It absolutely can, I have several portable apps with self updating ability built in, when I use them it prompts me if I want to update or not, I personally appreciate that in certain cases.

I do dislike when they throw config files all over the place, so cleanup becomes very messy if I need to remove something.

Again tho, natively on windows there isn’t a great way to do that anyway, the windows store sucks and not many will use the package managers via cmd either.

Noob, I snort pure Colombian and run LFS directly in my brain using the power of cryptic dreams and messages handed down to me directly by Lord RMS

Actually, to be clear, I don’t think FOSS purity is bad. I just mean that denigrating what others are doing because they’re using something non-free while they’re making steps in the right direction is dumb and counterproductive.

To my mind, FOSS is the only way forward for a healthy, functioning society, and the fact that so much of our digital landscape is being gradually replaced with it is to me evidence of that. I think the end goal should always be pure FOSS, but that doesn’t (necessarily) mean immediately jumping to all FOSS; it just means taking steps to cut out proprietary software wherever you reasonably can.

Oof GTK is probably one of the worst dependencies you can try and port to Windows.

What I’ve done in the past is use something like Onno Setup which can call a script during install.

Or, and this is new to me, use the Official tools to build a package for windows on whatever Linux distro you are on. From what I’m reading, it should package GTK with it.

Edit: also, I need gtk because people around me who uses windows aren’t going to use CLI program at all.

If that’s the reason - maybe you can use TUI instead? In Windows, it’d open a CMD window which your users will be able to use. Not as pretty as actual GUI, but easier for Windows users to use than a CLI.

Another option is to use one of the numerous Rust-native GUI libraries (like iced or Druid, to name a few). None of them are as big as GTK/QT - but they are easier to get running on Windows.

My antivirus is extra paranoid: it scans new files as soon as they’re unzipped or as soon as I try to run them for the first time.

A lot of exe files are secretly zip files. zip files can contain arbitrary data at the end of the file. exe files can have arbitrary data at the start of the file. It’s a match made at Microsoft.

This is true for the code part, but executables can also contain data does compress well and maybe not be compressed inside the EXE (e.g. - to avoid the need to decompress it on every run)

Time to install flatpaks. It’s the future of userspace programs on Linux anyway, you’ll get newest versions there the quickest.

That is part of the deal with Debian. You get stable software… but you only get stable software. If you want bleeding edge software, you’ll have to install it manually to /usr/local, build from source and hope that you have the dependencies, or containerize it with Distrobox.

If you go to a butcher, don’t complain about the lack of vegan options.

That’s because the official instructions say to install it through snap. Which is just snap install blender. You may have problems with flatpacks (I don’t to be fair) but that might be outside of the scope of this comment. Or just go to the app store and download it haha

And if you really want to install the deb package there are instructions to add the PPA.

You can also just download any binary file you find online and run it. Or use any install.sh script you happen to find anywhere.

Package managers are simply a convenient offer to manage packages with their dynamically linked libraries and keep them up to date (important for security). But it’s still just an offer.

You can still do that on Linux. Just download it and run. You can even compile it from source if that’s your thing.

However, because there is a much greater variety of Linux distros and dependencies compared to Windows or MacOS versions, it’s better to either have a Flatpak, AppImage, or package from your distro’s repo. That way you’re ensured that it will work without too much fiddling around.

Do you know about AppImages? Seems like those meet the need you’re complaining about.

You still have to set the executable flag for them, but you can do that through the graphic user interface. No need to open a terminal.

The difference between a package manager and an app store is that the package manager allows you to pick your own sources. You can even run your own repository if you wanted to.

I don’t really like the way software installation is centralized on Linux.

you have to interact with their ‘trusted sources’.

I would prefer if it was easier to simply run an executable file on my personal Linux machine.

I’m sorry but which is it? Do you want centralized software installation (this is literally how all of Microsoft Windows works.) Or do you want independent release software? (Those are the ‘trusted sources’ you seem to detest.)

And there are plenty of programs that run as an independent file on Linux, installers even. They just aren’t labeled .exe.

Either you want Linux and want independent control of your desktop system and environment or you want to be spoonfed everything as a Windows or OSX user. So which is it??

Software Installation is all but centralized on Linux. Sure, there is your store or package manager, but both Apple and Windows do have that, too. But you can always add any source you want to that store (flatpak is great), find an AppImage, some doubious install script, find your own packages and manually install them (like .deb), use Steam or sometimes, like with Blender, download, decompress and run it.

As others have pointed out you can do this, but there are at least two major advantages to the way Linux distributions use package managers:

1. Shared libraries - on Windows most binaries will have their own code libraries rolled into them, which means that every program which uses that library has installed a copy of it on your hard drive, which is highly inefficient and wastes a lot of hard drive space, and means that when a new version of the library is released you still have to wait for each program developer to implement it in a new version of their binary. On Linux, applications installed via the package manager can share a single copy of common dependencies like code libraries, and that library can be updated separately from the applications that use it.

2. Easy updating - on Windows you would have to download new versions of each program individually and install them when a new version is released. If you don’t do this regularly in today’s internet-dependent world, you expose your system to a lot of vulnerabilities. With a Linux package manager you can simply issue the update command (e.g. sudo apt upgrade) and the package manager will download all the new versions of the applications and install them for you.

The entire point is you don’t need to wait through a slow installer, you just open discover or software center and install whatever software you need. In addition to being easier and more intuitive its also more secure (you’re less likely to receive binaries from a malicious actor)

Developers can just make a flatpak, publish it on flathub (or their own repo if they want, but ugh) and that’s it. Then you install the app through GUI package manager or by executing flatpak install <package name> command in your favorite terminal.

That’s literally what I wrote on a satirical post about moving to Windows! https://lemmy.world/comment/14612934 Except I was being sarcastic and you’re being serious.

Yea, I’d expect it to be named “propertty”