// QEMU UART registers - these addresses are for QEMU's 16550A UART
#define UART_BASE 0x10000000
#define UART_THR (*(volatile char *)(UART_BASE + 0x00)) // Transmit Holding Register
#define UART_RBR (*(volatile char *)(UART_BASE + 0x00)) // Receive Buffer Register
#define UART_LSR (*(volatile char *)(UART_BASE + 0x05)) // Line Status Register
This looks odd. Why are receive and transmit buffer the same and why would you use such a weird offset? Iirc RISC-V allows that, but my gut says I'd still align this to the word size.
Backwards compatibility aside, why bother implementing additional register address decoding? Since the host already doesn't need to read THR or write RBR they can be safely combined. Some UARTs call this a DATA register instead.
The UI in terms of space and usability looks great. Two "modern" things I don't want to miss: Good font rendering and a fast application launcher (mod -> type a few characters -> enter). What I dislike the most on modern UI, and maybe absolutely hate, are all those super slow animations. Just gimme the damn thing, I don't need those animations. (Yes I know on most plattforms I can disable them, but this often takes quite a few steps)
Agreed on the animations but that isn't my top of the list because as you observed, those can usually be disabled. The most annoying aspect of modern interfaces is total inconsistency in looks and behaviour across different applications. Even common action icons vary in style, colouring and shape from one application to another. Title bars are hijacked for whatever fanciful ideas the app designers had in mind, scrollbars and other basic widgets are rarely drawn using native desktop components, tab ordering is a dream of the distant past and so on and that's if a given app even responds to the tab key.
You would've really hated software in the early 90's -- every single thing had an aesthetic of its own. It was actually quite wonderful, and a lot of style/"personality" embedded in these design choices.
Are you thinking here of pre-multitasking desktop usage, stuff like DeluxePaint, Scream Tracker, that kind of thing?
Certainly the late 90s was the heyday of desktop consistency on Windows, in the 95/98/ME era, I think driven largely by the conventions Microsoft established in Office. And I believe Mac OS gave pretty good platform-level guidance then too, so things were generally okay with a few exceptions— stuff like media players that have always been more on the fanciful side.
It is my recollection as well. Most applications used VB, Delphi, MFC etc. that all had the "native" OS look and feel. There were some exceptions like WinAmp and others, but from what I can remember most applications were more consistent than today.
Those toolkits would usually reimplement the "native" look and feel from scratch, or nearly so. It was uncommon to rely directly on the basic OS widgets.
MFC was Microsoft, so that was definitely native, and I think a lot of stuff used native even just for performance reasons. I remember getting very frustrated around then when something would want me to install the JVM and I knew I was in for a laggy mess of an application that would have bad font rendering, strange little buttons, and its own file picker.
Don't modern versions of Windows contain at least 5 different widget frameworks? Like, Win32, Ribbon (I think engineered for Office as you said), WinForms, WPF, WinUI 1/2/3... I think Apple just has Cocoa (Carbon is long gone), AppKit, UIKit, and SwiftUI.
MFC is a thin wrapper around Win32. Delphi's VCL is a much thicker wrapper but still using (mostly) native widgets; ditto for VB6.
So, no, it was quite the opposite - it was uncommon to not rely directly on the basic OS widgets. Off the top of my head, the two toolkits that I remember that didn't do that were Borland's OWL (which quickly died out in post-Win16 era, since Delphi/VCL was strictly better), and Qt, which while not using native widgets tried to approximate that look and feel as much as possible.
Even in Java land, their first take - AWT - wrapped native widgets. It wasn't until Swing that they moved on to rendering their own, and it was widely derided as looking inconsistent with other apps as a result of that.
Early 90's, especially on Amiga -- where a proper windowing desktop env ("workbench") coexisted with the wildest custom UIs. Maybe it was the roots of the machine, heavily used as a games console and a demoscene workhorse? It seems like at that time there was so much creative design effort put into UX -- and it didn't seem to get in the way, maybe because each genre of software was kind of on the cutting edge back then, establishing what would eventually evolve into conventions. Mod trackers, image editing, disk copying, etc. Maybe it's a bit of nostalgia, but it felt really immersive to pull up a piece of software you were familiar with; each UI was so distinct and purpose-built, but it also had.. flourish? style? soul? Not so much now.
Nah, the Windows ecosystem never even got close to being consistent.
MS Office had its own UI toolkit and routinely invented new UI paradigms that weren't exposed in any Windows API, leaving people who wanted to look native scrambling to reimplement. This was particularly the case for toolbars. MS Office first invented the so-called "coolbar" and then the ribbon. Internet Explorer also rolled its own toolbar styles in ways not supported in the base Windows API e.g. toolbars with large icons and sliding sub-sections. Inventing custom toolbars was practically a sport on Windows; Netscape also did it.
At the time the most popular media players were WinAmp (totally custom and themeable to boot), RealPlayer (custom UI https://andrewnile.co.uk/blog/remembering-realplayer/), Quicktime (custom UI) and Windows Media Player (mostly but not entirely native).
Even the base utilities that came with Windows weren't consistent with each other. It wasn't uncommon in the Win 9x era to find programs still using Win3.1 style file dialogs ... a few are still buried in Windows today!
The problem got worse when you examined the artwork. The stock icon library in Windows was anemic, so dev platforms frequently had to expand the core library with their own. Delphi apps could be easily identified by the distinctive icons in their buttons (https://zarko-gajic.iz.hr/wp/wp-content/uploads/2017/02/delp...).
Restyling window decorations was also very common. Microsoft themselves did it routinely, for example their flagship Encarta encyclopedia app had totally custom widgets and window styling: https://winworldpc.com/product/encarta/1999
Windows apps of this era were much like web apps are today: they shared some common code for things like rendering menus, buttons or widgets in their settings screens, but the main UI users interacted with were almost always custom widgets that were extremely varied between apps. Win32 was nearly impossible to style compared to HTML so this represented a large investment of developer time, but a custom branded UI was believed to be worth nearly any cost. This is something fundamental to how humans work and is pointless to fight, a lesson the web platform fully embraced giving it an advantage over other UI toolkits of the era.
It was still much more consistent than things are today, though.
With respect to toolbars / coolbars specifically, one thing to remember was that those weren't kept for Office/IE use only, but rather shipped as reusable components ("common controls" etc), and so other apps could and did pick them up. Indeed, well into late 00s, the common fashion for Windows apps was to try to look like the most recent version of Office wrt menus / toolbars.
Also, I do recall that those apps which tried to look flashy with fancy custom styles etc were often perceived as unprofessional, and quite a few people (myself included) deliberately avoided them where possible - and it wasn't difficult to do, with natively styled alternatives readily available. I distinctly recall my own late-90s Windows desktop, and it was very consistent.
IIRC most of these controls only shipped in comctl32 much later than they launched, and comctl32 came with the OS which had many-year upgrade cycles. So in practice everyone invented their own because they didn't want to wait.
I'm sure people who care could make a consistent desktop, but our memories differ on how popular that was. The theming craze was a 90s thing, so even when apps didn't roll their own brand like MS themselves did, apps often let you apply custom themes to change the look.
In theory it's easier than ever to do that. You could create a browser extension that used user stylesheets to restyle websites to have a consistent look and feel. People make that effort to build ad blockers but not to build consistent looks, so I guess there isn't that much demand.
Yes, it’s wonderful having to figure out 50 different UIs designed by 50 different artists with 50 different ideas of what a drop-down should look like and how it should work.
The mental overhead from having slightly different drop shadows and button sizes is minimal, and I think pretty overblown by people who prefer identical UIs for everything regardless of form factor (which is of course a valid take).
But someone could just as easily respond to today's UIs with "Yes, it's wonderful that every single app looks identical, as if it was all designed by one pretty boring artist with no creativity whatsoever" and that would also be a perfectly valid take.
Predictability is a feature. That does include button sizes. To me that's where the asymmetry is. The "boring artist with no creativity" complaint is aesthetic, whereas predictability is a functional concern.
It’s more like I have to actually think about where the program options are. Or that on firefox nowadays I have to snipe a tiny small spot on the title bar to move the window, instead of clicking some button that doesn’t look like one.
When I first used MacOS I was surprised about tute consistency to access the settings of every program, even third party, with the same shortcut cmd+,
Do you have a solution in mind? While a platform (OS) can provide a UI toolkit and provide a HIG, one cannot stop language and programming tools vendors or programmers from doing whatever they like.
I'm starting to think that it would take replacing basically everything that's happened on Web frontend development since XMLHttpRequest with an alternative system that's still standards-based, platform-agnostic and Web-centric, but designed from the ground up as a GUI toolkit instead of a markup language for hypermedia formatting.
Because with the current status quo, the platform that dominates everyone's mindshare is HTML/JavaScript/CSS. Which has a really rudimentary concept of UI controls, and human interface guidelines that spend 90% of their effort on begging people to manually implement usability features that we used to get for free with native GUI toolkits. And I think that we might need to get away from that mess before it's possible for anyone to have any energy left over for worrying about HCI on the level that we used to in the late '90s and early '00s.
But realistically, you’re never going to stop a motivated app designer who is dead set on making their app an unique snowflake art project rather than a tool that users need to use.
There are also some classes of apps where the platform UI kit is insufficient. Immediate examples that come to mind are kiosk software, trading software, games, etc.
It’s to be expected when almost every app is electron or some web wrapper, that all consistency is lost. The only way to get it back (kinda) is to avoid those apps.
But for a lot of use cases that's not possible anymore, that is, a lot of the applications using the OS' UI libraries are no longer maintained.
And a lot of people have to use the applications as supplied, e.g. Slack or Microsoft Teams. Which can be accessed via a web browser, sure, but dedicated apps for these are also nice because they have a dedicated spot in the app switchers.
The OS / UI toolkit should be strongly opinionated, making the consistent, happy path easier to develop and making customization possible but with great effort.
But people would claim that it's easier for a company to take a bunch of web frontend devs and have them develop a UI which rides on top the already-existing Electron. Which is why we have such a plague of bloated Electron apps - because companies are lazy and don't care about the end user experience.
I like the idea of trying crazy and new ideas, but this looks like they just thought corners weren't round enough, and that people will pay money for a file manager that has no sharp edges and won't integrate with your OS.
Nobody forces you to install and use apps made of a different toolkit (or version of said toolkit) from the one shipped with the desktop.
You can use only Cocoa apps on MacosX, qt6 apps on a kde plasma 6, gnome/gtk4 apps on a gnome3 desktop or whatever is the equivalent in the windows 11 world.
> What I dislike the most on modern UI, and maybe absolutely hate, are all those super slow animations.
Slow animations are a way to hide latency, they are essentially loading screens. Apple is really good at it, or at least it was with the early iPhones, and a reason why iPhones felt so smooth compared to their Android counterparts while not being actually faster. For me, it is an impressive technical feat and it took years for Android to catch up (see: "project butter"), and in the end, it was mostly by brute force, i.e. putting ridiculously overpowered hardware in smartphones.
Remove the animations or make them faster (you can do that sometimes), and the lag may become apparent.
Why you have latency to hide in the first place is another problem. There may also be some clueless designers who put slow animations for no good reason, maybe because they are just copying Apple, not understanding why Apple did it in the first place.
There are also some animations that that have utility beyond eye candy in communicating to the user what’s going on, which is particularly important for non-technical individuals.
For example the animation associated with minimizing windows in most desktop environments makes it crystal clear where your window went after you press the minimize button, even for novices. Removing that animation makes the interaction significantly more confusing.
Those animations are a no brainer because they both communicate something meaningful and don't get in your way. While the animation is happening you can keep working with whatever you were doing. I think a great UI can both be animated and allow you to work unhindered, if the designers put their mind to it.
Well, the iPhones were in fact faster. Faster at playing the animations, at least.
I worked on an app in the iPhone 4S and Galaxy S II era and we wanted to use the same trick on both: smoothly animate the view switch between user interaction event and the API response. It worked super smooth on iPhone, and it was jittery as hell on Android. In the end we left the animation on the former, and move the users straight into the loading screen on the latter.
> Slow animations are a way to hide latency, they are essentially loading screens.
Except that most of the time there really isn't any latency to be hidden, the action becomes effectively instant once you remove the animation. Starting a new app (or switching to an app that was evicted from memory) is the main exception and that's quite rare.
> Remove the animations or make them faster (you can do that sometimes), and the lag may become apparent.
This is my number one trick on Android phones. Enable developer options and change the animation speeds from 1x to 0.5x. It makes your old phone feel new.
> Apple is really good at it, or at least it was with the early iPhones, and a reason why iPhones felt so smooth compared to their Android counterparts while not being actually faster.
Is that why iOS animations always feel so slow to me? Modern phone hardware can do things so much faster, but the animations are still utterly sluggish in my opinion. Worse, there's no way to speed them up; even with reduced motion, slow movements are simply translated into just-as-slow fades, which are somehow even more obnoxious.
> Slow animations are a way to hide latency, they are essentially loading screens. Apple is really good at it, or at least it was with the early iPhones, and a reason why iPhones felt so smooth compared to their Android counterparts while not being actually faster.
Now it got flipped. I turned off animations on my Android phone, and it's great. And now every time I have to use iOS (for app development), everything seems to be moving in slow motion.
And you can not turn it off! Apple in their infinite wisdom doesn't provide ways for app developers to disable animated transitions.
My No.1 pet peeve is the scrollbars. Somehow every modern UI designer hates it, and hates it deeply. They always want to get rid of it. And TBH I'd prefer the other way around - how about we get rid of them instead?
There’s a similar disdain for menubars which I really can’t understand. The disorderly and abbreviated hamburger menus that most often are used as a replacement are just worse on every single axis except for maybe visual appeal. They throw out what could be the single strain of consistent usability across apps in favor of looking good on a PowerPoint slide and web marketing blurb.
Two of GNOME's recent updates have made searching menus for a rarely used item incredibly painful:
1. Replace the menubar with a hamburger menu; in some cases the hamburger menu then contains file/edit/&c. so it's just a spurious extra click
2. Require a click to see the contents of a submenu and a click to go back
Fortunately my most-used GNOME application (Evolution) has an option to restore the old behavior for both of those, but I literally cannot think of the motivation for these two changes that clearly make things worse. The only halfway plausible idea I have heard for #2 is that the GNOME UX designers think that submenus are bad, so if you make them hard enough to use, developers will stop putting them in their applications. #1 is probably partly a looks thing, and partly a "too many people have fewer horizontal lines on their screens than I did in 2004[1]" thing.
1: That's when I got a 1600x1200 monitor; people today with 1080p screens have only 56 more lines than the 1280x1024 monitor I had been using since the previous millennium
The GNOME project seems like it has a strong desire to converge all on-screen UI onto mobile-like patterns. There are some mobile conventions that can be brought over to desktop without impairing usage too much but I think that perhaps it’s starting to cross too far over to the mobile side of the line.
It’s unfortunate because in other ways I find GNOME/GTK more agreeable than KDE/Qt (layout of controls within windows is consistently better in GTK environments/apps for example, Qt apps have a tendency to feel slapdash/haphazard/“engineery”) but I don’t like the increasingly strong mobile influence.
I agree. Hamburger menus aren't any better than menu bars. It seems like an example where design has more importance than function.
My alternative to the menu bar would be a search bar that allowed me to search in a Google style everything related to that program: functions, features, shortcuts, and documentation.
File | Edit | View | etc. is not the right choice for every program.
This is where it’s useful for the menu bar to be system-owned rather than the responsibility of individual programs (whether that be a global bar as in macOS or attached to window decorations). That would make it easy to implement a toggle that hides menubars either globally or on a per-app basis and enable a Unity-type shortkey-summoned HUD to be used instead.
Hamburger menus work much better on mobile screens that are horizontally constrained, are less visually intrusive when not in use and don't require at least two levels of nesting like classical desktop menu bars do.
Even on mobile, in most situations there are better UX choices than hamburger menus. Nearly every app using them I’ve encountered has put as little thought into navigation and UI hierarchy as possible and are awkward to get around in. They’re the ones where one has to go down a winding ever-changing maze to access the desired functionality.
Most would be better served by surfacing the most commonly used screens as tabs and most commonly used functions within those tabs. Ideally 2-4 taps is all it should take to get anywhere, and there should only ever be a tiny handful of niche things that take 5+ taps to access.
Agreed. And the worst part is that you could use a (well designed) menubar with a keyboard by using the Alt-key combinations together with cursor key menu navigation and similar techniques. But you don't have that luxury in hamburger menus. You are forced to use a pointing device such as the mouse, or if you are lucky, a completely non-standard key combination to bring it down. Awful.
Certain long emails I get don’t show the scroll bar at all in iOS Mail, and I get low grade anxiety not knowing how long the email is or how much more is left.
I’m also perplexed why the mail developers would allow such a thing or what kind of bug causes such behavior.
My hunch is that they are just using an existing framework that does that, and it may require some digging into the configuration to disable that (or worse, have to change some code). Since this is never going to be an urgent thing it will never be fixed.
I work in MacOS VSCode frequently, and whenever I open a large repo with a huge number of files, it's PIA to find the scrollbar. I have to hover the mouse above it to make it appear, but how can I hover above it without knowing where it is?
> Certain long emails I get don’t show the scroll bar at all in iOS Mail, and I get low grade anxiety not knowing how long the email is or how much more is left.
Yet in iOS you can swipe vertically some pixels and you will see the scrollbar telling you this exact information.
> What I dislike the most on modern UI, and maybe absolutely hate, are all those super slow animations.
This is what drives me crazy on macOS. Specifically, the animation for switching between virtual desktops. When I hit Ctrl+1/2/3/etc I want it to switch instantly, no animation - not slide into place. It's even unresponsive until the animation finishes.
I use the Aerospace tiling window manager for macOS just so I can move my apps to different spaces and opt+key move to them. usually vscode in opt+1, firefox in opt+2 and discord in opt+d
I have to use a Mac for work and it drives me crazy. I disabled everything I could find, enabled reduced motion preferences or whatever it's called and still got some animations.
My "solution" is to use Aerospace [0], which reimplements window management. That's the only way I found to not have animations. Unfortunately i still feel some delay when switching windows compared to i3wm/sway in Linux.
I hate the default animation speed of Androids that come with Pixel phones. They are too long and makes it feel like the phone is slow. One of the first things I do after buying a new phone is to half the animation durations using developer settings, and the phone feels much faster.
Agreed! Whenever I use someone's phone I instantly notice how sluggish it feels with animations turned on. If I offer to turn them off they often get surprised at how much faster the phone feels after.
My pet peeve: Animations are a crutch used by designers who think they need them when in fact they should just have improved the UI so users don't get confused about the origin of a popup or window. The only justified use of animations in UIs that make sense is in scrolling, everything else is just adding latency to hide your incompetence.
> If I offer to turn them off they often get surprised at how much faster the phone feels after.
If you're using Android there's also a "visible touches" option you can turn on in the Developer settings. It's a big UX enhancement of its own and IMHO should be promoted to the Accessibility settings (together with the options for speeding up or disabling animations).
It provides a quick feedback loop when you're trying to poke at stuff on-screen with your fingers. Which is nice since it lets you know quite seamlessly (1) how good your aim is, which gets kind of critical when using fingers on a touchscreen compared to a mouse. and (2) whether a tap has even registered in the first place, which is often problematic for many users who may be wearing gloves, have drier skin that doesn't register as easily, etc. etc.
Animations are pretty, I am not disputing that. And there is probably a large segment of users who like them. It is, in my opinion, a preference of form over function as animations almost always make a particular workflow slower to complete.
A large part of the charm of these 90s UI recreations is precisely the lack of antialiasing and other niceties we expect of modern UIs. There was another project recently on HN that uses modern font rendering with a Windows 9x look, and it's just not the same, IMO. SerenityOS comes closer to what I remember, though it still doesn't quite match the look of MS Sans Serif(?).
I agree that it is charming when I want to tinker with stuff. But if it's for "real world " usage, I don't think that being charming matters a lot. I mean, it really depends, not everything has to be about "serious" usage but if it's the intended goal then I think that good font rendering is something that still matters a lot.
I still use chrome sometimes for example just because it seems to have a better font rendering (on Linux but also on Windows) than Firefox. It's completely irrational in a way but it does matter sometimes
AFAIK there was a case being made at the time that the animations were to provide predictability in performance, users get into a bit of a rhythm and it was better to slow everything down a bit to lower the overall variance of response. This made more sense when HDDs were slow.
> What I dislike the most on modern UI, and maybe absolutely hate, are all those super slow animations.
I get it, and I agree. But what I personally hate the most on modern UIs is hiding things. Why aren't my scroll bars visible when I'm not interacting with them (and even when visible, are ridiculously small and low-contrast)? Why does IntelliJ hide the buttons for interacting with tool windows until I mouse over where they should be? Why does MacOS hide my application launcher bar by default? Stop hiding things!
These statements always catch me a bit off-guard. Is there no such thing as a cancelation period in the US? When my employer wants to kick me out, he needs a good reason for that and I'd still be paid for 3 months. Which is often even longer, depending on how long you belong to a company.
In my experience, it doesn't matter anyway. You can be paid to be sat at home and, while the worry of finances is kicked down the road, the big dark questions come home to roost very quickly.
In Canada in general either employer or employee can terminate without having to give reason. Typically it's either a few weeks of notice for termination but the employer can choose to require the employee to depart immediately and instead payout severance for equivalent time instead. There's nuance province to province though.
In the US it's similar but AFAIK it does vary state to state. To my knowledge there isn't any law that requires what you're describing in North America.
I have seen quite a few times in my career large US tech corporations specifically choosing not to open a satellite EU sales office or a dev office in DE because of the horrendous labor laws.
Sure, very nice for the workers. But foreign money chooses to skip DE because of this.
If you asked them about it though, they'd say the labor laws were excellent and the American labor law is horrenous. It's about perspective, but let's face it - America is way more set up for innovation than any other Western country.
There is a WARN[1] period before the employee is officially laid off, but their access to all the corporate resources are cut off immediately. From the employee perspective, they have lost everything the moment they are being told that they are laid off. It doesn't matter that they are still getting paid.
I started playing with vim, but neovim is another level. It's my goto for coding over vim. I don't expect them to put more effort into anything, because it just works as a editor/IDE.
No offense here, but I am wondering what's the relation between the topic being discussed here and vim/neovim?! I also checked your profile and again, wondered why you are putting this: "Only admins see your email below. To share publicly, add to the 'about' box." as your bio!
> I am wondering what's the relation between the topic being discussed here and vim/neovim?!
It's a similar religious matter?
> I also checked your profile and again, wondered why you are putting this: "Only admins see your email below. To share publicly, add to the 'about' box." as your bio!
Why not? I have "Error retrieving about text" in my "About" field on a certain sms replacement chat app...
> Xilinx tech support page and a forum post explaining how to get the Xilinx ISE to run under Windows 8 or Windows 10. Unfortunately Xilinx no longer maintains this development suite, but has also considered it unnecessary to support their Spartan 6 platform in the successor software suite, Vivado…
Wasn't the Spartan 6 also supported by the open source toolchains? I just did a couple seconds of search and I could only find the yosys support, but no nextpnr. Sad. Xilinx/AMD should open that up
6 figures? I only take jobs with free fruit basket. That's the minimum.
PS, my references: I have never used cursor, I am quite bad at vibe coding and don't enjoy it at all. I rarely even use AI for help. But I am quite decent at FPGA design and embedded developemnt. If you have a job for me in germany or remote in europe, then I will pay you with very bad humor every week. Possibility of using linux+neovim is a requirement though. (Yes, I really need a job)
It’s pretty normal for programmers to make over $100k in plenty of companies in much of the US. Big tech companies may pay a reasonable amount more when you consider equity, especially in high cost of living areas.
Programmer salaries have diverged a lot between the United States and Europe in the last 30 years. (This comes up basically every time remuneration is discussed on this forum). Hardware engineer pay has not seen such rises in the United States or Europe though I don’t know how much FPGA work ends up in the hardware bucket rather than the software one.
Some financial firms will pay well by American standards in Europe and also make use of FPGAs, though I don’t know how many you’d find in Germany.
Programmers are also well paid in Switzerland.
Maybe one thing to add: ‘bad humor’ is a bit hard to parse in English – it sounds like a cross between ‘bad humored’ which roughly means ‘unpleasant to be around’ and ‘bad sense of humor’, which is what I think you intend.
Not sure either of those capture the nuance of bad humour… ‘bad sense of humour’ can imply not finding anything funny or taking offence to jokes. ‘Bad humour’ in this context implies bad at being humorous, e.g telling dad jokes
Yeah sure, I know that a lot of people make 6 figures. The comparison to switzerland is not that great though, giving the cost of living. There is not much difference of me getting a 100k CHF in switzerland or 60k€ in germany
I shared this sentiment. But since I just host some personal fun projects and I got really lazy when it comes to self-hosting, I found great pleasure in just creating the simplest possible docker containers. It just keeps the system super clean and easy to wipe and setup again. My databases are usually just mounted volumes which do reside on the host system
I'm sure you find some joy in it or just like to explore what's possible. But just a reminder: Pieter Levels is running his million dollar businesses on a single VPS (if that's still correct). But yeah, if you like it, why not.
"Premature clustering is the source of all evil" - or something like that.
But I just wanted to comment something similar. It's probably heavily dependend on how many services you self-host, but I have 6 services on my VPS and they are just simple podman containers that I just run. Some of them automatically, some of them manually. On top of that a very simple nginx configuration (mostly just subdomains with reverse proxy) and that's it. I don't need an extra container for my nginx, I think (or is there a very good security reason? I have "nothing to hide" and/or lose, but still). My lazy brain thinks as long as I keep nginx up to date with my package manager and my certbot running, ill be fine
%20(1).jpeg){kind=link}