going "ehh these complexities don't matter in the context i'm working in" is the same as going "ehh these people don't matter in the context i'm working in"

I go – somewhat controversially — further than that. If you inadvertently do that, and your software incidentally doesn't support other cultures well, that's one thing. Happens to all of us. But if you repeatedly brush off the problems Unicode was trying to solve because you think they're not important to you, that's a form of racism.

Almost invariably, it's people who write in English, or in another language using the Latin script, making assertions of the "why is Unicode so complicated?" or "it could easily fit in 2 bytes if you leave out all the needless complexity?" kind, where "needless" is needless for them, not for the people that complexity was designed for. Really, when you interrogate that any further, you realize what they're asking is "why won't those people just write in English instead?"

I sympathize that getting Unicode right is hard — so hard that most languages and frameworks don't even implement something like a string.getLength() API "correctly". They look at amount of bytes, or amount of code points, rather than what you probably want (but is much harder to compute): amount of visible glyphs on screen.

But just because it's hard doesn't mean we shouldn't strive to do it.

These pseudo-C# annotations to assembler look nice. For example, add r8d, 64h is commented with r8d += 100, which far more closely matches the way I think.

The feeling that others have used your app far more than you ever will is quite amazing.

I briefly dabbled in using WebAssembly (mono-wasm¹) to add .NET support, and made a proof of concept, but never quite bothered to optimize and merge it into some public website.

IHateRegex looks neat, but has significant differences from .NET (and doesn't appear to support named groups at all, a feature I use heavily), and has some very odd/annoying behavior with arrow keys, so, back to one of the two others I go.

(via)

Today in features so hard to discover, most users (including me) will conclude they don't exist, so why even bother implementing them at all.

(For me, it only shows tabs from one device, which is already broken enough, but also, that device hasn't been my current phone for weeks.)

I'd love for someone to benchmark the result of building Roslyn.sln on:

• macOS, running dotnet CLI
• Windows on ARM, virtualized, running dotnet CLI
• Windows on ARM, virtualized, running Visual Studio 2022, emulated

Haven't yet gotten a clear picture of just how bad the performance drop is.

(Assume a 32 GiB RAM machine with, say, 16 GiB reserved for the VM. Maybe 20, if that helps.)

I remember noticing this with my 2013 MacBook Pro back then, and I also noticed this much more recently with my iPhone 13: Apple really invests in those internal speakers, and the 2021 MacBooks Pro are no exception.

Why they care so much, I'm not quite sure. Maybe because they can. Those speakers aren't good enough for professional needs, and perhaps never will be. But perhaps they're just good enough now for professional preview purposes?

When psychological safety exists in a team, people will say more things like:

• I need some help
• I have an idea
• I think we have a problem
• I disagree
• I'd like to challenge our ways of working
• I made a mistake
• I made a really big mistake

If you use Homebrew on your Mac, you can create a Brewfile and significantly reduce the time to set up your next Mac. I haven't done this, but it's tempting. Here's Casey Liss's.

(I have done the equivalent on Windows, using Boxstarter, which in turn uses Chocolatey. Especially neat on servers, I find.)

(via)

Mac OS Classic had an app¹ called Scrapbook. I never really understood its point as a kid at the time, but Uli proposes that it was an early form of multiple clipboards. With that, the name and concept suddenly clicks.

Modern macOS still has Finder's Edit → Show Clipboard menu item (is that where you would have expected it to reside?), but other than syncing the clipboard across devices, it hasn't changed much.

This 2004 e-mail thread between Microsoft execs, frustrated over the relative progress of Windows Codename Longhorn (a.k.a. Vista) compared to Mac OS X 10.4 Tiger is interesting, but it also raises some questions on whether the execs even knew what they were talking about. Longhorn, of course, was a troubled software project in some of the same ways Apple's Copland (the original, never-shipped Mac OS 8) had been decade earlier.

Vic bemoans Spotlight being a WinFS rip-off, but Spotlight's focus is on search, and WinFS at the time didn't even have search.

Vic also thinks Core Video and Core Image are an Avalon competitor, but Avalon (a.k.a. WPF, the Windows Presentation Foundation) is a UI framework akin Apple's Cocoa/AppKit, whereas Core Video and Core Image are libraries to help build video and image editors; apps like Acorn and Pixelmator build on Core Image.

In many ways, the thread seems to miss the point. What the Mac OS X team at the time did was this: they shipped, and then iterated, and then shipped again. The Windows Vista team, did not; instead, they kept expanding the scope. What was supposed to be a stop-gap release before Blackcomb instead subsumed all ideas and the kitchen sink, including pie-in-the-sky projects like WinFS. Meanwhile, Mac OS X was incomplete: it lacked major features. 10.0 did not play DVDs or record CDs.² 10.1 was extremely slow. 10.4 still did not have a good answer for "how do I backup?". And so on. But almost every year, Apple iterated, and delivered, and it felt to users that things kept getting better.

If the Vista team had shown some humility and focused on features that were actually ready and polished, much like the 7 team successfully did later on, its story might have been quite different.

Was Vic too far removed from the Vista team(s) to understand its features? Or, worse, should he have had a better grasp of them but didn't?

I genuinely think (and thought at the time) that WinFS sounds intriguing on paper. All it needed is a good user story.

The San Antonio, TX zoo is debuting a jaguar skywalk, and I feel like I've seen that movie in 1993.

What if Cyan made a game set between Myst and Riven and called it… Mydst?

Speaking of Mac OS X, at some point, its Help menu gained a search feature, which goes through the entire menu bar of the current application and helps you discover features. The usefulness is fantastic, and the way it looks is cute.

What if the notch on the MacBook Pro caused the mouse cursor to obscure the camera?

Oh yeah, it's fall.

Rebrands rarely work.

The good news about Meta is that these kinds of rebrands never work. Philip Morris will never be known as Altria. Alphabet will always be Google, divisions be damned. Xfinity will always be Comcast. Qwikster, well, Netflix was just smart and killed Qwikster.

(As a somewhat tongue-in-cheek counterpoint, I bring up that nobody calls IBM by its original name any more.)

Steve Shives discusses how Trek fans are a fairly broad spectrum, and some may find Lower Decks very funny, while others barely react at all, and that's OK.

Personally, of the Kurtzman-era Trek shows I've watched, I've found Lower Decks's characters the most intriguing so far; certainly more so than most of Discovery's, but also more so than many of Picard's. I want to know more about people Boimler and Mariner and the others. And I occasionally find the show quite funny! However, having watched all of season 1, I haven't found any of the stories particularly interesting, and some of them outright dumb.

And yet: I'm happy there's lots out there for others to enjoy.

The first message was sent over ARPANET (which became the Internet) 52 years ago today.

To be clear, the main application processor continues to be ARM. Moving a dedicated chip to a different ISA is much easier, as you don't have to support existing (or any) third-party code. The main process, OTOH, has to run existing Android apps, and moving those to RISC-V isn't quite as practical — yet.

I suspect we won't see Apple moving to RISC-V. Rather, if they do move away from ARM — such as because they're unhappy with Nvidia having bought them, or because they disagree with further the evolution of the ISA — I expect them to diverge, slowly at first and then very rapidly. They've already started making custom extensions, and there's no reason we can't eventually have "Apple Silicon" that is in fact quite dissimilar from ARM. Just not any time soon, as Apple does benefit from compatibility with existing binaries. Look for this a decade or two from now.

(via)

I've been lucky in the 2+ decades of using PayPal, but I've always hurt of the capricious ways in which it screens transactions or outright bans accounts.

Such as, say, when you mention a trip to Cuba.

Or a Persian rug.

The MacBook Pro "notch" is the gift that keeps on giving. This time: what if you don't just split the menu bar but rather the entire screen?

TopNotch (via)is a tool that makes the notch nearly impossible to see, by turning its surrounding menu bar black as well.

Xiaomi phones apparently have a similar setting built in.

Depending on how popularity of this feature plays out, I wouldn't be shocked to see a first-party option in 11.1 or 11.2 that simply shifts the menu bar down.

Dropbox still doesn't have an ARM-native client on macOS.

Some of the early staff responses in that thread aren't great:

This idea is going to need a bit more support before we share your suggestion with our team.

It's not an "idea", and there isn't really any discussion to do, other than on priorities. But it's been a year and a half since Apple announced the transition.

A later comment announces a timeframe of "first half of 2022", which is not a great look, but better than nothing.

I imagine Mac users are no longer the key demographic Dropbox is focusing on, and haven't been for a while, but such a response will only further accelerate people moving away.

Not great.

In more bird news, you don't quite expect this wingspan.

The bird also seems quite aware of how impressed the humans are.

Best as I can tell, there are multiple issues at play:

1. if there are more menus (the left-hand portion) and status items/menu extras (the right-hand portion) than fit in the menu bar, the overflow strategy macOS employs is to keep removing status items until at least all menus fit; i.e., those status items cannot be seen or used, but at least the current app can be used in its entirety. This strategy, probably introduced in Mac OS X 10.10, was never great, and maybe it's time for Apple to change it. But it's not a notch- specific issue; the notch just exacerbates it because it removes horizontal space. If your screen resolution were low, you would equally have run into it sooner or later.
2. if the menus don't fit to the left of the notch, macOS will continue to the right, which arguably looks and feels a bit awkward.
3. if the application does custom drawing in the menu bar, macOS won't handle that at all. In that case, there may be glitches, and for affected apps, users can tick the "Scale to fit below built-in camera" checkbox. This checkbox is only intended (and visible) for apps whose developers haven't already said that their app is compatible with the notch.
4. lastly (and keep in mind: none of the above is a bug, just behaviors not everyone agrees with), we do have what appears to be an outright bug on Apple's end: if a status item is very wide and some of it fits, but some of it overlaps with the notch, macOS will partially draw it, rather than discarding it.

For 1, tools like Bartender and Vanilla exist. Those also make 4 largely moot. I'm hoping for a first-party solution here, though.

I don't think any of this makes these computers "not shippable" or a poor "design", as Quinn seems to imply by his two tweets.

Regrettably, the macOS TV app still has a lot of issues, including GUI glitches.

I was able to reproduce both of Uli's — the minimum width is too small to properly fit the Search field without overlap, and, on a Retina Display (a.k.a. on virtually all of Apple's recent computers), when dragging columns to rearrange them, they suddenly double in size.

Another macOS annoyance I run into a lot is how it handles desktop backgrounds. Really, all I want is to set, for all displays, a solid color that matches my mood. This seems like a System 7 (1991) kind of problem, but I cannot reliably get macOS 11 Big Sur (2020) to do this.

First, the UI to set backgrounds is glitchy: just the other day, the Folders section in one display was ordered differently in the other. And the "well" in which you can drag an image reacted for all displays, but for one of them, the actual picture didn't change, until I clicked somewhere.

But, more annoyingly, any of the three displays will, at random, show a different color. I click a shade of blue for all three in the morning, and by noon, one of them will show yesterday's shade of forest green.

I get the impression that Apple views this as a rather frivolous feature, and that bugs in it don't matter. In the grand scheme of things, they don't. But a company that prides itself on how good the selection of Apple Watch bands is should also be able to show the correct desktop background.

Adobe is adding NFT support to Photoshop. This is, of course, easy to circumvent and mostly an exercise in pointlessness.

Now, if you do want some kind of digital watermark in your image, Photoshop does in fact have tools for that.

Universal Basic Income trial runs come and go, but Los Angeles doing one seems like a fairly big deal. 3,200 households get $1,000 a month for a year.

(via)

Because software rarely operate on "files in folders" anymore, "export" is increasingly the way software exposes data. But usually you don't want a dead snapshot; you want to "use this data elsewhere"—which requires repeatedly exporting & reconciling.

Silos and exporting are less convenient than files if you want multiple apps to interact with the same data.

(via)

A pre-release benchmark shows the Intel Alder Lake-P Core i9-12900HK outperforming the M1 Max, by 3.7% at single-threaded and 3.9% at multi-threaded tasks.

Surprising, but also quite possibly misleading. What does it take to actually reach that performance? Does it fit in a chassis that passes as a "laptop"?

I suspect these are lab numbers. I do not expect a Dell XPS or a Microsoft Surface to achieve them. Plus, by the time these ship in volume, we might already see an Apple M2 — likely with poorer multi-threaded performance, but perhaps with better single-threaded performance.

That said, competition is good, and I'm glad Intel is seeing the light on a heterogenous setup.

(via)

Almost without fail every year, there's a rumor that an upcoming Apple Watch will feature non-invasive blood glucose monitoring.

Perhaps where there's smoke, there's fire, but also, perhaps wishful thinking rather than actual information is driving the rumor mill here.

(via)

A Twitter rulebook.

I have a few quibbles with it, but perhaps rule 3 applies:

Never argue with anybody about anything. Really.

The iPod team delivered a shipping product within about 5 months of Tony Fadell joining.

That's mighty impressive. A "startup"-like team within a company with the kinds of resources a startup wouldn't have. I cannot help but wonder if Apple would still be this "agile" today with a new product launch. That's a tough needle to thread.

(via)

If you really needed that level of performance (you almost certainly did not), you could buy the $2,000 Apple Afterburner card for your $6,000 and beyond Mac Pro two years ago, and accelerate some specialized tasks, such as ProRes.

Today, the M1 Pro has that built in, and the entire 14-inch MacBook Pro costs $2k.

The march of progress.