Dang. This is the future I wanted to see for mobile devices.
Really bummed where we're at right now in the mobile ecosystem. It's crazy to me that even now iOS/Android still obscure your directories from you, and navigating them is treated as something to hide from the users despite being an integral part of the operating system.
Really wild how these companies leaned into disrespecting their users. A lot of people might say "It's to help the users and remove touch points they don't care about", but the truth is that in making these decisions they have trained users to ignore these mental models and completely hidden or removed the opportunity to normalize these aspects of the operating systems and the device.
My android devices have all come with basic file system browsers, but even better, I can choose from a plethora of full featured apps for full file system access. This is one of the issues that has kept me away from iphone, though.
Modern Android (11 onwards) no longer allows full filesystem access for file manager apps. Apps can opt into exposing some of their files for management by implementing a horrible Java API designed for cloud storage services. Some vendors' forks probably disable this change but that's not something that can be relied on.
Meanwhile iOS also has sandboxing but at least you can still meaningfully use normal goddamn filesystem APIs and there's now a system file manager.
Minecraft for Android has two storage locations in-game: Application (default, not accessible to users) and External (accessible to users by connecting their phone to a computer).
I cannot backup my worlds without either using a third-party tool, or paying Microsoft $4 for the privilege of uploading my world to their server (Realm), changing my storage location to External and then downloading it back. Ridiculous.
I seem to recall very Android phone I've had came with an app called "Files". If not named "Files", they always had some kind of baked-in file browser. The last few phones have been Pixels, I don't remember exactly the specifics of the one on the Motorola devices I had prior, but on the Pixel devices its this app:
No ads. Sure, it tries to group things by categories and searching at the very top, but its still easy to see "Storage devices" which lets me navigate the actual folder structure. My Pixels have come with USB-C male to USB-A female adapters which allow me to plug in things like flash drives, memory card readers, cameras, external hard drives, even other Android devices and access the files through the Files app.
Yep, this is the main reason I use Android. It also means I can modify things on the OS level to e.g. fabricate false GPS data and supply it to apps that insist on GPS access to be functional.
On iPhones and stock Android phones, apps "know" when you deny permissions, and can even know when you use the mock GPS feature.
Out of curiosity, are you referring to rooted Android, or some variant that provides the user some protection from apps "knowing"? GrapheneOS seems to provide some of this (Storage Scopes, network permission, etc.) but not all.
Custom mod of LineageOS, not rooted stock Android, although you might be able to accomplish it with rooted stock and some frameworks like Xposed, they can be detected by some apps, and some apps actually do detect those frameworks and ban users. WeChat is particularly notorious for this kind of detection. If you get banned you not only are disconnected from communicating with your family and friends, you can't even book train and flight tickets, you can't buy food at many places that only take mobile payments and don't take cash, just for a stupid framework. So it has to be done on the OS level.
I was thinking about getting an xperia, I asked about rooting/custom rom and apparetly unlocking the bootloader will delete keys that the camera needs to run. Seeing as it's designed as a cameraphone first and foremost, that's a bit of a major feature to lose.
Why, as a user, is that story (even without losing the camera) would any of that sound better than just having the app store require that apps follow some approval guidelines about what they are allowed to do with privileges and what they are allowed to do when denied privileges?
Unfortunately, you can't really have sideloading and an app review process. If sideloading is trivial/normalized for users then it'll be trivial/normalized for bad actors wanting to exploit those users.
According to the App Store rules for iOS, An app can’t withhold functionality when you deny it GPS access unless it’s required for the app’s functionality.
Why would I use an app that requires unnecessary permissions?
The SSH client Prompt uses GPS permissions as a round-about workaround to ensure your SSH connection is not reaped and closed by iOS while in the background.
To stay in touch with friends and family who only use a certain app for communication and social event planning. I don't have the energy to convince them and an entire country to move off said app.
To get permits to certain wilderness areas that require GPS permissions to even apply.
To get on restaurant waitlists before arriving. (Some restaurants insist on your phone-reported GPS location being within a certain radius to get on the list. Welcome to Silicon Valley in 2023)
So exactly why do you need access to the file system? You have access to a location where you can share files between apps. Any application that supports access to file picker gives you access to any of your installed document providers like iCloud, Google Drive, Box, an attached USB drive, a network drive and the local device.
And direct access via the Files app.
The one thing missing is that admittedly there is no way to add your own music to the music library on the phone itself.
I pretty much never use it on my Mac, all the apps I use show the filesystem in a sidebar and have a fuzzy search feature. Even Apple apps (like Xcode) have that so clearly they're aware of how bad the file picker is.
The difficulty I had just copying things from one phone to another recently was ridiculous. You’re right, this should be a basic function. I have files I’ve copied from computer to computer that were originally from the late 80s.
I have this problem also with Windows, it's one of the two purposes for which I made https://dro.pm. If it's too large for email or there's no email client installed, I wouldn't know for the life of me how to copy things on Windows without spending >5 minutes setting up shared folders, firewalling, and then undoing the changes when done. Uploading to a website and typing over a very short link is done in ten seconds
Not as convenient if you want to copy a large number of files and don't want to zip them all, but in most cases I just want to transfer one file or link or command (text) and this is still my go-to solution more than a decade after I made it, works on all OSes with either a command line or a visual browser. Beat that, airdrop / nearby share / Huawei share.
Cool service, probably shouldn't be used for anything you don't want to be public. The URLs seem to change incrementally, and https://dro.pm/g is self referential for that reason, probably. You can go letter by letter to look at various thicks people share.
Yes, I'm hoping that's obvious as it'll be inherent to such easy-to-use links, though the site has such low traffic and I know from access logs that nobody is guessing these links in an automated fashion, that a few seconds of exposure are almost negligible, depending on what it is you're posting.
If the links were unguessable (or had a password option, as has been proposed many times), you'd also lose much of the convenience. But unguessable links is an option if you use the second tab, effectively the same as a password.
Especially if you use "delete after opening", the custom link can be extremely short-lived (no one will guess that in time) and you got an indicator of compromise as well. (I personally trust this enough for ID cards and passwords, but of course that's with the benefit of being the owner and knowing all of the above plus that there is no secret logging.)
Fwiw, I also like the "social network" effect so far. I've discovered interesting things on dro.pm by looking at what people use my service for. Currently, /q has an interesting youtube video that I wanted to continue watching on my phone for example :)
I still don't understand how move and copy were allowed to be fundamentally broken for so many years. File management is a very, very basic OS task. I should never, ever need to Jailbreak a device to make backups of files.
You can choose whatever words you want for it, but there's a difference between the targeted polished UX of "there's an app for that", and a morass of software that may or may not serve your needs that kind of sort of works together if you hold it right.
Is that true? mv isn't an application. To me, the GNU toolchain is more like an application, but the individual parts like GNU ld, GAS, ar, GCC, etc. are just programs, not applications themselves. And even then, application feels like it implies GUI, maybe?
I'd say something like GNUCash is an application, but hledger feels more like a program. Brewtarget is an application, but bc isn't.
For the same reasons Android just actively broke it. Sandboxing is a matter of user security, if you let apps violate sandboxing then you can't have a meaningful permissions system.
You can allow certain common kinds of sandboxing crossover - apps that want to touch the camera roll, apps that want to touch wireless/location data (which are one and the same), ones that want to use the contacts, etc. But at the end of the day if you don't fall into one of these buckets you're going to have an awkward experience going into the OS and approving moving stuff into/out of the sandbox.
It'd be nice if you could allow complete crossover for certain apps - so, VLC can touch all of the podcast app files, or whatever. But I can understand why they don't even want to cross that rubicon.
You've gone and conflated user and application permissions. I was talking about owner permissions. The device owner should always have the ability to backup any file on the system. Only Apple controls the entire system. They can and should give this control to the owner.
modern users (for example, kids) are able to get by with their files in the default folders of the applications and no concept of a directory tree. And with performant search they don't worry about it, they just search for the file. I don't know that that's due to disrespect but it's a thing.
To call it an abomination is annoyingly dramatic and "they don't know any better" is frankly elitist and I am really tired of these attitudes from HN.
There are people in the world who grew up with and are used to platforms that have filing-cabinet abstractions (files and folders) and let them run their own code and customisations. Similarly, there are people in the world who grew up with devices that don't expose files and folders as readily and instead present things through apps and search.
How many of the latter category could even accurately define what a "file" is? Do they even need to? Probably not, because they get just fine with a different set of abstractions and tools. Just because that isn't to your taste doesn't mean that they don't work fine for other people — there's a world of less technically-minded people out there who get by just fine with an iPad at home for the things they want to do. To them, computing has _never_ been more accessible than it is today and not having to worry about things like folders and files is actually a superpower.
The whole point of the filing-cabinet abstraction is to reify my access to something in some app. It's not a purely technical concept, it represents an affordance that users do care about. We went through this before; programs used to be written without any generalized notion of a file, and you'd have to work with things like "unit" records represented on punched cards, and "datasets" meaning collections of unit records. There's no reason to go back to that kind of chaos.
It's not a "filing-cabinet abstraction". "Files" and "folders" might as well be opaque and arbitrary words to most people. In a computing context, they have no or almost-no association with files and folders IRL. They don't even correspond to those concepts. In real life a "file" is literally a folder full of documents.
>How many of the latter category could even accurately define what a "file" is? Do they even need to?
All of them, because they don't just use phones. The same people have to write things for school and upload those .doc files onto a school website to have them automatically checked for plagiarism and marked, and if they ever work in an office they use files every day.
And trust me, people are NOT being empowered by these devices. They actually find them super frustrating every single day. I know non-technical people. They hate modern devices. The old systems did the same thing every day, day in and day out. They didn't change every five minutes. They didn't hide all their state.
These are non-technical people, who couldn't even connect their phone to WiFi without help from someone. Yet they had no problem with plugging their computers into the wall 10 years ago.
They can't work out how to print things today, what with the awful array of proprietary printing apps made by the incompetent software developers working for printer companies, yet they had no problem printing things out when all you needed to do was plug the printer into the back of your computer once when you bought it and forevermore press "Ctrl-P" and "Enter".
They can't easily use these newfangled tablets and phones to write documents or spreadsheets. Yet they had no problem using desktop office software 10 years ago. They had no problem with the concepts of files and folders. They do have a problem when they just can't work out how they're meant to get things from one device to another, because it's all opaque. They have to resort to things like sharing a document to themselves via email, then opening the document up in the email app and opening it in another app. That was the simplest solution that was obvious to my mother, and to be quite fair to her, that probably WAS the simplest way to move a document from one app to another. Having a single, universal, underlying "filesystem" metaphor that you learnt once and which applied to the whole system was great for non-technical people who just want to get on with using their computers.
So no don't start with this "has never been more accessible" or "not having to worry about things like folders and files is actually a superpower" crap. It's the opposite.
Honestly, for a lot of my personal files and documents I really don't care about the folder structure. Don't get me wrong, I do, but only because on most systems its the only reliable way to actually organize the files. But a directory structure isn't always the optimal way to organize things for a lot of my access patterns.
Like pictures. Maybe I'm wanting to browse by date. Maybe I'm wanting to browse by events and albums. Maybe I'm wanting to browse by photos with these two people in them. Maybe I want to browse by pictures with boats in them. So should I arrange my photos in directories by date, or by album, or by some kind of classification and then date (family/2023/bobs_birthday)? It sucks! Its a terrible process. And by count of files, photos and videos are the vast majority of the files I care to access!
And then its a similar thing with stuff like music, or movies, or all kinds of things. Folder structures are very rigid, things like links are all kinds of fragile and annoying to work with often.
For the most part, the vast majority of my files I'd prefer to have them all in some kind of database with tags and be able to query them based on the context of what I'm doing. Open a photo editor app? Ok, show me photos, probably recent ones first until I put some other search in it. Not just a file browser sitting in a home directory in a folder full of other folders with all kinds of other files that aren't even pictures.
IMO, for personal files, folders are bad. I don't mind a lot of these interfaces that try and hide the underlying file structure, so long as when I need to I can muck about in there as well, if it still exists.
I mostly agree with this. I don't think it's "files and folders" that are bad, but a file browser isn't the best way of browsing photos. A file browser is great for documents and you need to have it because you need that "lowest common denominator" browser as a backup or when things go wrong. But for things like photos what most people probably want is:
1. If you plug your SD card/camera into your computer, sync all the photos onto your computer and back them up remotely before wiping them from the SD card.
2. If you take photos with your phone, they should automatically be synced to your computer and to a remote backup.
3. When you want to browse your photos, you just go to a 'photo browser' and browse by tags, date, etc.
But one of the problems with this stuff is that those proprietary kinds of databases have a habit of eating your data and being difficult to troubleshoot. I remember with iTunes it would swallow up music you put into it and hide it away in some internal database somewhere. I would have much preferred if it had its own browser but that the underlying data were stored in ~/Music/Artist/Album/1.Track.mp3, so that if anything went wrong it would be easy to pull all that data out. I think the way it actually did it was some big store of files with names like ab4f8241def.mp3 and a database mapping artist/album/etc. to those files.
I have all my (legitimately-acquired) TV shows and movies stored across a variety of computers, and then xbmc has media sources that are things like nfs://10.0.0.1/tvshows and nfs://10.0.0.2/tvshows. Of course the only way I ever consume any of that data is through... a big aggregating media browser that automatically downloads metadata about my media from TMDB etc. But it's all still stored as /media/tvshows/Silicon\ Valley/Season\ 1/S01E01.mkv, yknow? It would be awful if I couldn't access it like that.
> But one of the problems with this stuff is that those proprietary kinds of databases have a habit of eating your data and being difficult to troubleshoot
Fully agreed there. It would be nice to have some real widely used standards in this space. I'd love to just have a single API available to query any of my datasets wherever they may be. When I open a picture picker I should see files from my NAS, from Google Photos, from my local device, from any remote device I can access at the moment, from a NextCloud instance, etc.
Why is visually navigating a directory tree superior to telling the computer "show me photos of my family" and having it do a better job than I manually good?
Throw everything in your home/office into a huge drawer and try finding that one tax receipt from 2012. Sounds fun?
Now imagine tree-organized file cabinets. You look up the labels: Personal finance, tax stuff, 2012. Found it.
If done right, its basically O(n) x O(log n). With N being very big and your seek time being very bad.
Its not because people lost the ability/knowledge/patience/etc to do this that makes it the same as the mess we have now.
Sadly, I think its more likely we'll increasingly rely on AI helpers to do the mental work for us (like already happens on Google Photos) instead of putting in the effort with something like a file structure again. I think its gone.
iOS has come a long way, though. There's a local file system I can access and move files around with. I can run Syncthing on it (with Möbius). I can even run Linux on it (iSH), including a package manager. I can use qpdf to decrypt a PDF I downloaded with Safari, and I can even run Python and compile C code, on the device, with acceptable performance.
It's not the same as a real computer (no daemons or cron jobs), but it can do a lot of the things I want to do when I'm on the go. The only thing I'm missing is running multiple apps side by side, but I guess that's hard on such a tiny screen.
You can't do any of these things without doxxing yourself to Apple, as you can't install apps without an Apple ID, and you can't get an Apple ID without a phone number.
Signing up for an account with your personal details is not "doxxing yourself". Doxxing is when someone else publicly reveals your personal information on the internet.
A counter-datapoint of one, from me: I don't miss the file system browser at all. I have been a little bewildered by the iOS "Files" app, and the addition to the sharing widget of "Save to Files" - I never use it. I'm not saying it doesn't have a purpose and that no one would ever use it, but I don't miss it and don't have a need or want for it.
Sometimes I want to collect a couple of files and documents from emails, and make sure they are both available and together in a place. That's about the only time I use it, but I find it pretty handy in those occasions. It's more of a temporary place though, rather than something permanent.
But we have things like the Raspberry Pi, which are very low cost and completely open to tinkering. Add a keyboard and clip-on display then you have a tiny but capable development machine, in a smartphone like form-factor.
Really bummed where we're at right now in the mobile ecosystem. It's crazy to me that even now iOS/Android still obscure your directories from you, and navigating them is treated as something to hide from the users despite being an integral part of the operating system.
Really wild how these companies leaned into disrespecting their users. A lot of people might say "It's to help the users and remove touch points they don't care about", but the truth is that in making these decisions they have trained users to ignore these mental models and completely hidden or removed the opportunity to normalize these aspects of the operating systems and the device.
Really wish it wasn't this way.