Goes well with their equally horrible hardware.
Time to realise the Mac era is over. Sadly. It hasn’t been a focus for the company in over half a decade.
This is using Catalyst, right? UIKeyCommand is a UIKit name, not an AppKit one. And Catalyst doesn't exist prior to Catalina. They've added a whole other (temporary?) operating system. It's perhaps not surprising this has a lot of bugs. Classic wasn't perfect, either.
That won't help you escape Catalyst bugs, except that you won't be able to run any Catalyst apps. But you can just stop running Catalyst apps, now, without reinstalling your OS.
I managed to do it during the beta. Since Catalina installs the OS on a separate partition inside an APFS container, I wiped the OS partition, then booted into one of the recovery modes (there are 3, I think) to reinstall Mojave on the partition containing my data. No data had to be wiped, and most kept working. Music library won’t work with iTunes but I got around that with Apple Music iCloud Library. A few other things won’t work, like Mojave Reminders won’t sync with iOS 13, but otherwise it wasn’t so bad. Think I had to reinstall Homebrew and Node and that was about it. After that I just removed the now-blank partition that formerly housed Catalina and renamed my data partition back to its regular name.
Unfortunately I do not recall which recovery mode I booted into. I tried a few, each offering a different OS to install. One would reinstall the current OS, one would reinstall the version that shipped on the Mac, and one would reinstall the current public build, I think. Wish I had a better recollection of which it was, but since Mojave is no longer the latest public build it may be moot. If you have the Mojave installer downloaded you might be able to just install it right on the data partition and wipe the Catalina OS partition afterwards.
Whatever you do, have a full backup before you even think about starting just in case, of course.
I am a fan of Twitter. It is a much better social network then Facebook and Instagram. But when it comes to development quality, Twitter reminds me of Ebay. The technology just does not seem solid. Every other month, something goes wrong.
I wonder if there is a systematic difference between companies with solid development and companies with wonky development. Is there a different hiring philosophy in place? A different type of tech stack? A different management structure?
Just a week ago, Twitter lost over 20% market cap because they noticed that they were showing people device-personalized ads even though they were opted out of that. After they fixed it, revenues went down.
In other words, a stunning 10 billion dollar of Twitters market cap was based on the illusion that the tech was working as expected. While in reality it was wrecking havoc.
Is it really possible that a feature that makes up a quarter of the companies value is not being tested?
I've thought a lot about this, having worked at some of these companies with both good and bad quality. The best conclusion I can come to is whether or not the executives use the technology on a regular basis.
At eBay, the execs didn't use the platform much. In fact, when I worked there, they had to give us a $10 credit just to get us to use it (and then stopped when the users said it wasn't fair that employees were bidding with someone else's money).
At Facebook/IG, the execs use the platform often. At Netflix the execs use it pretty much every day. At Twitter, you don't see a lot of executive use.
You can even see it at Apple. The quality on MacOs has been falling, quite possibly because all the execs are using iPads and iPhones as their daily drivers, and rarely using laptops.
I think in the end it boils down to visibility to people who can affect change. Either your execs need to use the product often, or the engineers need to be empowered to make changes without exec approval (or both, like at Netflix or Facebook/IG).
Are you implying that Netflix software quality is good? As someone who uses the web interface in a variety of browsers, I disagree in the strongest terms possible.
I think you may have inadvertently proved my point. The web interface is one of the least used interfaces. It probably doesn’t get used by many execs or engineers in the company.
I’m taken aback by all the people saying they never use the web interface - what in the world do people use, then? Is it really true that people overwhelmingly use Netflix on tiny phone screens?
Edit: I am stupid and forgot that smart TVs exist (despite owning one)
The TV mostly. Many TVs have built in Netflix apps, and also lots of thinks that you connect to a TV like AppleTV, Roku, Firestick, Blu-Ray players, etc.
I haven’t used the Netflix web app in years. I can’t think of the last time I’ve seen others use it. Perhaps the execs don’t use the web app? The iOS apps work great for me. Especially the offline download feature- great for planes.
Anecdotally, I’ve only seen or heard of one or two people in recent years use Netflix on a web browser and it being a regular thing for them. I’d expect the vast majority to be via TV or mobile devices.
The bar for Linux is “does it work at all?” If it’s not completely broken, then everything else matters little, because it could just as easily be broken. (Source: daily Linux user for... like ten years now?)
> At Twitter, you don't see a lot of executive use.
Umm? I worked at Twitter for 4.5 years and execs used it a lot. The _board_ largely didn't use it, but everyone else did, to sometimes amusing results.
> The quality on MacOs has been falling, quite possibly because all the execs are using iPads and iPhones as their daily drivers, and rarely using laptops.
iOS, iPadOS, and watchOS were pretty awful as well, and presumably every executive is using those. And I can think of at least a handful of senior executives who probably use Macs daily–they're also probably the ones with the most direct control over the people who can fix issues in the software they're using.
> At Netflix the execs use it pretty much every day.
If true, they need to be fired. If they experience how terrible the Netflix UI is and refuse to fix it, they're guilty of either gross neglect or sabotage.
I'm honestly shocked around how bad the experience is on the site. Mobile video especially. 30% of the time the video says "my account may not be allowed to see this video", even while logged out. Refreshing the page fixes it. But even when the video is loaded, it takes 2-3 taps on the play button to actually play.
It's not even the edge cases that they have issues with, it's the basics on the site. There are some serious quality control issues happening at Twitter.
Twitter... is a much better social network then Facebook and Instagram.
Off on a tangent: Are to elaborate? I've found Twitter extremely hard to use - people tweet and retweet so frequently, it always looks like a deluge of short, mostly meaningless, fluff. At least with Facebook, the longer bits of fluff are in a single post instead of spread aross and intermingled with other fluff.
But, maybe I'm doing it wrong? I really don't know. These days, I mostly stick to Instagram - it doesn't seem to have the negative impact on my psyche as either FB or Twitter.
The way Twitter works for me is to select very carefully whom to follow. And to turn off retweets from everybody. If somebody tweets fluff, I immediately unfollow them.
It completely escapes me, what you can do with Instagram. You cannot do text posts. You cannot put links in your posts. I have never seen a single intelligent discussion taking place on Instagram. What do you use it for?
It's not a platform for long form discussions or sharing links (That's why you can't even use links except one in your bio and everyone saying "link in bio" if they want to share something. There's even dedicated companies working around these restrictions and providing redirection services etc.
I also only follow people with very high signal to noise ratio, and it works for me. I will use the option to "Turn off retweets" from people who retweet useless fluff but tweet good stuff.
What I would love is a button that temporarily removes the the retweets and/or tweets of a person from my timeline. Some people will on certain days get really passionate about an issue I don't care about and don't want to see 20 tweets from them, when their daily average is 1.
I think the knee jerk joke would be Twitter rarely has intelligent discussion too. Obviously that’s not true and I’m sure you follow people who discuss worthy things. But the toxic parts of Twitter are so bad, which is what most of it is, is so bad, unless you’re keeping on top as much as you are, it seems like you get exposed to some of it...that’s not an issue most of the time on IG. It’s there for a chill experience.
The big difference between Twitter and FB/IG is that it's explicitly oriented towards following people you don't know, not your friends. There's great, meaningful stuff on there (amongst all the meaningless fluff, but that's just Sturgeon's Law). The political stuff is toxic but if you dodge that and follow interesting people it gets a lot better.
The 280-char limit is IMO a bug these days, not a feature, and people route around it somewhat awkwardly with threads.
> I am a fan of Twitter. It is a much better social network then Facebook and Instagram.
I mean. You have the right to your opinion. But... really?
I understand that the sentiment is “Twitter is mostly what you make it“, but it was so incredibly toxic that I was stunned to the point where I almost became radicalised against the ideology that was pervasive at the time.
Even now I am sensitive to it.
Facebook in comparison never had strangers piling on me telling me I am worthless and invalid based on my race and gender.
Not that any of them are fantastical beacons of acerbic communication which puts such a high emphasis on civility as hacker news does. But to call twitter the best of this breed is, I feel, speaking from a far different experience than my own.
> Facebook in comparison never had strangers piling on me telling me I am worthless and invalid based on my race and gender.
Uhm.. That seems to be just as prevalent on FB (or anywhere online) as far as I can tell.
As for Twitter vs. Facebook: I feel that Twitter is far less cluttered for me and "follow" (and unfollow) is a much better way to select my content then be"friend" (and then unfriend). On Twitter I just pick interesting accounts to read and now my feed is full of stuff that is largely interesting. On Facebook it's .. complicated (and I reduced my friend list to 2, because it wasn't worth the effort).
Instagram? That's just shallow/vain in every way and doesn't even compete IMHO - it's an entirely different beast (that I don't understand).
Do you have a sense of how much of this was due to how Twitter is designed as a medium vs due to the culture formed by the people who tend to use it?
I know both of these factors feed each other, but I ask because I’ve noticed several Twitter design decisions that seem to lead to angry mobs. Short tweets mean low-context, sorting feeds by engagement encourages piling on, and the rapid feedback loop is perfect for amplifying righteous anger.
It's probably both, but for me, I follow a pretty non-political group of people, but Twitter will absolutely insist on presenting me inflammatory and polarizing tweets just because someone I follow or someone popular engaged with that tweet. Facebook does not do this, at least for me it doesn't.
I'm more stunned at Twitter's UX. For example if a tweet of mine has x likes, I click on the link to see who liked it, and am often just greeted with a blank modal. I believe it has something to do with the likes being private, or likes from people who don't follow me? I actually don't know, and feel they could use this blank modal as a chance to tell me.
Their icon buttons on desktop have no tooltip. For some of them you can get a sense of what they do by hovering over them and looking at the url in the corner of the browser. But others that just have click handlers on them give you no clue at all. I could go on and on about how perplexing I find Twitter's UX to be, these are just two examples.
> I am a fan of Twitter. It is a much better social network then Facebook and Instagram
I used to follow a lot of blogs of people on Google Reader, when it was shut down everyone just chanted "What's the big deal? Just use Twitter it does the same thing".
Now to follow the same creators I did before I have to experience everything with a side order of hot takes about American politics (I'm not American).
Twitter is one of the worst things to happen to the internet, I can't have any sort of healthy interaction with the internet and actually keep up with what I'm interested in anymore. It sickens me that every time a platform or service shuts down the answer from Twitter addicts is always "just use Twitter".
> Just a week ago, Twitter lost over 20% market cap because they noticed that they were showing people device-personalized ads even though they were opted out of that. After they fixed it, revenues went down.
I'd like to clarify what actually happened here (throwaway for paranoia).
The actual bug was that the account signups flow was accidentally modified to automatically opt out new signups[1]. When Legal reviewed the ads targeting and discovered the opt out setting wasn't being respected, they made ads start doing so. Two this later they discovered the sign-ups bug.
[1] I know, a bunch of people would rather it was always like that.
I honestly wouldn't read anything in this bug in terms of development quality. Catalyst is buggy and poorly supported. Developing for Apple products is often-time an exercise in workarounds, hacks, and undefined behavior fun. These issues only get worse when you get to be the size of Twitter and the app architectures starts to get more complicated to support a large team. The iOS application patterns are not super evolved or friendly to scale.
Given Twitter’s size, you’d think they would be able to create good apps.
Instead, they killed off or bought and killed most of the alternative apps which provided significantly better experience with a fraction of resources.
I think it all depends on their mindset on developing the platform and the mindset of new hires - are they being incentivized to "shake X industry" or brainstorm features every other day, or do they look at what they have and figure out how to make it even better. Is Twitter made to make it easy for users to do some thing, or is it being made to do what some other company did, but "for X"
I find most SV companies, like game design, falls apart when it is used as a vehicle to test out a philosophy instead of implementing an idea users actually want.
According to levels.fyi, the highest paying job at eBay is about the same as a slightly-above-mid-level engineer at FAANG, so yeah, you're probably right.
There's also some weird, non-determinism in the beginning where the 'c' after a long chain of failed 'b's overwrites the previous 'a', which he is later unable to replicate. Not to mention preventing arrow-key navigation of the password field is the most annoying UI decision they could make. What an absolute mess.
So I am able to replicate this reliably but at first I misunderstood what was happening.
What happens is this:
1. Type some stuff in
2. Turn on “reveal password”
3. Type one of the “banned” characters
4. Switch “reveal password” back off
5. Type another character
6. The whole existing password is deleted and replaced by the letter you just typed.
So the reason I end up with two Cs in a row at the beginning is because I actually typed in another. After I type in the C that causes the field to reset, I erroneously believed that it had deleted both the B and C, leaving only the A, so I say “C” again out loud and type another C, believing the text entered is now “ac”.
In reality, it had cleared the field, replacing both the A and B with a C. So when I typed in that second C there were then two in a row.
The reason I didn’t understand what was going on was because I was just counting dots and assuming which letters were deleted at first.
Oddly, when this field-clearing behavior happens, it also hides the “reveal password” button until you add a second character to the field. That should be visible with only one character but in this particular situation it hides instead.
Aha! I missed that. How long did it take you to work out the bug here? How did you stumble on to this originally? Was it an organic error, or do you do this kind of thing for a living?
I was typing in a password (which I have since changed, lol) that contained one of these letters and it kept rejecting it. Assuming I was typing it wrong (I’m prone to that), I very deliberately typed in every letter one at a time and noticed that no dot appeared when I got to that character. I was gonna tweet about that one letter but then I started wondering how many other letters would do the same. So I tried them all.
My day job used to be internal technical support at a software company, and now it’s engineering.
Before that I was at 9to5Mac reviewing apps, and even before that I was a pretty prolific beta tester for stuff like iOS jailbreak tweaks, so I’ve got a LOT of troubleshooting and debugging experience that comes in handy here haha. I live to break things.
You can sometimes "inspect element" and then paste into the markup representation of the page ... but often then scripts will break on the sort of sites that prevent pasting.
Reminds of when I was not able to log into Skype, because when I typed in my username, which was something like firstname...familyname, it kept replacing the three dots with an elipsis automatically, without me noticing.
What could possibly be the reason for this? I mean...there's got to be some kind of validation that's going wrong here but I don't see any validation scheme that would include standard alphabetical characters.
> Root cause is Catalina regression that fails keyboard inputs when a UIKeyCommand is registered for the same key. UIResponder chain regression most likely.
That doesn't seem like the root cause to me. It sounds like they have events set up for the UI prior to login which doesn't really make sense. The regression in Catalina may have exposed it but the regression is not the cause of this, as far as I can tell.
While you may not believe they made the best design choice, there's nothing functionally wrong about the choice they made. According to the docs, their method should work just the same. So I don't think it's fair to say that's the cause of the bug.
>nothing functionally wrong about the choice they made
Sure there is... The app would capture those as shortcuts instead of as user input. You know...exactly what's happening here. If this interferes with writing tweets also then that should be easy to figure out.
As I said above, I'm not saying it's not because of the regression but it's more likely to me that Twitter messed it up. Otherwise, why is no other app experiencing this issue if it was done "according to the docs".
> Otherwise, why is no other app experiencing this issue if it was done "according to the docs".
Well, like you say, it's not common to use shortcut keys without modifiers right? So it doesn't seem that surprising to me that nobody noticed the regression.
The application is brand new, and is built on a brand new macOS API. Nobody broke anything. The point is whether the bug is due to Twitter’s usage of the APIs, or the APIs themselves.
Ugh, yucky. Surprising lack of quality control on Apple's part - surely they should have tested typing into a text field at some point during the testing; don't most programs register a shortcut key or two?
But there are also some UI fails on Twitter's side here too - like clearing the password field after clicking Show Password and typing more, and preventing arrow key navigation within the field.
"For those that are going to revert back to Mojave as a solution, a number of people are noting that this does not fix the issue (it didn't for me), indicating there is perhaps some sort of firmware update taking place with Catalina that will not be rolled back when you install Mojave."
>don't most programs register a shortcut key or two?
Yes which is exactly why I think it's likely he's being disingenuous about the root cause of this. Shortcut keys typically need modifier keys and, even in the instance where they don't, there has to be some check from the app to determine whether an input field or the main window has focus.
To be clear, I'm not saying it's 100% certain that this isn't an Apple issue but, based on Twitter's track record with their apps, this likely is on Twitter.
It's entirely possible where in Catalyst if a UIKeyCommand is registered that uses a non-modified letter, it overrides text input. Having said that, I can't imagine why Twitter for Mac would have solo-letter key commands registered during the login flow, and if they did then I can't imagine why they wouldn't also interfere during tweet composing. Or rather, I can imagine, but the scenarios I'm constructing would imply extremely contorted view controller hierarchies that make no sense.
Can’t imagine why this would cause it to delete all input in the field when one of those characters is typed in using “reveal password” mode though. If you enable reveal, type in a string containing b, and then switch off reveal mode and add another character, it deletes the whole password. If it was just dropping inputs that would be one thing but this seems to be aggressively refusing to allow them in the field.
It sounds like its specific to secure text fields, which I assume is why it doesn't interfere with tweeting. But thats a fair question about setting those up before login. My guess is those registrations aren't happening at the login VC level, or maybe in some global container VC. But I don't think that matters too much, because there are certainly use cases where you might want to show a secure text field even when logged in and you want shortcuts registered.
I am and that's why this is so surprising to me. No other app is reporting this type of behavior and Twitter is known for shitty Mac apps. I highly doubt that the regression could cause this unless it was done incorrectly to begin with. Otherwise, UIKit/AppKit would be doing the same for every other app. We may hear about that in the next day or two but, until that happens, I have a feeling this is all on Twitter.
That doesn't mean much if they implemented it in the exact same way they did for the full app. Unless we see this happening in other apps, I'll default to the implementation being the problem rather than the API itself.
Every time I see things like this, it makes me feel glad I chose Win32 (actually Win16, not that it matters all that much...) as my first platform to start learning to write GUI apps for, and didn't switch over to the Apple side of things after that; the "move fast and break things" mentality was already somewhat noticeable in Apple back then, whereas Microsoft (the "old MS", not the "new MS") was more concerned with "protecting your investment in the platform". In my experience, core functionality bugs like this just don't occur with Windows, probably because there's far less rewrite-churn[1].
I know Windows has had various UI frameworks come and go over the years, but they're all fundamentally layers on top of the underlying stable Win32. For Mac I'd not really dug into how to write GUIs, but I know there's things like Carbon, Cocoa, etc., and those seem to be the base itself, and change significantly or are even entirely removed at the whim of Apple. Is there any stable GUI API on the Mac at all?
Note that the open source releases often get things scrubbed or ifdef'd out of them.
I mean things like TODO comments, not deliberate censorship.
I fail to see your point.
You don't notice anything wrong with the insane complexity to do almost nothing? A similar near-nothing app would be 1/10th the size on Windows, if not smaller.
I see nothing wrong with that app, since it’s purposefully convoluted. Apple’s application frameworks are written in and meant to be accessed through Objective-C so trying to write an app in C is going to require a lot of hacks and additional complexity. A simple application can be written in under a dozen lines of Swift or Objective-C code. (For an idea of how insane what you linked is, consider trying to use a C++ library from C: you’d need to hand-mangle the function names, and if you want to pass in parameters that were C++ objects you’d need to lay out their vtables and fields correctly, and…you get my point. It’s not anywhere near representative of what the normal development experience would be like.)
It's not "purposefully convoluted", and I'm not talking about "normal development experience", but rather the larger amount of non-essential complexity it shows; and as anyone should be aware, complexity is always an opportunity for bugs, regardless of whether it's something you would see or write or even be aware of in the "normal development experience". Combine that with Apple's attitude of constant churn and it's not hard to see how bugs can multiply easily.
Using a C++ library from C is not hard because objects map to structures in a straightforward manner (and in fact early C++ compilers were C++ to C translators), and you would need to use structures anyway, unless the library was obtusely designed with nonessential complexity. That's how the Win32 API works too (with the exception of the bloated COM stuff.)
Cocoa and the Foundation framework have been around longer than MacOS / OS X, i. e. more than 20 years. How many different concepts has Microsoft called ".NET" since then?
Regardless of the flavor of the month, the difference is Microsoft is still supporting most of their old frameworks. You can still go write an MFC app today, for example. (VB6 is a major unfortunate exception, I will admit).
Maybe some over-zealous shortcut handler that's preventing keystroke events from bubbling up? It's interesting that shift-b produces "B" but just b with caps-lock doesn't produce any characters.
but why would it work with a revealed password then? and why would typing one of those characters, hiding the password and then typing anything else clear the entire password? I'd really like to know what the heck is happening here
It reminds me of the days when I used WinRAR for a lot of things. The password field would only allow entry of ASCII characters but the program will accept any character by copy and paste, leading to a world of pain where certain archives can only be opened on a specific OS with one particular version of WinRAR because the internal parsing seems to depend on both when it comes to exotic characters. Good times and I think everybody have learned to avoid this sort of hiccups by now.
https://twitter.com/nolanobrien/status/1189623346033381376?s...