okanat

... which is "ripping off OSX". But you don't want to hear that, do you?

UX being copied by OSes from each other is a good thing.

I actually like Ribbon and I disagree. It made functions easier to find, for people who didn't have a muscle memory of those functions. It actually taught them how to use the program and exposed conditional menus like picture options better than the toolbars that appear and disappear.

Of course people who got used to traversing 3-level menus got annoyed but they still have all the same keyboard shortcuts. However, spending enough time, you also get used to Ribbon and it usually requires fewer clicks not more.

I still find a tastefully done Ribbon is better than a menu and a toolbar combination. The problem is, not everybody has put the same effort that Microsoft did which created monstrosities like Solidworks Ribbon.

I think the sweet spot is WPF or actaully its FOSS version: AvaloniaUI. It is high level enough to be cross-platform and not stupidly verbose, however it is rather light pn resources. However, Win32 still wins in absolute minimalism.

Although I started with 98, I think Vista / 7 was the peak. It was user friendly enough for newbies but didn't sacrifice advanced controls. Just organized them better to bring focus to the more used parts.

The syntax and meaningful spaces are annoying. The dynamic typing is abused to hell in everything including the standard library. It has wacky paradigms like "exceptions instead of conditionals" or having "else" clause for loops.

It lacks decent packaging and by design cannot handle multiple versions of dependencies or perform metadata-only/declarative resolution. To resolve the dependencies of a Python package you need to install it. Many Python packages resolve their dependencies imperatively depending on the current system state. When you need binary dependencies like numpy it gets even worse since libpython doesn't have a stable ABI.

It is one of the most underdesigned languages but it holds a dear place due to being easier to translate from pseudocode to actual implementation. This resulted many academics to pick it up and write advanced math libraries in it.

> Or just in general how these people seem to completely forget that desktop usage is largely just an afterthought for Linux and the main usages are phones, servers, small devices and supercomputers.

I think you're right in that most paid developers use/develop for Linux ecosystems do for server or embedded, where they use Linux out of "laziness" since it provides a basics. Then the engineering team implements everything else GUI, webservers etc on top of it. At work I'm a firmware engineer but I can also write quite complex Linux stuff and I occasionally contribute to the Yocto recipes. You really see that Linux ecosystem is really for people who prefer to dig stuff. Dig decade-old email threads, read C source the whole time. Fix and patch weird bugs and pile workaround on top of workaround by forking things. Linux ecosystem really isn't designed to scale to be deployed into wild-west PC universe and still lives in a "here is the source; good luck!" world.

> Even ignoring the future though. It's just the understanding of the “now”. Things like initially libinput having no way to disable mouse acceleration and the developer at first resisting it until receiving a mountain of complaints and finally relenting because “Everyone would want mouse acceleration always on, right?”.

I mean it can be argued where to split and keep libinput as the lower-level library for the embedded engineers (TBH I don't know if embedded engineers really care). However, what I don't understand is: if they are going to split things in a more granular way, why not provide a higher level library for DEs where functions for acceleration etc. are provided as extras.

At the same time, I have a hunch that GNOME could be the reason (AFAIK libinput dev is employed at RedHat with other GNOME devs) and GNOME likes things "lean" i.e. featureless. One of the reasons Wayland has completely surrendered window decoration and we have to use workarounds and hacks like XDG Decoration is GNOME's insistence on offloading everything to clients and vetoing every single compromise (including XDG Decoration, but Valve pushed it AFAIR).

Again I think designing Wayland as a "allocating rectangles on CPU or GPU buffers" protocol/library is okay. But why don't we have a unified higher level "Wayhill" where ex-GNOME DEs define things like what a decorated window is [^], what happens to the crashed application, unified resizing, actual display management and unified configuration so it can be shared with the login manager. Now KDE implements one thing, wl-roots implements another (wl-roots is probably closer to my "Wayhill" idea but it kind of caters to more terminal-driven stuff, I do prefer pure GUI-driven desktops) and GNOME just opposes everything.

I see this kind of "provide the lowest level abstraction, nothing else" as a paradigm in every single component in Linux. It is really tiring both as a desktop user and as a paid developer. Since it is a socio-technic concept, rather than a purely technical, I don't really see that we'll solve it with "more FOSS". The overall mindset needs to change or we need to bootstrap a new ecosystem (which is an uphill battle).

[^]: Yes Firefox and Chrome would look wacky (basically dual decorated) since they use some lower level GNOME/GTK stuff (file dialogs, cairo rendering etc, GTK themes/libadwaita), however this kind of push could have helped browsers to be more agnostic as well.

Would you consider writing Docker/moby in a different language, if you were to create it scratch today?

Oh good to know and I stand corrected. I guess if you can opt-in with a syntax (like dyn does in Rust), this kind of in-the-middle generics could be a nice addition to systems programming languages. However, we need a GC-less solution of course.

Not really. The modern NT enforces its specs unless you opt-in or Microsoft opts your Hyrum's Law monstrosity in.

If you look at the compatibility tool you'll see multiple checks not only executable name match. It can check version headers and other signatures.

Sorry but scientific method has been the default for "Truth" for many academic subjects. It is reliable and idempotent. If all of the human information and most of the knowledge workers are removed from the Earth tomorrow, many will arrive in the same conclusion in most of the natural sciences and maths. The language and representation will be different but prime numbers will still be prime numbers and the matter will still be made of one neutral and two "interactive" elements.

Social sciences like politics, history and economy are harder to replicate with the scientific method, however repeated observations, statiscial information as well as qualitative analysis with evidence-based research can still provide replicable analysis.

> If anyone has a better idea how to determine what is true and what isn't - i'd certainly be very interested

This kind of excuse is often cited to shut down any discussion, especially in history, politics and economics. I find it deeply anti-intellectual and dogmatic. As long as evidence and counter-evidence is followed, it should be possible to discuss alternative interpretations and solutions.