I think this is less a "macOS lacks the feature to prevent focus stealing" and more they never even imagined the scenario where it takes time for the app to open. Instead of even considering that a user may add inputs while waiting for launch, potentially signaling that they are no longer requesting focus from the program they clicked on, they simply designed for the optimistic case where there never IS time for the user to give these inputs before their intention of opening the program is completed. Not saying it's any less bad from a user's perspective, but I think there's a difference in mindset here.

In my experience (admittedly a bit dated now) Mac apps always take a few "bounces" to launch and be ready. There are a lot of nice things about MacOS but snappy application launches isn't one that stands out for me.

Having owned a good number of Macs and one hackintosh in the past, that would be very bizarre assumption to overlook. App icons bouncing while something loaded was never a rare thing.

IMO, nothing should ever steal my keyboard focus. As soon as I've typed a single character, whatever happened before, too late for you.

Seems like the programmers at LinkedIn never imagined that the network isn't reliable, and can't invalidate the session on actual logout; it only happens when whatever page is replacing the logout page finishes loading.

Is it could be a bug (IMO yes, a credential disposition / security bug), or it could be a "feature tax" to keep the tracking context hot for whatever page you logged out of LinkedIn to avoid being tracked...

It’s probably the lack of distinguishing between a window obtaining focus because the application just launched and the window of a running application requesting focus for some other reason.

Nevertheless, the OS could conceivably track whether there was any focus change since the process was launched or since it received the last key-press or mouse-click event, and only allow the focus change when there wasn’t (and otherwise use a non-focus-stealing notification mechanism).

As always, Apple has designed for the happy flow, and nothing else. Mac apps have never opened instantly, this is nothing new.

They can open nearly instantly, but they need to be well-built native apps. Lumbering behemoths like Adobe CC or anything Chrome-based is going to be a bit slow, and even smaller apps that pull in a big foreign UI toolkit are going to take a tick or two longer to start.

It’s too bad that the art of building lightweight applications is becoming a lost art.

same with iphones and switching between cellular and wifi connections. I'll be walking away from or toward my home or my office and there's just barely a wifi connection, but hardly enough for any requests or responses to even make it through, yet Apple never allows the phone to fall back toward the cellular connection.

I can't believe in all these years they haven't implemented a dual setup where they connect the wifi and do some kind of test connection or request to confirm it's a stable signal before entirely switching away from cellular data.

> Apple never allows the phone to fall back toward the cellular connection.

Maybe you’ve turned off WiFi Assist? It’s an option to allow mobile data to supplement WiFi.

Nope. it's enabled.

That's the thing. I think parent question should have been the opposite.

If it's enabled, it tries as hard as can be to 'assist', even if the Wifi is bad, and it should have given up a long time ago.

On the current M series hardware, in my experience, it's rare for the icon to land from its first bounce before the software is up and ready.

Focus management has never been Quartz's strong suit.

If anything, they've already gone above and beyond by allowing users to even make inputs while an app is launching.

Perhaps its a slight overcorrection but it's better than the past where an online shop I gave my credit card info to would reply to a forgotten password request with an e-mail with my password in it in plaintext. Security used to be awful and people suffered because of it.

I wonder how many accounts have ever actually been compromised by a MITM reading passwords out of plaintext emails. I would guess it's very, very small.

I believe the bigger failure mode in that scenario is the fact that the site was storing user passwords in plaintext (either directly or effectively). Users tend to reuse passwords, so when there's a breach, more than just the user accounts for that site would be compromised.

No site should be able to give you your password because they should never store it. They store the hash of the password and whatever password you provide is then hashed and compared to the stored hash.

That way, when their security is compromised, the attacker doesn't get your password.

Any site giving you your password is doing it wrong because they should never know it.

NSA probably read and stored a highly significant fraction of those emails.

This is a very odd take in the world where your device will connect to any old wi-fi.

On top of that any takeover of a device like an ISP router or BGP misrouting is now an automatic compromise.