manually address apps with unsaved documents, quitting apps in one fell swoop still greatly reduces decision making because I no longer have to individually consider whether to quit a given app. It’s going to be quit the same as the rest so all I have to decide is where I should save the open documents, which in itself compels a good end-of-workday habit that I should have been doing already. When I start work the next day, the previous day’s work is saved and my Mac is effectively reset with just a handful of apps and windows open.

Having used this automation throughout the summer, I can now say with confidence that managing windows and tabs in macOS is once again truly satisfying. Navigating between apps doesn’t feel like work anymore and features that never appealed to me with dozens of windows and tabs now make sense. I can find that one window using Mission Control. I actually use command-[number] to jump to a specific tab in Safari! By reducing the cognitive effort involved with quitting apps, I have reduced desktop clutter, which in turn has reduced cognitive clutter to the point where my Mac is once again a tool that helps me focus because it’s no longer like a foyer full of boxes I have to carefully sift through, but an extension of what is currently on my mind.

Those who have been following the rollout of Apple’s new Liquid Glass theme accuse Alan Dye and his team of designing user interfaces that look good in marketing materials at the expense of usability. That’s a fair criticism, but I don’t think “marketing” is the right way to frame it. In my mind, marketing interfaces are a separate issue. They are designed to push users to do something they wouldn’t otherwise. Liquid Glass hamfistedly just tries to look cool.

Looking cool isn’t a bad priority for an interface and it’s certainly a way better priority than marketing. Interfaces built for marketing necessarily come at the expense of usability because their priorities typically come in conflict with those of users. Streaming services are the best example of this, where promoted shows are given priority over those already in progress. Cool looking user interfaces, on the other hand, aren’t inherently at odds with users. iPhone OS looked cool and was immensely usable, and I would argue even Aqua was still very usable even before the transparency and pinstripes were rightfully toned down.

“Marketing UI” is an unfair term for something like Liquid Glass. Trying to look cool at the expense of usability is bad, but it’s way less egregious than actively interfering with users. A better term, in my mind, is “Hollywood UI”. Hollywood has long given computers made up user interfaces, some of them very cool, others not so much. Regardless of their coolness, Hollywood UIs can look like anything because they are ultimately just another prop or set piece. They don’t actually have to work.

That Liquid Glass looks cool in marketing and elsewhere isn’t really the problem. iPhone OS and Aqua looked good too. The problem is that Alan Dye and his team seem more interested in making interfaces that merely look good rather than those that can survive contact with the real world, probably because designing props is a helluva a lot easier and more fun than designing tools that actually work.

Seems like Microsoft is still migrating features from the old Windows Control Panel to its newer Settings app. Here’s Sean Hollister, at The Verge:

But the Control Panel still can’t die. The latest features to migrate, as of today’s Technical Preview: clock settings; time servers; formatting for time, number, and currency; UTF-8 language support toggle, keyboard character repeat delay, and cursor blink rate.

While it is indeed hilarious that Microsoft is still migrating stuff out of Control Panel to Settings over a decade later, my gut sense is that Windows 11 has had to pay down technical debts the same way people have to gradually pay down their financial ones, in installments that span multiple years.

The Windows 11 PC that I use for gaming and Plex isn’t my daily driver. Because I don’t use Windows for either work or other personal needs, take what you’re about to read with a huge grain of salt. That said and in my limited experience, Windows 11 has been gradually getting noticeably better, and dare I say, nicer¹? Windows Settings is nicer than Control Panel. The new right-click menu is nicer than the old one. Part of me wonders why Microsoft has been so gradual with these rewrites rather than just releasing them in a more done state, but then again modernizing decades old components is never easy, especially when you’re trying to satisfy software compatibility and entrenched IT practices.

Taking such a long time to revamp these components does merit some teasing and probably some criticism, but I think keeping at it for over a decade shows a resolve that is also worthy of praise. The effort gives me confidence that the people in charge in Redmond truly care about improving the user experience of their desktop OS. I wish I could say the same about the people in charge in Cupertino.