I've wanted more restrictive and granular permissions for a long time. Web browsers don't need unrestricted access to the entire file system. Neither do games. Preview doesn't need to access the microphone. Word doesn't usually need to record the screen. PowerPoint should not be sending email or accessing the camera. TextEdit doesn't need access to my contacts or calendar. And I don't want any of these programs interrupting me with stupid notifications. Maybe some people want to give these extra permissions, but I want them to be off by default.
My complaint is that permissions don't go far enough; for example I never want to be interrupted by software update notifications, but there's no way to turn them off.
I'd prefer sensible (silent) defaults, such as disabling PowerPoint's access to my webcam.
Like Apple, you are thinking App-centrically. Stop. Apps are not the center of the users’ universe. The users’ own tasks are.
Therefore you need to think task-centrically—i.e. What does the user need/want to do?—and structure your security model to enable and support that.
This is the case on _every_ OS that lets users install applications.
So, how can we make sure users doing tasks aren't afflicted by applications that want to sneak in? You're saying to "structure a security model differently" leads me to think maybe users shouldn't have admin access on the machine? But then you get all the critiques people have of iOS.
I'm honestly asking, what's the alternative here?
Locking down the entire UX doesn’t work because day-to-day operations become onerous while the constant spew of “Are you sure?” dialogs teaches users to click OK without thinking.
Instead, put the sensitive stuff into its own “secure desktop”, and police all movements between that and the “standard desktop”. Even stick in an “unsafe desktop” for doing risky stuff, and encourage users to install new/untrusted software there; e.g. let them install and run anything without nagging, but keep apps completely sandboxed. Boundaries should be clearly delimited, e.g. give the high-security desktop its own window chrome and fixed background image that can’t be mimicked on the others.
It’s deliberately coarse-grained, unlike the current mess, and that’s the point. That makes it easy for users to mentally model and match scopes to tasks, and limits nagging to where it really matters: when crossing scope boundaries with data in tow. Simply switching scopes and working within a scope should not nag, however.
Look, I’m not saying it’s guaranteed to work. But at least I’m thinking about it from the point of view of end-users, not vendor developers and lawyers. It’d need prototyped and tested in both lab and real world, and sure as guns users will discover exciting new ways to screw things up; which means learning and adapting to account for that.
But that’s UX. Either do it right, or go get a job flipping burgers where the worst you’ll do is hospitalize a few dozen locals with E. Coli. TWO BILLION now trust large chunks of their lives to their “personal” machines. That is a lot of people to screw up.
As an "expert" software engineer, I completely trust absolutely zero software running on my computer, in my browser, or basically on any device anywhere. I begrudgingly put up with it all because there is no alternative. Adding friction is good because it will eventually push (force?) developers to use more modern, safer development patterns.
[0]: https://blog.robenkleene.com/2019/08/07/apples-app-stores-ha...
I have a bunch of Applescripts which start with:
tell application (path to frontmost application) to
...so they can control whatever is on screen at the time.Since Mojave, any time a new application happens to be in front, I get prompted to give Bash permission to control that app. And then the entire set of approved apps seems to get wiped sometimes when I edit the Applescripts. There's no way to approve apps ahead of time, and there's no way to blanket approve control of any app.
Good point. Apple doesn’t care. If they did, they would’ve put some fricking effort into it. Not only is 10.15’s “Automation” security model pathetically thin and annoying (I’ve used those APIs; they’re crap), it doesn’t even protect users as it claims.
Allowed a script running in Terminal.app permission to access your Contacts? Great, now ANY script that runs in Terminal can do whatever it likes to your address book.
Once again, anything that does not fit cleanly into Apple’s naive App-centric model gets neglected, overlooked, or stomped on.
Honestly, I think now they’re just waiting for AppleScript usage to die out on its own, so they can kill off the old and crusty (but very powerful) Apple event IPC system entirely. So it goes.
--
TL;DR: Apple is in the commercial business of selling Apps. So where’s the benefit to them in you making your own?
:(
Or more likely it forces the retirement of known good software, and requires you to use unknown, newer tools that have had much less time for security and behaviour auditing my the general community over the years.
I haven’t reenabled it yet, and I don’t think I will be doing for the time being. Apple’s gone too far here. Protect me, sure. Make me take an extra step to run an unsigned binary. For critical modifications like installing kernel extensions, make me run a terminal command, that’s fine. These measures would assist me in maintaining my own security, and they certainly shield my computer illiterate mother. But treat me like a hostile actor on my own computer and strip me of my superuser privileges? No.
I’d rather be a pirate than join the Navy.
Similarly, an excess of popups doesn’t increase security. It annoys users to the point they stop thinking about individual security decisions.
There’s an important intersection of security research and HCI that doesn’t get discussed as much as it needs to be.
Similarly, an excess of popups doesn’t increase security. It annoys users to the point they stop thinking about individual security decisions.
It's particularly annoying b/c there's a subset of craftsmen developers who pay close attention to this problem, who realize these cures are worse than the disease, and who reverted years ago to not using these flawed security patterns. Many here on HN, for example.
And then there are Apple engineers who seem to be so far behind the curve that they're just now discovering and implementing patterns that others not only discovered and tested years ago, but discarded years ago.
It's like corporate 9-to-5'ers with no pride of craftsmanship or motivation to stay current with their field made these decisions. Either that or clueless management ordered it, which in any other company besides Apple would be my first guess.
QFT. Because there are not enough upvotes in all the world.
The annoyance is still there/potentially larger, but it's easier to make an informed decision about something specific, rather than the usual "enter your admin password to allow... potentially anything, probably persistently. Or don't, then no app for you." binary which seems easier to just blindly go along with, since auditing is the only way to find out what exactly the privileges are for.
The question is rather where sheer request volume overwhelms gains from actually knowing what's being asked for, but better tools letting user set a base comfort level as far as common requests like disk access etc should help with that.
I am however struggling to articulate my opinion in a way that doesn't boil down to essentially "the issue with the dreaded Vista popup was that there weren't ten different varieties!", heh. Nah, just comes down to asking once vs. every time.
It's not exactly new in Catalina, but it's worse. Apple just does not even think about the fact that you might want to do things remotely or automatically.
The execution is where Apple is fucking up, yet people are hating on the concept, as if poor UX is somehow mandatory for something like this. Worries me a bit.
I believe there should absolutely be intentional, physical, steps to allow full remote access. Intentionally setting the terminal for full disk access is no less crazy than intentionally exposing a port through your firewall.
I'm not "upset" that it needs authorization. I think it's a design flaw that you can't do it via a terminal, you don't even know that that's why your process has stalled, and you have to click a box every time.
The problem I see with these current changes is that they appear to be reactive—“X, Y, and Z are potential threat vectors, so we need to lock those down”—rather than taking the long view—“what does the future security model need to look like and do?”—and plot a clear, reasoned course to that from where we are now.
The result is an Application-level security model, when what users need is Task-level security. Users still get all of the pain of change, yet none of the intended benefits—all that’s happened is we’ve gone from one unfit-for-purpose platform to another.
For instance, just consider high-security operations such as internet banking. No way on Earth should we be using the same web browser application for sensitive financials as we use for reading all the daily gossip, browsing teh prawns, etc. Split that out into a dedicated, hardened version with no support for extensions, external automation, and its own isolated password managment. Whitelist security-sensitive URLs so that they open in that version, and restrict opening common-or-garden URLs to the low-security version only.
Isolating sensitive User Tasks within their own “high-security desktop” environment would take a lot of the security pressures off individual apps, obviating the need for these endless “Do you want to grant permission for X to access Y to do Z?” requests. Which, as any security fule kno, are worse than useless at protecting users; protecting only the platform vendor from taking the blame when the inevitable fuckups occur (aka Blame the Luser).
That’s the user-security model Apple ought to pursue, but they’ve been stuck so long in their classically atomic Application-centric view of the universe I don’t think they can imagine a platform where atomic Apps are subservient to the big smooshy probability cloud of user tasks that surrounds them. Partly because that nice simple neatly-packaged atomic App model is what generates revenue, but also partly because it requires giving up control; and both developers and marketeers like keeping control very much, and fear very much ceding it to anyone else.
That takes courage, and Apple hasn’t had that sort of courage for a decade.
But putting Control into the hands of Users, in a form that those users can use, is the very core of UX, and nowhere more important than protecting users from “themselves”.
> But I don’t know a single expert Mac user who is not seriously annoyed by the heavy-handed security design of Catalina. Not one. Every single expert user I know is annoyed. That is a bad place for MacOS to be. MacOS 10.16 needs a serious course correction to fix this, and if 10.16 goes the opposite way — growing even more heavy-handed in restricting professional Mac users from just using their machines as they want and expect to — I genuinely fear for the future of the Mac as a platform for serious computer users.
https://daringfireball.net/2020/02/my_2019_apple_report_card