Over at Pixel Envy, Nick Heer wrote the perfect sequel to my earlier post about the years of softball coverage of Windows on ARM. Whereas I focused on what people were writing about Windows on ARM, Nick smartly looked at what folks were speculating with ARM-based Macs.

Wellborn’s selection of quotes from enthusiastic press coverage of Microsoft’s lukewarm ARM efforts reminded me to go look for some reactions to the early rumours and the announcement that Apple would be switching to its own processors. I want to do this not just because these things are funny to read in hindsight, but also because they illustrate why media and analyst coverage often gets this stuff wrong in the first place — especially when it comes to Apple.

What I still can’t get over is the numerous takes that framed Microsoft’s failed ARM efforts as bad for Apple. The most common of which was effectively “ARM-based Macs might suck because Windows on ARM sucks.” Daring Fireball’s John Gruber took the high road when commenting on one of the pieces I mentioned from The Verge that used Windows on ARM as a cautionary tale for ARM-based Macs.

As with my previous item linking back to Rik Myslewski’s 2008 take on Apple’s acquisition of P.A. Semi, I am not trying to dunk on Dieter Bohn here. 18 months ago, these were all perfectly reasonable concerns.

I don’t think Dieter deserves to be dunked on for suggesting ARM-based Macs could be slower or have other trade-offs. I too wrongly speculated that may be the case when emulating x86. What I think merits dunking is suggesting that Windows on ARM could be used as some sort of model for what could happen with ARM-based Macs given how much faster Apple’s chips were compared to Qualcomm’s and Apple’s proven track record of migrating architectures. Everyone knew Microsoft had slower chips and everyone knew Windows had really never meaningfully run on any other architecture than x86.

While I think it’s gone better than anyone imagined thanks to the insanely fast performance of Apple Silicon, the fact is that the Mac’s transition to ARM was always going to be entirely different than that of Windows, and that was obvious from the get-go.

Tim Sweeney spoke at something called the “Global Conference for Mobile Application Ecosystem Fairness”. According to MacRumors, here’s what he said:

What the world really needs now is a single store that works with all platforms. Right now software ownership is fragmented between the iOS App Store, the Android Google Play marketplace, different stores on Xbox, PlayStation, and Nintendo Switch, and then Microsoft Store and the Mac App Store.

Sweeney isn’t just talking about Apple and Google there, he’s talking about all console makers. My bet all along has been that Tim Sweeney wants the Epic Games Store to become the Amazon of software, and that ambition necessarily goes beyond Android and iOS.

I am writing this on a new 16″ MacBook Pro, which has an ARM-based M1 Pro for its CPU. It was delivered this past Tuesday, October 26th, 2021, almost exactly 16 months after Apple announced the transition away from using x86-based processers built by Intel. By most accounts Apple’s ARM-based M1 Macs are the fastest laptops money can buy. Even running software compiled for x86 seems to just work.

Rather than add yet another review to the pile, I wanted to instead look at how the other traditional desktop platform added support for ARM-based chips, Windows. Specifically, I wanted to see how Microsoft’s effort has been covered. Most of the links and excerpts are from The Verge, since they seem to have more Microsoft coverage than anyone in the last five years.

When Windows on ARM was announced at the tail end of 2016, Tom Warren had a largely positive take:

Microsoft is demonstrating its desktop apps on ARM capability with Adobe’s Photoshop software today, but any of the millions of desktop apps will just work according to the company. “I think people are going to have to experience the devices for themselves,” explains Myerson. “I think that for many people it’s going to be a very delightful experience.”

Unlike Windows RT, which had a desktop mode that didn’t run apps, ARM on Windows 10 will look and feel just like regular Windows 10, and also work mostly the same way. That means the laptops that arrive next year won’t be confusing, and will bring impressive battery life that we haven’t seen in Intel-powered laptops. It also means that phones could eventually support full desktop apps, enabling Microsoft’s PC as a phone Continuum concept to become a lot more powerful.

The rosily mentioned “Continuum concept” involves Windows Phone 10, a soon-to-be-discontinued platform that had less than 1% marketshare when this article was written.

Anyhow, flash-forward to April of 2017 when Tom Warren reports that ARM-powered Windows laptops won’t arrive until Q4, but still reiterates Microsoft’s plan to support “traditional desktop applications”.

Microsoft first announced its plans to bring Windows desktop apps to mobile ARM processors last year. The software giant is planning to allow partners to build laptop devices first, with a version of Windows 10 that will support ARM chips directly by including an emulator in the operating system. Devices will be able to run traditional desktop applications like Chrome or Photoshop, and they’ll be based on Qualcomm’s Snapdragon 835 chipset.

While many were hoping we would start to see ARM-powered Windows laptops soon, Qualcomm has revealed they won’t arrive until the end of the year.

Well, it wasn’t until March of 2018 that Dan Seifert got to write about using an ARM laptop running Windows, but don’t call it a review.

The model I’ve been using won’t be sold in the US and has more RAM and storage than what we’ll be able to buy, so this isn’t a review of the device itself, but more of a look at how the platform that will run on various devices from HP, Lenovo, and others.

This non-review says Windows ARM shows a lot of promise and starts off by stating that the core of Windows 10 seems to perform well, and lists off various benefits of ARM, LTE support, better battery, etc… Seifert waits until the end of the article to point out some limitations, like web browsing…

Performance in Chrome is rather bad, with sluggish load times, stuttery scrolling, and slow transfers between tabs. You’ll have a much better time sticking with Microsoft’s Edge browser, but that brings its own issues, namely poor compatibility with certain websites and a tendency to get overwhelmed after a few hours of use, requiring a restart of the browser.

…and app compatibility.

I did download and run Adobe Photoshop for laughs; it works about as you’d expect it to: slowly. The limitation against x64 apps means that virtually any modern game will not install on this computer, not that it’d run very well on this hardware if it did install.

The first actual review of a Windows ARM laptop that one could presumably buy doesn’t appear until six months later, almost two years after the original announcement. It’s here that Dan Seifert wastes no time getting to the reality of the situation.

On paper, the Book 2 is the Surface Pro I’ve been looking for: an always-connected, ultraportable computer that will last more than a full day away from an outlet. But, as is often the case, what’s on paper doesn’t always line up with reality.

But even this very critical review leaves the reader with this strange caveat toward the end.

Those who only have to do a handful of tasks for their work might not have an issue with the Book 2’s performance, but anyone that juggles a lot of web browsing with other productivity apps will likely hit the Book 2’s wall rather quickly.

Y’know, all those people who don’t use their computer to browse the web or use apps.

Anyhow, in October of 2019 Microsoft announced it’s own Windows ARM device, The Surface Pro X. Chaim Gartenberg’s coverage dedicates a whole sentence of skepticism before citing Microsoft’s power claims.:

The biggest question remains whether the SQ1 will be able to offer comparable power to a traditional x86 processor, though. Microsoft is making a big deal of the power it’s managed to get out of the SQ1 (and the GPU, which it also worked on together with Qualcomm). The SQ1 is a 7 watt processor (compared to the 2 watt output that Microsoft says most ARM processors offer), and the company says that it’s stretching that power as far as it can, getting “three times more performance per watt than the Surface Pro 6.”

Dieter Bohn’s review came one month later.

Not that my expectations were very high — other ARM laptops have been dog slow. But the core of Windows 10 runs just fine for me. It’s the full version of Windows 10, by the way, not some RT or S version. At the extreme, I have had several apps open — including two different browsers with a dozen or so tabs open in each — and nothing ground to a halt.

[Chrome] runs on the Surface Pro X, but I wouldn’t say it sprints. It is discernibly slower than the 64-bit ARM version of Edge on this computer — that’s been true even on Intel computers, but it’s a little slower here. Still, totally usable.

Not that the bar is low or anything.

Microsoft updated the Surface Pro X in October of 2020, just before Apple released the first M1 Macs. Tom Warren gave it a largely positive review.

In fact, this new Surface Pro X blasts past the Windows Hello facial recognition screen and straight into the Windows desktop with ease. Microsoft says it hasn’t made specific changes to Windows Hello, but everything feels a lot smoother to me.

Brownie points for logging in fast?

Spotify brought the Surface Pro X to its knees last year, but I’ve noticed it now launches without killing the rest of the system. It still takes a while longer than I’d expect to render album art, but it’s greatly improved over what I was experiencing on the original Surface Pro X. Discord also feels a little better this time around, and I’m seeing less performance issues during calls. I can also install apps like Clatter, ShareX, or Tweeten without errors or problems.

And apps are not entirely broken.

Adobe also promised a year ago to bring all its Creative Cloud apps over to the Surface Pro X, but we’re still waiting on a release date… While the Adobe wait continues, you might assume that the company’s Creative Suite desktop apps would run on this latest Surface Pro X. Unfortunately, they do not. I was able to run Photoshop on the Surface Pro X last year, but now you can’t install any Creative Suite apps at all.

While Adobe did release a beta ARM version of Photoshop for Windows and macOS a month later, I think this is still worth highlighting. Microsoft specifically demoed Photoshop as an example of “desktop apps that will just work” when Windows on ARM was announced almost four years prior.

Compare this to The Verge’s coverage of Apple’s chip transition, which went from announcement to shipping hardware in a mere five months. The site dedicated not one, but two articles dedicated to tempering readers’ expectations. The first, by Dieter Bohn, came just before the event and is comically titled “What Windows can teach the Mac about the switch to ARM processors“.

It could be a rough ride

Speaking of things Apple wouldn’t want: ARM-based Windows computers are slower. Unless you’re able to stay within those Chromebook-esque constraints, things get real chuggy real fast. We’ve all been assuming that Apple’s much-vaunted prowess at making fast ARM chips for iPads will translate well to Macs, but there’s no guarantee that’s true until we get to test them ourselves.

The second came after Apple announced the first M1 Macs in November. Jay Peters published a piece titled “There’s a question mark hanging over Apple’s Arm Macs“.

If you preorder, you’re hoping Apple has finally figured it out

“Finally.”¹

The reason they can natively run iOS apps is because the new Apple M1 is based on the Arm instruction set, just like your smartphone, instead of the x86-64 instructions used in Macs and Windows PCs. But the reverse is also true: we’re currently taking Apple’s word that existing Mac apps will work well when they don’t run natively. Yesterday’s was the second presentation in a row where we saw canned demos and unlabeled graphs instead of actual benchmarks and performance comparisons.

It’s not just Microsoft that has struggled with Arm-based computers. Samsung released the Surface-like Galaxy Book 2 in 2018, but you likely won’t be shocked to hear that The Verge’s Dan Seifert had problems running certain apps. My colleague Cameron Faulkner ran into similar issues while reviewing the Lenovo Flex 5G in July. Microsoft is still working to improve Windows’ app compatibility on Arm with x64 emulation. We’re optimistic, but we’re not close to recommending Windows on Arm over Windows on Intel options.

The Verge’s most prominent coverage of Microsoft’s struggles² with Windows on ARM came in articles about Apple’s transition.

This really exemplifies something I’ve been thinking for the last few months, that Microsoft hardware and even software announcements have been treated with kid gloves. It’s not just The Verge either. What first got me thinking about this was episode 36 of The Test Drivers, wherein Myke and Austin discussed technical issues they faced when interviewing Microsoft’s Chief Product Officer, Panos Panay.

The problem is that about half way through the recording [Panos Panay’s] camera overheated, shut-off, something… I’m not totally clear on what happened, and so we lost the feed, which also meant that the audio was all lost at the same time.

Technical difficulties during a video call aren’t uncommon, but this isn’t some schmo working out of a coffee shop with shitty Wifi. This is Microsoft’s Chief Product Officer whose hardware failed catastrophically while doing a podcast specifically to promote Windows 11. Given this is the person in charge of both Surface and Windows, I think it’s a safe bet that the camera involved was built into a Surface computer running Windows.³ Can you imagine the negative coverage if that had happened with Craig Federighi and one of these new MacBook Pros? Instead, the episode went on to reiterates how good Windows 11 will be. Because my old PC isn’t supported, I haven’t had a chance to checkout out Windows 11, but reviews seem to be luke warm.

Likewise and not surprisingly, I haven’t really used any Microsoft hardware products, but my rough impression is that the Intel-based Surface Pro line is good. People love their Surface Pros, and I get why. They have a considered designed, are competitively specced, and don’t come loaded with crap-ware typically found on other PCs. There’s a good chance that if pressed to buy a PC laptop, I’d buy this year’s Surface Pro 8.

I also get why people are excited about Microsoft in general. This new Microsoft surprises and delights by doing things that old Microsoft would never consider. They have Visual Studio for the Mac. They make PC hardware. They even include Linux support in Windows. This new Microsoft is exciting and different, but they’ve also been around long enough to show us who they are. Nothing exemplifies that more than Windows on ARM. I think it’s great Microsoft has spent five years pushing Windows on ARM, but no one in their right mind could say they’ve been as successful at it when compared to what Apple has just accomplished. The tech community likes to pretend that Windows on ARM and the Surface Pro X are viable, if not flawed, options when they’re really not.

We can still get excited about Microsoft, but it’s time to temper the expectations of what they announce with the track record they’ve shown, and we need to take the kid gloves off when examining what they ultimately deliver.

I already had an MP3 player that I used while jogging when the iPod came out, a Samsung Yepp. It could store and clumsily navigate between about 12 songs, all of which had to be transferred over the course of an hour via USB 1.1. By comparison the iPod could hold about 1000 songs, which were easily navigated on a screen via an intuitive wheel, and could be transferred in seconds over FireWire. I really think I was one of the first few people who wanted that original iPod. Being a poor college student at the time, the odds of me getting one were slim-to-none.

Then one summer day, I got home to find my not-so-precious Yepp had been mysteriously shattered. While there has yet to be a confession, I am pretty sure my mother was involved because she offered me cash to replace the broken MP3 player on the spot and without hesitation. It wasn’t enough to buy an iPod, but it was enough for me to easily cover the difference using money from my summer job at the time.

Flash forward to a few years later, when my mother decided to get my father an iPod for Christmas. Still relatively poor, my role in the gift would be to put my father’s music on the device. This was no small task. My father is an avid music listener and had amassed close to 200 CDs at the time. I persevered, knowing that being able to hold his entire library would be a magical moment for my father. It was¹. To this day he still prefers his iPod Classic over pretty much any other device, including his iPhone.

I couldn’t help but notice this article from the Wall Street Journal with the headline “The Apple-Microsoft Tech War Reignites for a New Era” wherein Tim Higgins and Aaron Tiley write:

On Thursday, Microsoft Chief Executive Officer Satya Nadella launched Windows 11 with what was widely seen as a swipe at Apple and the controls it wields over its iPhone App Store, but without mentioning the rival directly. Both companies are positioning themselves for an impending battle over the augmented and virtual reality market that is seen as the next major frontier in computing.

It taps into this sentiment that Windows 11 exists solely in opposition to Apple, but that really doesn’t make any sense to me. As successful as Apple has been, they aren’t an immediate threat to Microsoft. Fundamentally, one is primarily a consumer technology company and the other is a business technology company. While each has tried to drink the other’s milkshake, neither has meaningfully succeeded. The biggest and most immediate threat to Windows is not anything made by Apple. It’s Chromebooks. Look at this Ars Technica article from February:

Despite the fact that macOS landed in third, viewing this as an example of Google beating out Apple directly might not be accurate. Rather, it’s likely that Chrome OS has been primarily pulling sales and market share away from Windows at the low end of the market. Mac market share actually grew from 6.7 percent in 2019 to 7.5 percent in 2020.

Meanwhile, Chrome OS skyrocketed from 6.4 percent in 2019 to 10.8 percent in 2020. Windows fell from 85.4 percent to 80.5 percent.

Windows’s decades long monopoly in the PC space has roughly four trenches that I can think of:

1. Cheap commoditized hardware from partners who have to compete on cost
2. Integration with Microsoft’s market leading Office and Outlook.
3. Market share dominance, leading to a plethora of exclusive third party apps
4. Preferred vendor status among IT departments

For over a decade, Google has been methodically outflanking each of these trenches. Now over half of the “basic laptops” sold by B&H run Chrome OS, Google Docs is outperforming Office 365, and Android has more market share than Windows. I am not sure how well Chrome Enterprise is doing, but seemless security updates combined with minimal on device software seems like benefits that would appeal to most IT departments.

Time was most people needed Windows. Now it’s never been easier to imagine a world without it.

Let’s take look at some of the features Microsoft is touting with Windows 11.

1. A new design that just happens to trend toward the look of Chrome OS
2. A deep integration with Microsoft’s market leading Teams
3. Android app support
4. An added emphasis on security.

These aren’t features to lure Mac or iPad users to Windows. They are to keep Windows customers, consumers and businesses alike, from switching to Google. Even letting developers use their own payment processors, which has been touted as an assault on Apple’s App Store, is just as much an assault on the Google Play Store.

I am not arguing Microsoft and Apple don’t compete. Of course they do, but framing Windows 11 as some major shot in a war against only Apple is outdated and foolish in an era when Microsoft necessarily has a much bigger adversary to defend against.

While demonstrating my external display to a relative looking to buy one for their home office, I was reminded of how bad the display market has gotten for professionals. “Professional” is somewhat of an overloaded word. What I mean by “professional” here is someone who spends most of their day working on a computer and is thus staring at some display for most of the day. The display I use in my home office is the 4k HP Z27. This display has decent brightness and color, and the USB-C charging is downright fantastic. I would happily recommend the Z27 to most professionals with one caveat: resolution.

A few quick searches on B&H reveal that the most popular screen size for an external display is 27 inches, and that the two most common resolutions at 27 inches are still 1920×1080 (a.k.a. Full HD or FHD) and 2560×1440 (a.k.a. QHD). Simply put, QHD can show more detail than FHD because it has more pixels. This is particularly valuable on large displays where more detail translates to being able to show more text and graphics. FHD by comparison starts to look oversized at around 24 inches. For these reasons, most professionals of various sorts prefer QHD displays. Coders can see more code, writers can see more words, and illustrators can see more illustration. Because there have long been a plethora of QHD 27-inch displays made by a variety of companies, professionals could (and still can) easily pick and choose which model best suits their need. Coders and writers don’t have to pay a premium for the color quality needed by illustrators.

But by today’s standards, both FHD and QHD are considered low resolution given a 27-inch display. As I wrote above, the number of pixels in lower resolution displays is mostly about showing more stuff. By comparison, the number of pixels in high resolution displays is almost entirely about showing more detail. The high resolution displays found in most Apple devices have double the resolution and four times the pixels of their predecessors. Apple refers to these displays as “Retina“, because they claim the human eye can’t differentiate between the pixels in these displays at typical viewing distances. The default display found in the 27-inch 5k iMac is effectively “Retina QHD”, because its 5120×2880 resolution is exactly double that of its 2560×1440 QHD predecessor.

Whereas low resolution 27-inch displays have plenty of FHD and QHD options, no other Retina QHD displays exist on the market outside of the 5k iMac and the $1300 LG UltraFine. The only common high resolution option at 27 inches is 3840×2160 (a.k.a. 4k), which in Apple parlance is effectively “Retina FHD”. While both macOS and Windows 10 can effectively render QHD resolution given a 4k display, details are still noticeably less sharp. Regardless of whether you believe Apple’s “Retina” claim or not, this lack of sharpness makes sense for one simple reason: half pixels don’t exist.

Here’s what I mean:

A QHD display natively displays 2560×1440 pixels where each graphical point is neatly represented by 1 physical pixel.

A 5k iMac renders Retina QHD in points where each graphical point is neatly represented by 4 physical pixels.

A 4k display renders not quite Retina QHD, also in points, but because 4k is less than double the resolution of QHD, each graphical point is represented by 2.25 physical pixels. But because half pixels don’t exist, things necessarily start to get messy.

This leads to my caveat about resolution when recommending any 4k display, including my HP Z27. 4k is still better at 27-inches than low resolutions, but it necessarily requires compromise. You can either have more text and graphics at not quite Retina QHD or perfect, but oversized sharpness at Retina FHD¹. Professionals who simply want a Retina replacement of their current QHD display have exactly one $1300 option. Given this severe lack of choice, most of them will spend less than half of that by compromising on a decent, but not quite Retina QHD, 27-inch 4k display… like me.