picomac

Anyone who loves clever iOS games knows the drill. An indie developer comes up with a genuinely clever idea, suited perfectly to mobile gaming. They release it to great acclaim. And then the vultures sweep in with poorly-made clones or outright frauds. It's a dark and predictable pattern in the App Store economy.

Well, for once the tables have turned. For the past few months on the New York subway, I've seen people playing a mesmerizing game with a novel mechanic. It looked like a Breakout- or Arkanoid-style game, except each block requires multiple hits to clear, and instead of controlling a paddle you aim a railgun-like barrage of balls. It really drew my eye, but the longer I looked, the more obvious it became that it was poorly designed, full of ads, and generally low-quality. I decided to ignore it.

Then, a couple weeks ago, I saw Max Temkin tweet that he'd been testing a new game, Holedown. I clicked through and saw a ball-bouncer in the same genre. But it looked polished. It had beautiful typography, quirky anthropomorphic balls, and proclaimed to be "a ball-bouncer with depth". Instead of hitting plain colored squares, your targets look like rounded-off Tetris pieces. It costs $3.99, paid upfront, no ads, no garbage. Ten years into the App Store, it may be the first reverse clone.

Holedown delivers all the dopamine hits that I wanted from whatever that subway game is. (I couldn't even find it on the App Store, since I don't know its proper name and search terms like "ball bouncer" got me nowhere.) I'll admit that I was initially confused by the gameplay. Blocks move up the screen every turn — hit the top and it's a "crash" and game over. But my games kept ending after just a few shots, without crashing. I must have missed the tutorial text that I needed to collect crystals and upgrade my total shot count to keep playing further. Once this clicked, I was on a mad crystal quest.

Upgrades unlock more shots, deeper levels, and longer ball barrages. The game ultimately tops out in a sort of endless mode where you can earn up to 99 balls per shot and dig for an hour or more on end. (Or, if you're doing it the healthy way, set a Screen Time limit and resume where you left off.) You can also play the earlier levels, aiming to beat them in as few shots as possible. The simple decision to use roundrects instead of square blocks adds strategy — your best shots are by hitting an exact bank off a corner to get the balls pinging back and forth sideways, multiplying their effect. You can also squeeze a shot between blocks that are diagonal from each other and watch chaos ensue. And to keep things moving along, the balls automatically accelerate the longer they remain in play.

Throughout, there's a delightful soundtrack, but even more pleasing are the madcap plunking patterns of the bouncing balls themselves. One of my only quibbles is that while there's an option to turn off music, there's no separate switch for sound effects. That means that you can't play Holedown and listen to a podcast or your own music through headphones. Hopefully a 1.1 release will add that ability.

Regardless of whether you've seen its inferior inspiration, I strongly suggest giving Holedown a try. If the 30-second App Store preview isn't enough to convince you, or you want some advanced strategy tips, I'm going to be doing something new. Monday, August 6 at 7PM EDT, I'll be streaming Holedown live on Twitch. Watch for a tweet on @_picomac when I go live, or follow the link on this episode page. I hope you'll be able to tune in, and until then, happy bouncing!

The 2018 WWDC keynote brought tons of interesting updates, big and small, to all of Apple’s major OS platforms. (Well, except tvOS. Poor tvOS.) It also was the seventh such keynote since the death of Steve Jobs in 2011. Often, after these events, people who are disappointed in the outcome muse about which of their wishes would have been fulfilled if he was still alive. It’s a particular brand of resigned pessimism, and it rarely does any good.

So don’t worry, I won’t be engaging in those counterfactuals. And I expect very few others will, largely because there is so little disappointment following this year’s keynote. Nearly everyone got something they wanted, from iOS automation to UIKit on the Mac to a more independent Watch. Instead, I’ll ask: what did we get only because of the composition of Apple’s leadership today? In other words, what 2018 OS features might not have survived Steve’s absolute veto power?

I think the best place to observe this is in macOS Mojave, designed under the leadership of Craig Federighi. The (literally) biggest statement Craig made this year was the ten-thousand-point “No.” in response to whether macOS and iOS will be merged. Others have countered “Haha, but all the Mac apps will be iOS ports using Marzipan!” These retorts ignore the significant changes being made to the most central Mac app of them all — the one that is always running — the Finder.

Significant time in the onstage demo was given to Gallery view. It will be a great step forward, replacing the least useful (and likely least used) view in the Finder. Gallery view’s introduction immediately drew comments that “Cover Flow is back!” I understand why Cover Flow has been forgotten, but it’s not gone. In fact, it’s been in every version of the Finder from Leopard to High Sierra.

The question is: how and why did Cover Flow last this long? My theory is that it was protected by a double layer of nostalgia. Judging from his demos of it, Cover Flow was one of Steve Jobs’ favorite features of the iPod era. And in turn, he loved it so much because of the nostalgic feeling of flipping through racks of vinyl LPs, marveling at their high-quality artwork. Letting it go — even to be replaced with a more modern interface incorporating editing and automation tools — is one more admission that those eras are gone forever.

Fortunately, Mojave’s new features aren’t all just cold-hearted progress. Once the first beta was seeded, Guilherme Rambo noticed that beyond light and dark modes in Mojave, there are now customizable accent colors. This bit of whimsy is very similar to an appearance option from Mac OS 8, and breaks the 17-year dominance of blue and graphite over the Mac’s appearance. I feel like the enforced conformity of One True Appearance was a Jobs-ism; after all, High Tech, Gizmo, and the other wild window appearances set for Mac OS 8 were drastically scaled back immediately after his return to Apple. And while macOS will almost certainly never again allow the delightful insanity that Kaleidoscope schemes provided in the Classic era, this is another small area where letting go is welcome.

Ultimately, we can never know what might have been, but only what has come to pass. Even with privileged access to the inner workings of Craig Federighi’s team at Apple, it would be impossible to pinpoint how these features came to be, and what hurdles — real or imagined — they overcame to come into existence. But one way or another we got them, and I see them as signs of a promising future for the Mac.

I’ve been without a weather app on my phone for almost a week now. On May 24, through some combination of laziness, spite, and GDPR, Weather Underground’s Storm app shut down.

The writing had been on the wall for iPhone X users for months, even before official warnings came. Long after every other app I use on a daily basis had pushed an iPhone X update, Storm remained letterboxed. But despite the ugliness, it was still the only app that put all the weather data that I wanted in one place, and it stayed on my homescreen.

Then weird things began to happen. A minor release didn’t change the interface, but introduced a bug where users who’d paid to remove ads started seeing them again. The issue was resolved — at least eliminating third-party ads — but soon that slot began to plug “Storm Radar”, an app by a different developer but with very similar branding. You really want to use Storm Radar. It’s great. It’s by our new corporate overlords, The Weather Channel. You know, the people whose website used to be the best litmus test for the effectiveness of your ad blocker. Go on, try it, you’ll like it.

I did try it, and I can say with no reservations: Storm Radar is a bad app. It feels like a Windows app ported to iOS. Banner ads float in the middle of nowhere on the iPhone X screen (and the option to pay for their removal is gone). Its map tiles and color scheme are hideous. It has no 3D touch actions on its Springboard icon. And worst of all, it’s a one-trick pony. Granted, that trick — good predictive radar animations — is unmatched in any other iOS app. Except, of course, for old Storm, which separated past and future radar, avoiding the display and navigation bugs present in Storm Radar.

Having dismissed Storm Radar as Bad, I ignored it and continued using Storm. Until the dire warnings began. “Upgrade required!” (Remember, to an app with a similar icon, different name, and different developer — perhaps putting that message in violation of App Store rules.) Then “3 DAYS LEFT”, 2, 1, and 0. “This app is no longer supported.” It functioned for a few more hours, and then, it would seem, Weather Underground revoked their own API key so it displays blank maps, blank forecasts, blank graphs.

In one sense, this was a straightforward simplification of a bloated app catalog. After the merger, one company was managing three iPhone weather apps: Storm, Storm Radar, and Wunderground. They cut it to two, but only after creating the problem by developing Storm Radar in the first place. And they eliminated easy access to detailed data that weather nerds like myself love. I want to see the pressure in millibars. I want to see a graph of the dew point over the next week. And when there’s a tornado warning at 3AM and I’m debating whether I really need to go to the basement, I want to tap on a storm system and see the 0–3km vertical shear. (It was super high. I went to the basement.)

These features are now buggier and more fragmented than before. Storm Radar shows shear, but not reliably. WunderStation — yes, a fourth app, but only for iPad! — shows me the graphs I want. Nothing puts them all in one place as well as Storm already did.

And that’s the worst part: a company with great data has no incentive to deliver a good iOS experience. A parallel problem has arisen with Electron apps replacing native apps on Mac OS. And why should Weather Underground spend money developing great apps when the market price of an app is free? A lousy app or bad web-based experience or even nothing at all earns the same revenue. Weather Underground at least licenses their data through an API, with fees high enough that indie developers make it a paid option. But that passes the risk onto small companies that shouldn’t be shouldering it for a giant corporation.

So there’s some small hope that someone else will make the next great weather app, and charge me a fair price to send nearly all of that money along to Weather Channel HQ. But even more than when I discovered Storm a couple years ago, I wouldn’t bet on it. And that, rather than the missing icon on my homescreen, is the true loss.

It goes without saying that Siri has a lot of catching up to do. Anyone who hasn’t given up on it entirely will want to see significant improvements in iOS 12. I’m not too optimistic, so I’m asking for something smaller: just a little self-awareness.

I’m not talking about eliminating the dad jokes, though that would be wonderful. All I’m asking for is some simple connect-the-dots, making Siri not look oblivious when asked basic information about the device it’s running on. There are several pieces of information that are available onscreen in built-in, Apple-created apps, that Siri completely falls down when asked about.

If you’ve ever been to the Apple store for a repair on your iPhone, you know the first thing they ask is whether you’ve backed it up to iCloud. Genius Bar techs have memorized the four-tap sequence they need to access this information in Settings, but think how much time that wastes in the aggregate! Yet asking Siri, “When was my last iCloud backup?” produces one of three unhelpful responses.

“Everything you need to know about Apple products is at Apple’s website.”

What does that have to do with iCloud?

“Apple.com should be able to answer that question, and more.”

You know that icloud.com is a thing, right?

“Allow me to direct you to Apple’s rather fabulous website.”

For that one, they really punched up the smugness factor in the voice improvements made in iOS 11. And all of these are failures to parse the question. In fact, they are keying on the word iCloud and ignoring the rest. Just saying “iCloud” alone produces the same stock phrases.

I’ll grant that this request has some moderate degree of difficulty (remote data needs to be loaded) and limited use outside of a support context. But what about purely on-device data for one of iPhone’s flagship features: health and fitness? Don’t bother asking Siri…

“How many steps have I taken today?”

“I can’t answer that on your iPhone, but you can find it in the Health app.”

I am still mystified by this answer. My phone is unlocked. This data is 100% available to me. The steps are logged by the accelerometer in the phone itself (I don’t have an Apple Watch). And what possible definition of “on your iPhone” excludes the Health app which is on the homescreen of that phone? At least you can open the app by simply commanding, “Open Health.”

Even that didn’t used to be the case with the Home app. For years, most commands with the word “home” got routed to the HomeKit or Maps APIs, making it impossible to launch the app via Siri at all. I reproduced this as recently as a few days ago, but it seems to be fixed at the time of writing/recording. Well, unless you have another app installed that has a similar name.

“Open Home.”

“Which of these would you like? Home … or … Home – Smart Home Automation?”

“Home.”

“Just swipe up from the bottom of the screen to get home.”

Maddening. Don’t let Apple tell you that they’ve added support for follow-up questions to Siri. When the Siri prompt sounds, it’s actually starting again from zero. You can ask “What’s the weather?” in response to the follow-up, and it will go fetch the forecast. A simple problem that should be constrained to two choices is instead opened up to the entire universe of possible commands, inviting poor results.

So this is all I ask: Siri, please listen to yourself. Please learn about where you live. Do your homework, and we can talk again this summer.

WWDC 2018 is fast approaching and Apple-watchers everywhere are preparing their iOS feature wish lists. There’s plenty of low-hanging fruit available to make the iPad and iPhone better internet communicators and content-creation and content-consumption devices. But what about making the iPhone a better phone?

I know. “Who uses their phone as a phone?!” Like email, voice phone calls have become a bizarre, legacy, mostly inbound communication channel. As Google controversially pointed out recently, sometimes a phone call is the best or only way to find certain information or accomplish a task. But 90% or more of the calls to my iPhone are junk.

Apple must have realized that spam calls were of rising concern to their users, as they added call blocking capabilities in the CallKit API introduced in iOS 10. This led to an explosion of call-blocking apps in the App Store, some of them scammy in their own right, requiring access to all your contacts in order to function. A few are above-board, not needing your private data. One of them is WideProtect, which I’ve been using for several months now.

These mass blockers are useful for the latest pattern in spam calls here in the United States: spoofing a number similar to the one being dialed, in hopes that it will look “familiar” and get you to pick up. I started by blocking numbers that were extremely similar to mine, differing only in the last three digits. That didn’t stop enough of the spam, so I expanded the range to anything differing in the last 4 digits — 10,000 numbers in total. That seems like a lot, but I’m only using 1% of WideProtect’s blocking potential! CallKit allows each blocker extension to cover 100,000 numbers, but apps can have multiple extensions. WideProtect has 10 extensions for a maximum of 1 million total blocked numbers.

Blocking extensions are far more effective than my old tactic of putting repeat offender numbers into a single contact named “spammers”, but they don’t eliminate the annoyance. CallKit blocking prevents the phone from ringing, but voicemails still come through as notifications. These too are predictable: most are just 2 or 3 seconds of dead air, while others are a robot reading a script about senior care medical devices they want me to purchase. Needless to say, these are the most garbage-y notifications I receive on any given day.

So why can’t iOS eliminate them entirely? Smart handling of voicemail has always been part of the iPhone’s “revolutionary mobile phone” features (Visual Voicemail, as it was then called, was truly unheard of in 2008.) Today, iPhones analyze voicemails on-device to create transcriptions, which are available within notifications via 3D touch. Using that transcription data to suppress notifications could be dressed up as fancy machine learning, but really all that’s required is the type of basic pattern matching that has powered email spam filters for decades. And any good spam filter needs a safeguard against false positives, so offending voicemails could be sorted into the separate Blocked Messages view in the Phone app, rather than being blackholed entirely.

Perhaps the average iPhone user is accustomed to their device being a constantly buzzing annoyance box, and a couple spam calls per day are nothing amongst their dozens of _other_ spam notifications. But I try to keep those annoyances as close to zero as possible. So even though the phone is, most likely, the least important part of the iPhone, robust spam filtering would help it be the best phone available.

This is Alto’s Odyssey.

It’s been a couple months since the game’s release, and for some, it may be already forgotten. But many of the initial reviews placed the followup to Alto’s Adventure directly into the pantheon of iOS games. I agreed with those reviews then, and after dozens of hours of flipping, wingsuiting, grinding, and wallriding, I agree with them now.

At least, I agree with their conclusion. Many glowing reviews praised Odyssey for bringing unprecedented nuance and refinement to a “casual” game. But I would argue that these traits break Odyssey out of the casual category and into uncharted territory. Part of what keeps me playing Odyssey is the leaderboard, where I’ve bounced around the top twenty but never broken into the top ten. Sitting down to attempt that feat is not casual at all — a significant chunk of time needs to be set aside to aim for three, four, or five million points when they’re earned by the tens of thousands.

If Odyssey didn’t constantly delight, 40 minutes on the mountain would be a slog. I think anyone who views Alto as a one-button endless runner would feel that way. I’m not going to try to convince you that the balloons, birds, storms, and total solar eclipses make Odyssey the best one-button endless runner ever on iOS. No, that would be Tiny Wings. Why not Alto? Because it is, crucially, a two-button endless runner.

And there, if it lies anywhere, is Alto’s single flaw: the failure to reveal at the outset that the wingsuit and its control scheme is vital to the game. It’s all too easy for a truly casual player to entirely miss its importance, and thereby miss out on the incredible combination play that makes Alto unique and addictive.

Back when tvOS was launched — and Picomac was new! — Alto’s Adventure was hailed as the perfect game for the one-button Siri remote. There’s no doubt that the snowy and sandy environments look stunning on a big screen, but trying to toggle the wingsuit with a firm press is just a recipe for a crash. This is, of course, to Alto’s credit and the Apple TV’s detriment. The lament all along about Apple TV games has been that without a dedicated controller, they are forced to give up complexity and be shadows of what they can be on other platforms.

So if you still haven’t played Alto’s Odyssey, I cannot recommend it highly enough (at least on your iPhone or iPad). Or if you did give it a try and found it too simple, too casual, give it another go. And remember: save up all those coins for wingsuit upgrades. You won’t regret it.

A few weeks ago, David and Katie were kind enough to have me on Mac Power Users to talk (to the extent that's possible) about regular expressions. It's a massive topic and we barely scratched the surface, but having listened back to the segment, I think it gives Mac users a number of points of entry. Even so, there was one big topic on my outline that I forgot to cover: multiple matches.

As David said, the basic conceit of regex is "find and replace on steroids". I think even that description is limiting, since the basic model of find and replace is to find one thing and replace it with one other thing. But once you're proficient in the basic search syntax, you can go much further by learning the replacement syntax. (Frustratingly, this tends to vary from app to app. I'll describe things with BBEdit's syntax, which I'm most familiar with.)

If you want to match multiple components of your overall regular expression, all you have to do is enclose them in parentheses. Then, each of these can be referenced in order, left to right, in your replacement string by using \1, \2, \3, and so on — though it gets pretty unwieldy after 4 or 5. This can be used for mundane tasks, like reordering parts of a date string, but it's also endlessly flexible, especially because you can refer to a captured group more than once.

This past week at work I applied multiple matches to what would have otherwise been a tedious clicky-draggy spreadsheet task. I was setting up software for some computer-based tests, including mapping the number of questions answered correctly to a standard score. (Remember the SAT? It's out of 1600 points, but there aren't 1600 questions. Thankfully.)

I had an Excel spreadsheet that a coworker made, but the software required a CSV, with each row having the low raw score, high raw score, low scaled score, and high scaled score for different ranges. We had already run an analysis and decided these correspondences for every possible value, so I needed those pairs of columns to be duplicates of one another. Yes, I could do that in Excel or Numbers, duplicating columns and saving out CSV files via an Export dialog. Or I could make a much faster round trip to BBEdit.

The routine was simple: paste two spreadsheet columns into a text file, giving me numbers separated by a tab. I captured the first number, matched a tab, and captured the second number: (\d+)\t(\d+). Then, since I needed to both flip the column order and duplicate them, I replaced the string with \2,\2,\1,\1. And thanks to BBEdit making the "Replace all" command available even without opening the Find and Replace window, transforming new data only took one keystroke: ⌥⌘=. 15 repetitions of the process only took a couple minutes. (This was right on the threshold of "should I automate this?" If it was 100 repetitions, the answer would have been a definite "yes".)

Now, chances are you'll never have to perform this exact task. In fact, the chance that I'll have to do this again are near zero. But the chances that you'll have to rearrange some lines of text are pretty high. Even if you have to look up the syntax, just knowing that the apps you have on your Mac are up to the task will enable you to solve the problem and get the work done faster.

I've been crafting a vision of what the Mac Pro (Mid 2018) may look like ever since the line was declared definitely-not-dead in April. I don't have many pro needs; podcast production, light image editing, and the occasional Handbrake session are as intensive as my computing gets. Even so, I built my core idea around the one thing I struggled with when configuring an iMac: storage (especially the question of to NAS or not to NAS).

I'm sure Jony can do better than this.

There's a lot that's appealing about network attached storage. It offers redundancy, it's available from everywhere in the world, and it can be shared among devices while at home. But it has major downsides too. It's essentially a second computer and carries the price tag to match, even before you fill it with hard drives or SSDs. And because NASes like those from Drobo and especially Synology are computers, they run their own operating system…which isn't Mac OS. I don't care how lightweight and convenient it is; I want no part in managing a Linux box through a fake Windows GUI in a browser. Even if I took that leap, dealing with network drives still isn't a native Mac experience. (I see this frequently at work with our office-wide Samba share. You can't put files in the trash and Spotlight is nonexistent. It's no way to live.) So I configured my iMac with a 2TB Fusion Drive and hung a 4TB USB drive off the back.

Even though the drives I hooked up to my iMac are nowhere near full, I would love to see a device that integrates the mass storage of a NAS with the power and software of a high-end Mac — a Mac Pro-bo, if you will. I considered it a total pipe dream for a while. Remember, the last Mac to offer expandable, multiple drive storage was a 40-pound cheese grater from 2012. But the recent sneak peek at the iMac Pro actually gives me more hope than ever that we'll see something like it.

On the WWDC slides and on the iMac Pro website, there's a new schema for what Apple considers the pro setup, and it's a 5K display sitting next to a box full of hard drives. Given the current lineup, that involves an iMac display and a third-party box to the side. The converse — a trash can Mac Pro with a third-party display — proved less than viable. The big design question for the future Mac Pro is whether these two components, one centered around the display and one centered around storage, are the "modules" Apple referred to in the Mac Pro roundtable.

The other future of computing.

Swappable storage, backed by APFS and macOS would certainly be a unique product. Many assume that the Mac Pro has to be even more expensive than the iMac Pro, but Myke Hurley and Jason Snell pointed out on Upgrade that it might be best if the Mac Pro sat in the middle. Tim Cook's Apple loves to hit every price point, and there's a huge gulf between $2300 at the top of the iMac range and $5000. I imagine that $3500 could offer a lot in terms of CPU, RAM, and storage in a compact Mac — especially with no display and nothing but air in some storage bays. Max out the storage or add a putative $1500 Apple 5K display and the Mac Pro would be comparable in features and price to its all-in-one cousin, with room to grow. It's still just a dream, but maybe it's not so farfetched after all.

Picomac is back! I never intended to stay away from it forever, and there's just way too much good stuff from WWDC 2017 to talk about.

Rumors swirled before the WWDC keynote about whether Apple would bring major design changes to iOS 11. The answer is that Apple took the middle path: there's nothing like the major shift from iOS 6 to 7, nor will this year's iOS look exactly like last year's. There's also a clear direction of change. Nearly every new piece of interface is taking its cues from the 2016 edition of the iOS Music app. The theme of iOS's design is quickly moving to unify around the theme of bold header typography, rounded cards, and gentle translucency effects.

The App Store has gotten the Apple Music treatment in iOS 11.

But those are the terms we have to talk about it in: directions, themes, specific interface elements. Apple hasn't decided to craft a brand around its software design or give it a name. (Though perhaps that's for the best, given the dubious names that came out of this keynote.) As of earlier this spring, Apple is the only major operating system provider who doesn't have a branded design language. Google has been refining its Material Design since 2014, and Microsoft unveiled the Fluent Design System with Windows 10 in May. Beyond providing a quick way to describe these interfaces, they offer connotations, just like a good brand should. "Material" indicates touch and a tactile experience; "Fluent" shares a root with "fluid" and "fluency", getting at both smoothness and ease of use.

So the question is: does Apple need a branded design language to go up against Google and Microsoft? Need is a strong word, of course; iOS's interface can certainly match and surpass its competitors without a catchy name. To paraphrase Steve Jobs, design isn't what it's called; it's how it works. And what would Apple call what we can only call "the Apple Music look" or "iOS 11's design"?

The closest they've come to coining an interface term recently is "vibrancy" (and it's confusing evil twin, "dark vibrancy"). In truth, this effect is a major part of the Mac and iOS aesthetic, so Vibrant Design would be an apt name; it also fits with the splashy color and heavy font choices that are in favor now.

Another option would be to go abstract. A "California" design language would evoke the phrase "Designed by Apple in California" present on all their hardware and packaging. It would also offer more flexibility than a descriptive name. After all, interface design evolves at a rapid pace and is subject to trends in fashion. Today's fresh and vibrant design could easily be as passé as skeuomorphism and linen in a few years. A rebrand would always be possible, but that takes extra effort. So perhaps the best approach is for Apple to push its interface design along as they see fit…and to let the rest of us worry about what to call it.

Since the MacBook Pro with Touch Bar was announced a couple weeks ago, there's been plenty of speculation on why exactly it exists and what role it will serve in the Mac lineup of the future. The same kind of questions arose when the Force Touch trackpad was added to the MacBook Pro a couple years ago. There, the answer to "why this?" seemed fairly simple: as a non-moving part it eliminated internal space and made better use of its ever-increasing surface area, since an equal amount of force registers a click on any part of the trackpad.

The answer for the Touch Bar isn't so obviously pragmatic. It doesn't save space, as it occupies the exact same area as its predecessor function keys; the low end model that retains those keys reflects that exactly. And since the machines still aren't even in the hands of reviewers, the jury is still out as to whether the touch interface features will make the Mac more of a joy to use. Nobody knows if the Touch Bar will be liberated from the MacBook enclosure and put in standalone input devices, like when the Force Touch trackpad became the Magic Trackpad 2. (And unlike the Magic Trackpad, given the components involved, it seems nearly certain that a standalone Magic Keyboard 2 with Touch Bar will not be a $0 build-to-order option with the iMac of the future.)

Nevertheless, I think that Touch Bar technology will spread to all Macs, even if it comes at a cost to consumers. The reason is the problem that I believe served as the entire genesis of the Touch Bar technology: someone inside Apple looked at the glyphs on their function row keys and said "why do we keep letting these become obsolete?"

There was once a time when that wasn't a concern; Apple's desktop and laptop keyboards just had plain, numbered function keys. This was even the case when Apple started assigning essential features to them, like when Exposé was introduced in Panther. Its features were originally mapped to the blank F9, F10, and F11 keys, and I learned to use them by feel on my 2003 PowerBook G4. (It was super easy, as the function keys were physically grouped into F1–4, F5–8, and F9–11.) But the next wave of Apple hardware brought dedicated media playback keys in the function row, and they landed on the only glyphless function keys available: F9–11.

Nothing to see here, move along.

Obviously my keys didn't change overnight, but their associated actions did. Software "broke" perfectly good hardware. And other mismatches constantly arose. Hardware keyboards could easily outlive the features printed on them. Until just last year I was still using a keyboard that had a Dashboard key on it (yes, Dashboard is still technically alive, but just barely). And uniformity is impossible, as today's laptop keyboards with backlight controls put them on F5 and F6, while Apple's standalone keyboards — designed to work with any Mac — just offer blank space there. "This keyboard doesn't have backlighting, so, I dunno, have a couple extra buttons." It's inefficient design.

So rather than scrap the function keys as hopelessly broken, Apple is trying to turn a weakness into a strength. The function keys keep changing every couple years, and it's a problem. Instead, make them change every couple seconds and it's a delightful solution. Maybe someday the Touch Bar will vanish into a fully interactive keyboard. After all, the original Mac keyboard had no function row and neither does the iPad Smart Keyboard. iOS goes as far as dropping number keys, hiding them in a separate mode. I can't predict the future of Apple's input devices beyond the Touch Bar, but it seems a safe bet that they'll be capable of changing — quickly.