Riccardo Mori

A few notes after the Windows 11 preview, and on Windows in general

June 28, 2021

Software

The other day I watched the Windows 11 introduction event and I liked what I saw. The following observations are scattered, won’t probably be very cohesive, and are mostly meant to be bits of thinking-aloud I want to put out there.

Regarding my notes on the features of Windows 11 that were presented on 24 June, I will assume you know what I’m talking about. Here’s the reference page (at the time of writing) on Microsoft’s website. For a lengthier reference, you can read Introducing Windows 11 on the Windows Experience Blog.

The elephant in the room

Both during the event and in the aforementioned blog post Introducing Windows 11, there’s a patently false statement:

The web was born and grew up on Windows.

It was not. I was almost chastised for not having scrambled to point this out on Twitter. Yes, I had noticed that statement. Yes, I know it’s not true. The reason I didn’t run to Twitter and shout Impostors! Shame on you! is that other people in my timeline had noticed it and were talking about it. See for example this thread by Benj Edwards on Twitter, with several interesting responses. Read the Wikipedia entry for World Wide Web for a detailed historical perspective.

Fun bit of trivia: the father of a high school friend of mine was friends with someone working at CERN around 1990. The first time I heard about the World Wide Web was one Sunday in early 1991 when I was having lunch at my friend’s house and his dad mentioned what he had heard about the project. This will have a considerable impact, was his comment. Indeed.

In an exchange with a friend on Twitter, I said: [Panay] could have just written: “For many, the web was discovered through Internet Explorer and Microsoft Outlook” — not as dramatic a statement, but certainly truer. And I added: It’s a big blob of bullshit that was added to the narrative sauce according to which “Windows has always been an innovative platform”. It hasn’t.

It’s strange how some people have interpreted my words and thought I was trying to downplay the enormity of the original lie (The web was born and grew up on Windows). In my book, I usually don’t say that something is a big blob of bullshit if I want to minimise or downplay it.

Anyway, now that we got the elephant out of the way — hopefully? — let’s talk user interface, et cetera.

Window management

Snap Layouts is a very nice and direct way to spatially organise windows in your workspace. On the Mac I’ve seen something similar only in third-party utilities such as Moom or Divvy. It’s good to have such a versatile feature built into the system, especially if your operating system is called Windows.

For me it’s interesting to note how, on the Mac, I’ve never really felt the need to use such utilities for window management. I tend to just arrange windows manually and if I need to create different workspaces, I just group windows in different Desktops via Mission Control. When using Windows, however, I often find myself wanting some tool to automatically arrange the various windows I have open on the desktop. Which is a bit strange, since I don’t really have a different workflow than what I have on the Mac. In any case, with Windows 11 I shall take advantage of this feature without a doubt.

Snap Groups are an interesting addition to handle complex multitasking. Having Windows remember the layout of a bunch of apps and projects, and being able to access them directly from the taskbar is undoubtedly handy; especially when you end up having a lot of application and document windows open for each project you’re working on.

The ‘docking/undocking experience’, where Windows remembers the window layout of all the apps/documents you were using as you disconnect and reconnect your laptop to an external display, is another feature I really like. From 2002 to 2018, my main work machine has always been a Mac laptop (iBook and PowerBook at first, MacBook Pro later on) in desktop configuration attached to an external display. And I can’t believe the docking/undocking experience over the years has actually been degrading rather than improving. Before finally switching to a retina iMac in 2018, when disconnecting and reconnecting my MacBook Pro to my external LG monitor, not only was I forced to manually rearrange several windows and spaces, but I even had to constantly enter the Displays preference pane in System Preferences to re-select the correct colour profile for the display.

I wish Apple prioritised these very same features instead of focusing on how to change the look and feel of Safari (for the worse), or rethinking certain user interface elements or certain user interactions in ways that end up creating friction and hiccups in Mac workflows, in areas where there wasn’t really anything wrong previously.

Dead Tiles

I’m not a long-time Windows user. At least, I haven’t been using Windows continually since its inception. I started in the early 1990s with Windows 3.0, then went through 3.1 and Windows 95; then there was a period with Windows 98, and that was it for a few years. I knew my way around Windows XP, Vista and 7 mostly because during certain work collaborations I had to use Windows PCs.

I rediscovered Windows around 2017 by way of Windows Phone 8.1 and Windows 10 Mobile when I wanted to study the Metro interface on a few Nokia Lumia smartphones I’d been acquiring. And I was more impressed than I had anticipated (you can read my reflections here and here).

One of the features of the Metro user interface I really love are Live Tiles. I think they are a good, efficient concept. They merge together different user interface layers — an app, an updated status within the app, a shortcut to launch the app, and a widget — in a single element that is also visually customisable. After my rediscovery of Windows on mobile devices, I had the opportunity of acquiring a ThinkPad T400 with Windows 7 preinstalled for a very low price, and immediately updated it to Windows 8.1 Pro. Because I wanted the same experience on a PC. It’s perhaps for this reason that — unlike many long-time Windows users — I really like and enjoy Windows 8.1 and its user interface (well, most of it). And I’m probably one of the few who likes Windows 8.1’s full-screen Start menu:

Sure, maybe its being full-screen is a hindrance or an interruption when you’re doing something and have to access it, and it takes up all your screen. But Live Tiles are much better here than crammed together in a column of Windows 10’s Start menu, and the list of apps in Windows 8.1’s Start menu is much more legible than under Windows 10, at least for my ageing eyesight.

Which brings me to Windows 11, where Live Tiles are definitely gone. As a consequence, its Start menu has become utilitarian, but in a bland, boring, aseptic, business way. The Live Tile has been dissected and divided into its main components, and now you have apps on one side (the Start menu) and Windows Widgets on the other (the Widget expandable glass pane). The fun, the whimsy in the Windows interface has been neutered and now it’s all clean, calm, and played safely.

Old code, many UI layers

John Gruber linked to this certainly fascinating piece: State of the Windows: How many layers of UI inconsistencies are in Windows 10? — And it has truly been an interesting read. Many critics of Windows have been pointing out the same things over and over again: it’s a bloated system which has never really got rid of layers and layers of old code, and layers upon layers of past UIs. And for sure, after Windows 11 was presented, they all came out of the woodwork pointing out the same old stuff, that Windows 11 is yet another layer of paint splashed over this tall, Tower-of-Babel-like building that is supposedly getting more and more unmanageable, version after version.

I’m not a developer, so I can’t make insightful comments on Microsoft’s decision to retain so much legacy code over the years. From a pragmatic standpoint, I’m inclined to say that if doing so is good for backward compatibility and it doesn’t have a meaningful impact on a machine’s performance, then what’s the problem?

I’m running the latest Windows 10 Pro on a 2013 ThinkPad X240 with 4 GB of RAM. It has a 1.6 GHz dual-core Intel Core i5-4200U CPU (2.6 GHz Turbo boost) with a 500 GB 5400rpm hard drive. I could make things a bit faster by swapping the hard drive with an SSD, but since I use this laptop occasionally at the moment, it’s not a priority. Still, I have no complaints about Windows 10 performance on this ThinkPad. It’s not the right machine for CPU-intensive pro applications, sure, but it handles everyday tasks to moderate workloads very well.

I also have Windows 10 running on my iMac via BootCamp, and in many situations the iMac feels much more responsive than under Mac OS. So, once again, I don’t see (empirically) the presence of strata of old code as a particular hindrance in day-to-day use. I have been running Windows 10 on this iMac for five months now, and I never ran into a problem (conflicts, freezes, crashes, app instabilities, driver incompatibilities, etc. — nothing of the sort).

But I am also inclined to say that even the various layers of past Windows UIs aren’t problematic. Because their differences are, in most cases, purely visual. Yes, sometimes the æsthetic mismatch is indeed a bit jarring, but the inconsistencies stop there. In the way certain windows or buttons or controls look, not in the way they work. If they also worked differently, then I would agree that it becomes a serious UI issue.

We can look at Windows as if it were a hotel that gains two or three more floors every time a new Windows version is released, and guests are always accommodated in the newer floors at the top of the building. There, the environment looks fresh, clean, modern; and there is where usually guests stay 90% of the time. When guests need to access a few facilities on the lower, older levels, they may find furniture that looks dated, or doors and windows finished in a previous style, but none of these elements will work differently. They may pause for a moment in mild bemusement, but there is no severe usability impact. It’s mostly a superficial annoyance.

Am I excusing this? Not exactly. Homogeneity in user interfaces is important, and for some users, loading the older-looking Control Panel by clicking on certain advanced options within the Settings app (which is the Control Panel’s modern equivalent) can be confusing. I would certainly prefer for Windows to have the exact same visual style everywhere. But in the meantime, do this little experiment: take a look at the various examples of visual inconsistencies listed in the aforementioned article, then ask yourself if you’d really be able to tell that certain windows or controls come from a previous version of Windows without external guidance. Sure, certain icons can be easily spotted. But most of the times you’re presented with windows with more subtle differences.

Something I’ve always noticed with Windows applications is that, since there was never a strict enforcement of Human Interface Guidelines in the Windows world, third-party developers have often been more creative with their applications’ interfaces. When I was on Windows more frequently, I remember using several third-party apps whose UI, windows, controls, looked completely different from the system UI. This has been quite the different situation on the Mac front, where the majority of third-party developers have always stuck to Mac OS’s Human Interface Guidelines, in order to offer a ‘Mac-like’ app as much as possible. Still today, several third-party Mac apps look and behave as if they were built into Mac OS — some behave even better than first-party apps.

All this to say that visual inconsistency has always been part of Windows in a way or another, and that’s why a lot of Windows users — especially non-nerds — don’t seem particularly bothered by these inconsistencies. Again, I’m guessing they would be if these inconsistencies were more than just cosmetic.

And to anticipate a possible objection: yes, I tend to be much stricter in my criticism of Apple’s UI inconsistencies exactly because Apple has always been better at user interfaces and user interaction, it has established a higher standard since the early days of the Macintosh, and it’s extremely jarring when you notice that in recent years the company hasn’t maintained that higher standard as consistently as before. UI consistency has always been a big deal — rightfully so — when it comes to Mac OS, and when you find UI elements that are inconsistent not only in the way they look but also in the way they behave (see for example the Move to… element in Big Sur’s Mail, that looks either like a button or a text field, and it’s actually a menu), then it is problematic.

What needs rethinking: the Ribbon

Something that’s been unavoidable for me over the years has been Microsoft Office. Whether on Macs or PCs, I’ve always had to deal with it, usually because my clients worked with Office files and requested total compatibility in our file exchange. But in my days using Windows 3.1 to Windows 98, I actually liked to use the first versions of Microsoft Word. Despite its complexity, its UI wasn’t that bad, and over time I had learnt to master it. That is, until the dreadful Ribbon UI appeared in Office 2007.

I remember opening Word 2007 on my wife’s PC one day and being utterly dumbfounded by the drastic overhaul of its interface and controls. So drastic that it took me minutes to find where even the basic stuff had been reassigned. The user interface shock was so strong I kept using earlier versions of Word (and Office) on PCs and Macs for years. And whenever I had some free time, I would dedicate it to calmly explore the Ribbon UI to find my way around that mess. I’ve come to tolerate it in the latest versions of Word for Mac, but my proficiency in this application has severely diminished since the Ribbon revolution.

But where I find the Ribbon to be particularly intolerable and unnecessary is in Windows’ File Explorer (what would be the Finder on a Mac). Let’s have a look at a File Explorer window in Windows 10:

(1) This element looks like an icon that does nothing. No, if you click it, you invoke a menu with commands to handle the window, including (redundantly) the controls to minimise, maximise, and close the window.

(2) These two other mysterious-looking elements at first glance suggest they’re tied together. The first time I saw this, I thought that the downward triangle served to indicate that the document icon on its left had a drop-down menu, as opposed to the element I marked as (1) here. But no, these are actually two separate clickable elements that invoke two different things: one opens a panel, the other a menu. Also of note: these three elements on the top left of the window all invoke items with completely different styles and UIs from one another.

(3) File, Home, Share, View, Picture Tools — all these look like menu commands, and you’d expect them to behave like such, especially if you normally keep the Ribbon UI hidden when not in use. Instead, only the File command displays a menu. The others will invoke the messy Ribbon UI. (At least the File command is coloured differently). While we’re here, something I’ve never understood in this window design is why, when every colour and styling is so subtle, do we have this Manage label screaming at you in such a prominent position?

(4) The Ribbon itself: on a theoretical level, I understand the reasoning behind it. Instead of having a hierarchy of controls that is revealed only when a menu is invoked, here you have everything laid out in front of you. You should have quicker access to certain functions and controls. In theory. In practice there’s such a high amount of visual clutter that it defeats the purpose. The various elements inside the Ribbon do retain a hierarchy, but it’s often laid out both horizontally and vertically.

In parts of the Ribbon your eye is invited to follow horizontally — see the various macro-sections Panes, Layout, Current View, Show/Hide. But then you have elements inside these sections that invoke menus and submenus, and so you’re back to selecting things vertically. Also, have you noticed just how tiny are those triangles indicating the presence of a menu? Everything is certainly discoverable here, but the level of visual complexity is just off-putting, and the end result is that all these elements and controls, despite being ordered inside different sections, give an impression of utter general disorder.

(5) Again, the navigation arrows and the location bar here are just too small and not as visually prominent as they should and as they used to be in previous versions of Windows.

(6) Same goes for the Search field. Just crammed there, almost an afterthought.

(7) And what do we have here? Two more elements that essentially replicate a couple of View options. On the one hand, I’ll admit they’re handy, as they mostly avoid having to deal with View options from inside the Ribbon. On the other hand, they’re sooo tiny… And why put them there? Why not group all these icons together with the ones at (1) and (2) above?

Or better yet, why not get rid of the Ribbon UI altogether and go back to previous designs and refine them with a modern look?

In Windows 8.1, the Ribbon was already there, but at least the various UI elements in a window were more clearly defined via the use of more contrasty colours:

In Windows 7 things were better. No Ribbon, just menus, and the fact that each window resembled a Web browser window, navigation was more intuitive, as navigation controls were more prominent.

If we go back even further, in Windows 2000 Professional, all the UI elements and controls in a File Explorer window were possibly even clearer:

Look at it. There is nothing ambiguous in this window. Menus look like menus, buttons look like buttons, information is clearly laid out.

I’d really like for this Ribbon UI to disappear from Windows. I don’t know if Microsoft has something planned about it for Windows 11 or in the future, but in my opinion they should really rethink it and at least streamline it if they don’t want to remove it altogether. Elements are too scattershot: I would either stick to simple menus or a well-designed toolbar with a row of icons. And, while maintaining a clean, glassy look for windows in general, I would use colour and contrast to make the various components (title bar, status bar, navigation bar, etc.) stand out, something we already saw above in Windows 7 and 8.

I think this is more of a UI issue rather than some windows and buttons found in remote parts of the operating system that still look like they looked in Windows 7 or XP or whatever.

Conclusion: fanboyism? I’m past that.

As I said at the beginning, I like where Windows is going, visually, and (in part) from a UI standpoint. There is nothing particularly groundbreaking in Windows 11 — though if they really keep Windows updates 40% smaller and faster to install, as Panay said in the presentation, this could be groundbreaking enough, given Windows’ history in this department. But I liked certain details: Microsoft’s focus on offering advanced and efficient options to handle complex multitasking, the redesigned user interaction when using Windows in tablet mode, and the company’s willingness to be more open towards developers. (I’m not naïve, I know this is a strategic move to appear like ‘the good guys’ in contrast with Apple’s policies and attitude towards developers; but still).

In general, I had the impression that Microsoft seems to know their users and understand their needs better than Apple does with their (Mac) users. At least at its point in time. I still return to Safari 15 and ask myself, Do designers at Apple think people like to browse the Web this way? Do they think that shuffling tabs and the browser address bar around help or facilitate the use of the browser?

And I already know someone will email or message me asking if I’m becoming a kind of Windows fanboy all of a sudden. The point — if you haven’t understood it yet — is that I’m past any type of fanboyism. I observe, and talk about what I like and what I don’t like. These are not the times to be anyone’s fanboy. These are times to look at what big technology companies put in our plates, and ponder everything before happily shoving it down our throats. People need to get their work done with tools and solutions that respect their work and their workflows. That respect the users themselves. And if these tools and solutions come from different companies, then so be it. The ‘Us versus Them’ mentality should shift from thinking in terms of Apple vs Microsoft, iOS vs Android, Linux vs everyone else, My Favourite Platform vs Your Favourite Platform, and turn into Us, the users, versus Them, the big technology companies.

Safari 15 on Mac OS, a user interface mess

June 18, 2021

by Riccardo Mori

Software

Introduction: no WWDC-related observations?

Maybe you’ve wondered why I haven’t written my usual post-WWDC article sharing my observations about the keynote. Well, once again my main job absorbed most of my time and energies. But also, frankly, there hasn’t been anything extraordinarily good worth scrambling to write down.

Don’t get me wrong, plenty of nice things have been introduced at WWDC 2021 across all Apple platforms, but overall it felt like going to a nice restaurant serving you decent food. Did you eat well? You’d say yes. Did you eat so well it’s worth leaving a lengthy review online? Not really.

I will write something about WWDC, if I find an interesting-enough angle that’s worth writing about three or four weeks after the event.

The Mac — Saving the last for last

What’s really worth mentioning, though, is how Apple has screwed up the whole dessert.

The dessert is the Mac, of course, and in calling it a dessert I’m trying to give a respectable, polite label to something that was left for last. Deal with it, I’m still a Mac user first, the Mac platform is still the one I care most, so I’m giving it precedence over everything else.

There has been a wave of pieces written in recent times by pundits who now say that those who believed Apple did not really care about the Mac were wrong. Look, the new Apple Silicon chips! Look, new hardware updates! Even redesigned Macs! Whoa, crazy stuff indeed.

Why was Mac OS left for last at the WWDC keynote? Because essentially, everything new that’s coming in Monterey are features originally devised for iOS and iPadOS devices, and then adapted for Mac OS. And Universal Control I’d call an ecosystem feature. Nothing, to my recollection, was conceived specifically to take advantage of the Mac as a powerful, versatile machine and platform. Nothing was designed specifically with the Mac in mind. Nothing was designed as a Mac-first feature.

The utter user-interface butchery happening to Safari on the Mac is once again the work of people who put iOS first. People who by now think in iOS terms. People who view the venerable Mac OS user interface as an older person whose traits must be experimented upon, plastic surgery after plastic surgery, until this person looks younger. Unfortunately the effect is more like this person ends up looking… weird.

These people look at the Mac’s UI and (that’s the impression, at least) don’t really understand it. Its foundations come from a past that almost seems inscrutable to them. Usability cues and features are all wrinkles to them. iOS and iPadOS don’t have these strange wrinkles, they muse. We must hide them. We’ll make this spectacular facelift and we’ll hide them, one by one. Mac OS will look as young (and foolish, cough) as iOS!

Safari’s facelift

The new Safari. Image © Apple, Inc. and taken from the Mac OS Monterey Preview webpage.

For years, Safari’s interface has been a good balance between minimalism and functionality. The app chrome was shaved away, iteration after iteration, and it seemed it had reached its minimum but still viable footprint.

On my 11-inch MacBook Air (running Mac OS 10.13 High Sierra and Safari 13.1.2), the total vertical space taken by Safari’s Title bar/Toolbar, Favourites bar and Tab bar is 86 pixels, 64 if you hide the Favourites bar. The 11-inch MacBook Air’s display is only 768 pixels tall, but even in these ‘cramped’ conditions there’s still enough space left to a website for clarity and meaningful navigation. Especially if you hide the Favourites bar and use Safari in full-screen mode.

On my 13-inch retina MacBook Pro running Mac OS Big Sur 11.5 beta and Safari 14.1.2, the total vertical space taken by Safari’s Title bar/Toolbar, Favourites bar and Tab bar is 110 pixels, 82 if you hide the Favourites bar.

In both instances, the vertical space taken by the Tab bar is 28 pixel. Twenty-eight.

The point I’m making with all this pixel peeping is that these are negligible measurements. Getting rid of the Tab bar with the excuse that you’re saving space is the stinkiest bullshit I’ve ever smelt in a while. 28 pixels for any of the current Mac displays is nothing.

In The reshaped Mac experience I wrote:

Going through Big Sur’s user interface with a fine-tooth comb reveals arbitrary design decisions that prioritise looks over function, and therefore reflect an un-learning of tried-and-true user interface and usability mechanics that used to make for a seamless, thoughtful, enjoyable Mac experience.

What Apple has done to Safari 15 is no different. Ever since WWDC 2020 this damned quote by Alan Dye (Apple’s VP of Human Interface) has stayed with me, and most of what’s happening on Mac OS and to Mac OS always comes back to it:

We’ve reduced visual complexity to keep the focus on users’ content. Buttons and controls appear when you need them, and they recede when you don’t.

I’ll reiterate this until I’m out of breath: it’s not the right approach. Or at least, it could make sense if applied thoughtfully in certain parts of the interface. What’s going on in Safari 15 is that buttons and controls have actually been swept under a rug with an (···) icon, and they don’t appear when you need them — you have to look for them. And it makes no sense whatsoever that one would want to go looking for the Reload button in a tiny menu with a More… icon.

And what’s more arbitrary than the new Tab bar design? Two things any user, no matter their tech-savviness, has needed in a browser:

A wide Address bar to see exactly where they are, which webpage it’s loaded, the whole URL.
A proper Tab bar, with as much horizontal space as possible, to be able to open a lot of tabs and read at least a small part of their titles.

(And no, I don’t think that only nerds open dozens of browser tabs at a time. I’ve seen it firsthand in so many occasions and contexts that it can’t be just anecdotal data.)

And what have they done in Safari 15? They merged these two browser bars, of course, thus delivering the worst of both worlds:

On the one hand, the Address bar keeps shifting position and changing width (meaning it gets shorter) as you create more tabs. Please, turn on your common sense, and consider how ridiculous this is.
On the other hand, now that tabs have to coexist with the Address bar on the same line, they have less space. Meaning that on a relatively small 13-inch display, it’s probably enough to open just six tabs before things get cramped and tab text gets truncated to meaninglessness.

In other words, what a browser needs is horizontal breathing room, instead we have Apple doing things backwards, sacrificing horizontal space to give us what, 28 more vertical pixels? It seems as if the people in the design team are all working exclusively on 32-inch Apple XDR Pro Displays.

Many people are also commenting negatively about the new feature where Safari’s app chrome changes colour by taking the accent colour of the currently loaded website. Michael Tsai says: Having the page background color bleed into the tab area makes it harder to read, and it feels weird for the current page’s color to affect the way other tabs look. It also works inconsistently, even on the same pages on Apple’s site. At least there’s a preference to turn it off. And I agree with him, but this is a minor offence compared with the rest of Safari’s UI butchery.

By the way, in case you thought it was an original feature, go have a look at Vivaldi browser. While you’re there, take a look at how Vivaldi tackles the ‘too many tabs’ problem. Spoiler: by adding a second Tab bar. And I say, why not. There’s enough vertical space on current computers. And those users who usually deal with a lot of tabs prioritise having easy access to them and tab legibility, not having 20 more vertical pixels available.

While we’re on the subject of browser tab management, Apple’s proposed solution — creating groups of tabs — is rarely efficient and overall unconvincing. A little experiment: how many browser tabs do you have currently open? Let’s be conservative. Let’s say nine. How many are so tightly related among one another that you can meaningfully group them together? I bet none to very few. Unless you maybe start using such generic labels that you always need to check inside each group to see which sites you have there.

I have currently 18 tabs open in Brave on my main Mac. Each points to a website that is not related to any other and can’t be put under the same label or category. In everyday browser use, tabs are a messy affair that is kind of messy by design and users are generally fine with it. We need open tabs, we need to see what’s open at all times, and we need to be able to quickly jump to the tab we need in the here and now. In this scenario, pinning tabs is much more useful than grouping them and hiding them away. Your most visited sites have a persistent place in the Tab bar. Their tabs are reduced to show only the favicons because you know them well and there’s no ambiguity. And that’s really it. This way of browsing is not a problem in search of a solution, Apple. You have so many more UI issues to fix, instead you add some more by ‘revolutionising’ Safari. Stop — hiding — UI — elements — arbitrarily!

I dearly hope the new UI in Safari can be neutralised by some toggles in the browser preferences. But I also hope this thoughtless UI won’t give strange ideas to the designers of Brave, Firefox, and other browsers. I’m really liking Vivaldi at the moment, by the way. Its user interface can become quite complex, but the developers and designers have been open enough to give the user a great deal of flexibility, so that you can truly customise Vivaldi to your needs.

I’m also really liking another browser I’ve been beta testing for quite a while, about which I’ll hopefully talk in more detail soon.

Safari 15 on Mac OS, a user interface mess was first published by Riccardo Mori on Morrick.me on 18 June 2021.

No thanks to the App Store

June 5, 2021

by Riccardo Mori

Software

After months of blogging drought, Marco Arment publishes another great article, Developer relations. I kept nodding all the way through, and this passage inspired me to chime in based on personal experience as a power user / informed customer:

Apple further extends the value argument, and defends their justification for forced commissions, by claiming responsibility for and ownership of the customer relationship between all iOS users and each app they choose to use.

This argument only makes sense — and even then, only somewhat — when apps are installed by a customer browsing the App Store, finding an app they hadn’t previously heard of, and choosing to install it based on App Store influence alone.

But in the common case — and for most app installations, the much more common case — of searching for a specific app by name or following a link or ad based on its developer’s own marketing or reputation, Apple has served no meaningful role in the customer acquisition and “deserves” nothing more from the transaction than what a CDN and commodity credit-card processor would charge.

The idea that the App Store is responsible for most customers of any reasonably well-known app is a fantasy.

I’ve purchased or downloaded apps from the App Store since the beginning, and I can confirm everything Arment says.

Let’s start with iOS. Since 2008, I’ve accumulated about 250 apps for iPhone, iPad, iPod touch. It’s not an incredibly high number, and it includes apps that are now no longer available on the Store. Of course, the first years of the App Store were the most active for me as a customer: lots of new and interesting applications, lots of exploring. Those were exciting times indeed. As I tweeted recently, the sum of the apps I’ve purchased over the past four years doesn’t even reach the number of apps I purchased in 2010 alone. Then over the years things started to settle, and I started getting more and more selective. The bulk of the apps I still use today on all my iOS devices are apps I purchased around the 2009–2015 era.

But how did I find all those apps, anyway? Here’s a (surely incomplete) list:

Personal recommendations from friends
Recommendations on social networks
Recommendations on personal tech blogs I follow via RSS
News from app developers I was already following
Marketing efforts by the app developers themselves
Reviews from websites and portals that have been app-review-oriented from the beginning. Sites like Beautiful Pixels, and the now sadly defunct AppStorm Network in particular.

A unique resource worth mentioning is AppShopper, which has been both a website and an iOS app itself. Sadly it doesn’t work anymore, but when it did, I remember thinking that is how Apple’s App Store app should have worked. The app was efficient and well-designed, and it favoured discoverability by doing what Apple’s been trying to do since the major redesign of the App Store app a few iOS versions ago. At least 45% of the apps I have on my iOS devices were discovered thanks to AppShopper alone.

On the other hand, if I had to approximate a number to tell you how many apps I have discovered simply by accessing the App Store app and — without actively searching anything in particular — just finding an interesting app out of sheer exploration, then I’d say no more than ten apps in total. In the whole history of the App Store.

What’s remarkable is actually how many ‘misses’ I’ve got thanks to the App Store, especially in the case of Mac software. By ‘misses’ I mean when you realise you’d like to have an app for a specific task, you don’t find any recommendations via the usual channels, you resort to exploring the App Store in the hope you’ll find something useful, and you end up downloading one or more apps to try out and see if there’s a viable candidate. Apps that in most cases turn out to be underwhelming at best, or don’t really do what you wanted the way you wanted, or are just terribly-put-together pieces of software.

Search has always been one of the App Store’s weak spots. As someone still using several vintage iOS devices, it would be nice to be able to search for apps that are still on the App Store but have lower system requirements. Instead I have to resort to launching a browser and perform a search like this:

site:itunes.apple.com/us "requires iOS 5"

Given the sheer quantity and varying quality of apps it offers, a place like the App Store — both on iOS and Mac OS — should have some kind of Advanced Search facility, a way of filtering or otherwise fine-tuning your searches. Instead it’s still a mess, and even when you know the name of the app you’re seeking or the name of its developer, often it doesn’t appear among the first search results. To be fair, the situation has improved over the years, but the way sponsored apps and ads are given precedence still messes up search results a bit.

It’s a pity that app review portals like AppStorm have disappeared today. Its format was great: there was a staff made of curators who really cared, offering well-written reviews that were not too long, not too brief, with a final rating from 0 to 10. There were app round-ups, so you could have an idea of, say, which were the seven most notable productivity apps of the month. There were news and opinion pieces.

Even with the recent restructuring and redesign of both the Mac and iOS App Stores, even with the undoubtedly useful Today section, I’ve had a hard time finding new apps ‘blind’, just relying on the search & discover features of the App Stores themselves. Marco Arment is entirely right — the App Store has been little more than a conduit for purchasing and installing apps. In my personal experience, after 13 years purchasing apps, the App Store itself has had next to zero influence over my decision whether to buy an app or not. In the sheer majority of cases, I already knew what I was looking for, and I already knew it was going to be a good-quality app.

Apple Silicon Macs: wait

June 3, 2021

by Riccardo Mori

Tech Life

There has been one common theme in the readers’ emails I’ve been receiving for the past two months or so: people asking for buying advice regarding Apple Silicon Macs. While I intend to respond privately to each and every message, because everyone comes from a different, specific situation, I’ll write here my general take on the matter, one week before the WWDC.

If you want the TL;DR version, there you have it in the title of this piece. The longer answer is a reflection on the current state of things both from a hardware and software standpoint.

Hardware

Since the Intel-to-Apple Silicon transition started, Apple has introduced one new chip, the M1 SoC, and as one would expect from Apple’s history designing their own chips for iPhones and iPads, the M1 delivers an impressive CPU/GPU performance and power efficiency. I just wrote “as one would expect”, but in truth the M1’s performance vastly surpassed everyone’s expectations.

Currently there are four Apple Silicon Mac models powered by the M1: the 13-inch MacBook Pro, 13-inch MacBook Air, Mac mini and the 24-inch iMac. With the exception of the MacBook Air and entry-level iMac (both lacking 1 GPU core), these Macs all perform in the same excellent way. The models lacking 1 GPU core perform slightly worse but I guess that a regular user won’t really notice the difference.

At the moment, there’s no real performance differentiator between these machines; it’s not like the PowerPC or Intel days when you had, say, two Macs with the same CPU class but different clock speeds, or a different number of CPU cores. The only thing making a difference between one M1 Mac and another is perhaps if you choose a variant with 16 GB of integrated RAM instead of 8. At the moment, the main differentiator between these M1 Macs is their form factor and I/O connections.

Should you buy one of these Macs now? At this point the transition to Apple Silicon Macs is more or less halfway through, maybe even less than halfway. Apple still has to introduce the second wave of Apple Silicon machines, the prosumer/professional tier. Putting together various rumours, we should expect a 14-inch and 16-inch MacBook Pros, a more powerful iMac with a bigger display, a new Mac Pro, and there’s even a more powerful Mac mini on the way. There are also rumours of a new, colourful MacBook (I’m not sure if it’s going to be a redesigned MacBook Air, or if Apple is thinking about reintroducing a laptop that is simply called MacBook, without Air or Pro designations).

All these not-yet-announced Macs, supposedly starting to appear over the next months all the way into next spring (I’m speculating), should all be equipped with a more powerful chip than the M1.

So, let’s get back to the question, Should you buy one of the current M1 Macs now? Here’s my reasoning:

Do you currently own an older Intel Mac and you’re getting frustrated by its performance for certain tasks, a pro-level Mac would be overkill for your needs, and you can’t postpone an upgrade anymore? Then get one.
Do you think a colourful 24-inch iMac is just the right desktop Mac you were waiting for, are satisfied with what it offers and you think its price/performance is just right for you? Go for it.

Otherwise, wait. Even if you’re coming from a Mac mini and find the M1 mini appealing for the performance boost over your Intel mini at a relatively low price, wait for the purported M1X Mac mini. It could offer you even more performance and it may still be an affordable machine.

Wait until the transition to Apple Silicon Macs reaches the next stage. Wait until you’re starting to get a clearer picture. To see what the pro machines offer and at what prices. You will literally make a more informed choice.

Software

Apple Silicon Macs have of course a different architecture than Intel Macs, and even if an Intel Mac and an Apple Silicon Mac running Mac OS Big Sur look identical superficially, behind the scenes things behave differently.

You should really keep an eye on Howard Oakley’s blog, and I suggest you take the time to check what he has written so far under the M1 Macs category. In Would you regret an M1 Mac?, he writes:

M1 Macs have a complicated security model, quite different from that enforced by the T2 chip, which is difficult enough at times. Although this allows an M1 Mac to boot from an external disk without changing its security settings (in the way that you have to with a T2), recognition and use of external storage for such purposes is continuing to trip M1 users up, particularly when combined with relatively immature Big Sur installers. In recent versions of macOS 11 this hasn’t improved as much as it needs, and now even the NVRAM seems to have become locked down from the system.

And:

Macs, even those with T2 chips, have always offered the user choices as to how they’re used. Although the most popular is to boot each Mac you use into your account on its internal storage, many prefer to go around with their own bootable system on an external disk. This is ideal for someone who wants the same environment, apps and documents when they’re in the office as when they’re at home.

The good news is that’s now possible using M1 Macs, with some significant cautions in addition to the current limitations over updating bootable Thunderbolt disks.

You probably don’t want to try this with just any M1 Mac, though. Getting your wandering drive to work with a new M1 Mac is quite a fiddle, and not something you’d want to do every time you go into the office. With just two or three different Macs, it appears quite usable within the limitations now imposed by macOS.

These are just excerpts, make sure you read the article in full. Howard then concludes:

Are M1 Macs ready?

Yes, but you should still tread warily if you need:

more than 16 GB memory;

more than two Thunderbolt ports, or to use a dock;

to boot from an external disk;

to boot more than one Mac from the same external disk.

My very condensed take is this. Apple Silicon Macs, due to their different architecture, feature a different security model, a different boot structure, a different recovery system in case things go wrong, and despite we’re by now four minor releases into Big Sur, the situation is still a bit rough around the edges. I know we’re in the middle of a transition, but I wish things were somewhat smoother. And I wish Apple outlined these significant changes more clearly for the end user. New customers entering the ecosystem only now and through an M1 Mac won’t care much, as it’s all new territory for them, but users transitioning from Intel Macs should be made more aware of such changes.

In this regard, if you’re a regular user who uses their Mac for everyday tasks, has a minimal backup strategy mainly consisting of Time Machine backups and the occasional manual backup of the most crucial stuff, doesn’t tinker with their machine, and so forth — then the transition from an Intel Mac to an Apple Silicon Mac should be relatively painless.

Power users, developers, and assorted nerds should get an Apple Silicon Mac if they want, but I would suggest to not get rid of their trusted Intel Mac setup just yet. (I suspect they already know this, I’m just writing it down for completeness). Support for external disks seems to have been fixed in Mac OS 11.4 for M1 Macs, as Howard Oakley explains, but anecdotally I’m still hearing from people with M1 Macs reporting issues with some of their disks, and a friend recently told me via email that while he’s thrilled by the performance of his M1 MacBook Pro, at this moment he would feel uncomfortable if this were his only machine: I always feel that my experience is ‘on rails’ with this Mac, that whenever I want to try something more ‘experimental’ I’m either going to encounter some glitch, bug, or displease the OS in a way or another [winking face emoji].

Again, like we saw at the end of the Hardware section above, unless you’re in bad need of performance and your older Intel Mac is dragging you down, waiting a few more months shouldn’t hurt. At the upcoming WWDC 2021 we’re going to have a preview of the next version of Mac OS, and hopefully Apple has been monitoring all the feedback received about Big Sur, so perhaps later this year you can get an even better Apple Silicon Mac with an even more integrated and functional operating system. (Between you and me, I dearly hope Mac OS 12 is going to be a more ‘technical’ release, aimed at fixing and refining what doesn’t work in Big Sur, otherwise we’ll enter yet another round of ‘new features, new bugs, let the old bugs rot’).

What about me? What am I going to do?

This is what people often ask me when asking for upgrade advice.

I’m going to do exactly what I’m suggesting in this article — wait. I’ve always avoided the Early Adopter Train during a major Apple transition. I’m what you would call a power user, but for what I do I don’t need the latest and more performant machine, so I can play the waiting game without much hassle.

Whenever I decide it’s time to upgrade, I always aim for a slightly more powerful Mac than I need, because I plan to use such Mac as long as possible as opposed to upgrade frequently.

I admit I was curious to try an M1 Mac and was initially thinking about purchasing a base M1 Mac mini to use pretty much as a test machine while at the same time having good hardware in case I needed to perform the occasional resource-intensive task.

Then the 24-inch iMacs were introduced, and for a moment I thought I could go on and just replace my current 2017 21.5‑inch 4K retina iMac. But I’m not a fan of the design compromises Apple has chosen for these iMacs, I still think my current iMac and its assortment of ports are more versatile, and I have the feeling that I will like the new ‘pro’ iMac better anyway.

Then the recent leaks about the design and I/O connections of the purported M1X Mac mini (see my brief thread on Twitter) rekindled my interest in having such a compact and powerful solution in my setup, and at the same time are proving my point; that is better to wait if one can. If I had bought an M1 Mac mini on a whim, I’d certainly have regretted it now.

This stage of Apple’s transition to Apple Silicon Macs is like a very inviting round of starters at a banquet. If you eat too much of them, when the even tastier main courses are brought to the table you’re going to be already full.

Finally, I insist on waiting until more Apple Silicon Macs are introduced (and Apple shows their hand) especially if you’re a budget-conscious customer who can’t afford to upgrade their Mac frequently. You can save the money of an impulse purchase of an M1 Mac now and get a better Mac at a later stage.

The M1 iMacs: Unnecessarily thin

May 23, 2021

by Riccardo Mori

Tech Life

In my piece about Apple’s Spring Loaded event that took place last month, I said this about the design of the new, colourful M1 iMacs:

Apple has created what’s possibly the thinnest all-in-one desktop computer, and I’m sure they’re still patting themselves on the back. But was it necessary to produce a computer that is so thin that its power supply has to be external? I find that rather inelegant. My 21.5‑inch 4K retina iMac is less than 3 cm thicker at its thickest point, and has an internal power supply. I believe that Apple could have easily built a new iMac with this new, flat design, by making it 2.5 cm thicker and putting the power supply inside. It still would have retained a thin and elegant profile, and Apple would have spared us the external brick.

This design choice on Apple’s part keeps rubbing me up the wrong way. If the thinness of these new iMacs excites you to the point that you don’t really mind having to pay that thinness with having an external power brick, perhaps you’ll find my irritation a bit amusing. Or perhaps you’ll see it as a way to find something, anything, to criticise Apple or to be unhappy about.

Actually, these iMacs have features (or lack thereof) that are even more annoying to me, and that external power supply is just the cherry on top (or, well, at the bottom… of the desk… hidden somewhere…).

When I dislike things like this, sometimes I worry that maybe it’s a matter of age. That maybe I’m just that get-off-my-lawn curmudgeon who has seen so many designs and purported innovations in tech over the past 30 years, and therefore can’t really be enthusiastic about these colourful thin slabs. About these design statements that want to be so much more “how it looks” than “how it works” that, in order to do so, sacrifice I/O, sacrifice internal upgrades, and have to have some parts of their main circuitry placed externally because they simply can’t fit inside.

But I’m glad that someone like Marques Brownlee shares my perplexities about the 24-inch M1 iMacs. In his video review on YouTube, he’s not afraid of speaking his mind about some aspects of the iMac’s design that don’t convince him.

[The iMac] is trying to be a piece of tech that can fit into any room, and people would rather have something that looks friendlier than a metal box. Fine. Now, the chin on the other hand… This is probably where we’re going to start to disagree.

I get that it’s iconic to the iMac to have this chin, even though they got rid of the actual icon, the logo. But I just think it would have looked so much better without the chin. So, most of the computer is down there in the bottom. If you look at Apple’s videos, you can see their basic setup, which is the M1, and the logic board, and all that — most of the computer down in that chin. Then it’s just mostly cooling and speakers throughout the rest.

So, by shoving the computer down there, it let them make this whole computer razor-thin — 11 and a half millimetres thin for the whole thing! It’s thinner than a MacBook Pro, it’s thinner than an Apple Watch! This entire computer is actually thinner than the depth of a headphone jack, so they had to put the headphone jack on the side of the computer, otherwise it wouldn’t fit at all. It’s stunningly razor-thin, and it’s kind of amazing that the entire computer fits inside this sleek case — thinner than pretty much every other 24-inch monitor.

And that’s all amazing, and it clearly took a lot of work to make it this thin. But after I get it out the box and decide where I want to put it and lock it in place, I don’t care… I don’t care how thin it is. So I’m confident that nearly 100% of the people buying this machine would have been totally cool with it being, like, 20 millimetres thick, and being able to fit a lot of that computer stuff behind the display so that it doesn’t have a chin. And it would have looked way better and it still would have been impressively thin, so then that would have also given them more room for more impressive speakers. […]

They also probably could have fit the power adapter inside the computer like with all the rest of their desktops, but they really leaned into the external power brick situation with the iMac here…

Dave Lee, another competent reviewer with a good eye for design, doesn’t like the chin either, but more interestingly in his video review he mentions another aspect that’s worth pointing out. He appreciates the design of the magnetic power connector that attaches on the back of the iMac, but he also notes that the iMac is really light, 4.5 kg, and the bottom of the iMac’s stand is slick plastic, it lacks any grippy material. This, he says, is nice for sliding the computer around a little bit, in case you want to tilt the screen and show something to someone, or move the screen away from the sun, for example. But, because it’s a light machine, and there’s no grip on the bottom of the stand, if that power cable on the back is yanked, he thinks there’s a very real possibility that the iMac goes right over the edge of your table or desk.

Since Dave moves the iMac a lot in his video, you really notice just how lightweight this machine is, and just how easily it can be moved around — maybe even too easily. Even if you think the ‘yank the power cable and the iMac may fall’ scenario above is unlikely, think about other situations, like having your headphones plugged in, the headphones cable gets tangled or trapped around something (your chair’s armrest, perhaps), and you end up yanking the iMac, maybe even damaging the headphone port and surrounding chassis, given the position and the thinness of the computer. Do you think this is another extreme scenario? Maybe it is. But I have firsthand experience and other people’s accounts of this exact scenario happening with laptop computers (especially the ‘falling on the floor’ part).

Desktop computers with external power supplies

When John Gruber’s review of the new M1 iMac appeared on my feed, I went and read it right away because I was very curious to learn his take on it. Historically, all his product reviews have been thoughtful, detailed, and have often pointed out things I hadn’t considered, or made me see them in a different perspective, which is the kind of intellectual input I really love.

But this review was underwhelming and, as I commented on Twitter, with unusual fanboyish tones I’ve never really detected in his past product reviews.

His argument defending Apple’s choice to make the iMac so thin as to have an external power supply is unconvincing:

Is it a cheat to make the power supply external? Sure, maybe. But great designers know how to cheat right. The external power supply could have been internal, glommed onto the back behind the hinge somehow. There’s no doubt in my mind that this actual design — with the external power supply enabling the iMac itself to be as thin as possible and universally flat across the entire back — is a superior idea. The power supply will, in most people’s setups, be hidden out of sight. Why not hide something that’s even just slightly ungainly? All sorts of ungainly stuff, everywhere in life, is hidden out of sight. That’s a design cheat.

This reads like a brief from Apple’s marketing department given to a sales rep.

The external power supply could have been internal, all right; the iMac would have been a bit thicker, but still thin enough to be a striking design. It would have worked. It would have produced an equally elegant iMac with all its ports in the main unit. It probably wouldn’t have been a show-off design like this, but really, I don’t see any other reason behind the iMac’s thinness other than, We made it this thin because we can.

Is there any usefulness behind this absolute thinness? No. While I never liked thinness for thinness’ sake in computers, I could understand the argument in favour of having a thin laptop because it also means having a lightweight machine you can easily carry with you anywhere. In machines like the MacBook Air and the 12-inch retina MacBook, thinness can be considered a feature. In a desktop computer? It’s merely wow factor. It’s just looks. Is it good design? I don’t think so. Gruber talks about ‘design cheat’. It’s actually a compromise.

Gruber:

Why use an external power supply with a non-portable desktop computer? I think that’s the wrong question. The right question is: Why should MacBooks and iOS devices be the only devices that look as thin as possible? Apple has, in fact, done this before, but only sporadically. The 20th Anniversary Mac (1997), G4 Cube (2000), and the first few generations of Mac Mini (2005–2009) all had external power supplies. These M1 iMacs, I suspect, won’t be the last.

Well, in the case of the Twentieth Anniversary Macintosh, this machine featured a unique sound system by Bose, with integrated stereo speakers in the main unit, and a subwoofer built in an external Bass unit. The power supply necessary to power both the Mac and the subwoofer would certainly have been too bulky to be put in the main unit. Since the Mac came with an external subwoofer, it simply made perfect sense to put the power supply there.

The Power Mac G4 Cube came with a big and heavy power brick. You can see its size compared to the Cube’s in this image (taken from eBay):

This power supply wasn’t external just because this way Apple could make the Cube as compact as it was. If this power supply had been an internal component, the Cube wouldn’t have been just taller — its whole design would probably have been different and, essentially, much more similar to any other tower-shaped Power Mac G4 of the time. This power supply is very powerful, capable of powering both the Cube and an attached Studio or Cinema Display with the ADC connector. A power supply like this gets hot and needs proper ventilation. By making it external and in a perforated chassis, Apple could make a smaller and silent desktop setup (remember, the Cube was passively cooled like the iMac G3; it didn’t have an internal fan) where neither the Cube nor the power brick needed active cooling. This is a much better example of that ‘design cheat’ Gruber talks about.

In the case of the Mac mini, the reason for having an external power supply is right there in the Mac’s name. The point was to keep the Mac as small as possible because that was (and is) the defining feature of the Mac mini. Apple then managed to put the power supply inside the mini, because in later Mac mini units the internal components were smaller or removed (like the optical drive).

So, let’s recap:

In the case of The Twentieth Anniversary Macintosh, the benefit of an external power supply was to allow the Mac to have a unique and powerful sound system. It wasn’t just a matter of having a good-looking machine.
In the case of the Power Mac G4 Cube, the benefit of an external power supply was to have a compact and silent desktop machine that was still capable of directly powering ADC displays with just a cable, so that the user didn’t have to deal with two power bricks, one for the Cube and one for the display.
In the case of the first generations of Mac minis, the benefit of an external power supply was to have a small desktop computer. It wasn’t a purely æsthetic reason. The Mac mini was not designed to look cool, but to be usefully compact. Its compactness and small size is its defining feature and what makes it still a versatile machine today (Macminicolo will surely agree).
In the case of the new M1 24-inch iMacs, the benefit of an external power supply is… what? That so you can have the thinnest desktop machine ever? Okay. And that’s useful, how? Thin and light, some will argue. Sure, a light desktop computer is great when you need to take it with you when working out and about— oh, wait. Yeah, yeah, the answer is: the benefit of an external power supply in these M1 iMacs is that so they can look cool. It’s the looks. Just that.

Gruber:

The result of putting the Apple TV-sized power supply out of sight is that the M1 iMac itself looks like it’s just a display. Maybe that’s true of the Intel iMacs too, but the M1 iMac not only looks like it’s just a display, it looks like it’s a very thin display.

I own a 2017 21.5‑inch 4K retina iMac. It doesn’t look thick when I look at its profile. The curved, tapered design makes it look like a thin display indeed. Sure, not razor-thin as the new iMacs, but still, it’s only 5 mm thin at its edge. And it’s a very balanced design overall, both in terms of visuals and weight. It’s not a heavy desktop machine (it weighs 5.6 kg, essentially 1 kilogram more than the M1 iMacs), but it has sufficient heft you don’t feel it may topple if accidentally hit or if it gets yanked by a cable. And it has a very useful array of ports: four USB‑A, two USB‑C Thunderbolt 3, Ethernet, and an SD card slot.

The design of the new M1 iMacs… I don’t know. I even read someone who used the word daring, but what is the dare? Who gets to make the thinnest all-in-one desktop computer? Good, now what? What’s the purpose?

You may say, Sometimes things are just designed to look cool, and that’s okay, I guess. I think it’s a philosophy that works best with objects that aren’t everyday tools, but that’s me.

But even with some past Apple computers where it was clear that the primary purpose of their design was to look different and distinguishable from other PCs (think of the iMac G3 or the first colourful clamshell iBooks, or the Power Mac G3 Blue & White, and the following generations of Power Macs), there was at least one element in that design that made it also practical and smart. A common denominator of all those past Macs I’ve mentioned was their handles. All those machines had smartly integrated handles to lift and move them around. There was more of course, but even this small detail was enough to give their design a certain depth and purpose. These new iMacs are as razor-thin as their design is razor-shallow.