If someone wants really low ram consumption for a desktop. They should try out tinycorelinux which I have ran the whole system in <25/20 MB of ram from its most minimal option.
It's truly the most minimalist gui option just out there. It uses flwm & there own iirc very minimalist xorg server but most apps usually work
The one issue I have is that I can't copy paste text or do some simple stuff like moving my mouse on some text but aside from that, Tinycorelinux's pretty good
I put tinycorelinux on an old laptop a family member was looking to get rid of. It was the only OS I could find that still supported the ancient cpu.
It worked ok, but had a bit of a learning curve. I also had to run a couple commands every time I booted it up if I wanted to connect to wifi. I tried to get this to happen automatically, but wasn't having much luck. The password for the network also gets stored in plain text, so there was that. I didn't spend too much time on it, since it seemed like it was ultimately headed for the recycle bin and they just wanted to make sure none of their data was there, but thought if it worked decently well, maybe it could still be kept around and used.
NetBSD + xfce is also decent in this scenario
> The one issue I have is that I can't copy paste text
With pure X11 you copy paste via primary selection and middle click.
Hmm I have always been in the wayland world (KDE,Hyprland etc.) but I have been on xfce in mxlinux and I didn't need to do primary selection & middle click
I don't know if tinycore supports this. This was my biggest grievance because I had to create tmp files paste into it and then cat into it or something to work with this pain (which I feel like is pretty fixable/ maybe a skill issue from my side and honestly wishing for me to learn how to fix it)
that one issue sounds like a deal breaker to me. I mean, I copy and paste all the time. The one thing I wish tc would do is have a searchable package like most distros do instead of providing a large text file of all packages. Shouldn't be too hard to implement but whatever.
I mean fzf might help?
Can your "one issue" be tweaked by adding more RAM and allocating it thusly?
I'm using Void with 24gb ddr5 and frequently get system freezes during high productivity. Browser tabs in the background are often contributors, but working with openshot or odb crashes often.
I have several old nuc's and I might try tinycore on one. What do you or most others use it for, primarily?
I am not sure how my one issue can be fixed. It seems to be fundamentally an issue of their minimalist xorg server itself but I am pretty sure that there must be a way
> I'm using Void with 24gb ddr5 and frequently get system freezes during high productivity. Browser tabs in the background are often contributors, but working with openshot or odb crashes often.
Kdenlive's' pretty good for what its worth and I use Archlinux/cachy on an 8 gig system and browser tabs aren't that often atleast in here
> I have several old nuc's and I might try tinycore on one. What do you or most others use it for, primarily?
I used it to revive my 15 year old laptop and even ran complete modern firefox on it (its specs are 1 gigs 32 bit ram simple mini laptop) and ran wifi and ran firefox and ran pomodorokitty on it and I can sort of treat it as a second monitor
It's battery is removable so I am gonna change its battery as currently the setup takes time to install and I have to install it everytime I open/it shuts down which can happen quite a lot if I don't have it plugged in so currently its shutdown for over a month but I really liked the tinkering I did with when I ran pomodorokitty on it
I'm not sure if I understood your issue correctly but you can persist your configuration with all diskless (os is entirely in RAM) OSs as far as I know. This way you wouldn't have to install the setup after every reboot. Here is the guide for tinycore:
https://wiki.tinycorelinux.net/doku.php?id=wiki:persistence_...
Ah yes sorry forgot about persistence since I played with it some long time ago and the details were blurry
yes I probably could do that and most likely would on the laptop but I really wanted to tinker with tinycore a lot first so I was using the non persistence mode
I will probably do it later when I replace my old mini laptop's battery with a new (I think it costs less than a $ or so I have heard) but the procastination aspect is gonna have me do it to find a good shop around me to have the part etc. to probably and I am thinking of doing it after a few months but the mini laptop's still in my room :) (all be it off)
I've used it for mostly system rescue operations but it can do much more. Look at the package list for possibilities.
If the system totally freezes such that you can't even ssh in, that's just flaky hardware and you should replace it.
It's a newer Lenovo vpro, not because I wanted that, but because it's what I got. It came with 16g of reputable ram, then I added 8g ~1 year ago for $20, the exact same module which is now $120. Orher than a bad ram chip, what else would be the culprit?
I have 64gb in my linux machine and have managed to hardlock it a bunch of times exhausting the ram. Couldn't even REISUB a couple of times. The OOM killer stuff in Linux just doesn't work anymore by what I can gather.
Buying more ram is no longer an option, so I added a 128gb swap partition on nvme. I incorrectly assumed with 64gb I didn't even need swap. No crashes since.
If you don't want to move partitions around, you can add a swap file. ChatGPT or whatever can give instructions.
You should consider some sort of swap on ram like zswap rather than thrashing your nvme.
I used to run KDE and GNOME on a computer with 256 MB RAM back around the year 2000. Athlon 1000 Sempron and a Duron 800 (one of these machines started out with 128 MB RAM). KDE 1.x, 2.x, GNOME 1.x, 2.x. I don't remember the very minor versions. I tried a myriad of Linux distributions, and FreeBSD as well. I settled for Debian. Back then, we (me, friends, family, etc.) thought these DE's were very bloated. I remember KDE 1.x very vividly because I had to compile it myself (or look online for binaries), and I digged the CDE theme. The first lightweight DE (if you discount fvwm) I used on Linux was XFce, but that was later on. I pretty much started with KDE, tried a bit of GNOME, went back to KDE (I came from Windows 9x). In the end, I learned to appreciate GNOME, and MacOSX or Mac OSX as I used to call it back then (proper name was Mac OS X, I suppose).
My point is what you are used to is your reference point. The underlying OS isn't super relevant. On Linux, every distribution gets on par with each other eventually. On FreeBSD I used OSS and something like winmodem is just crap hardware. Nowadays my homelab and desktop have 64 GB RAM, while my MBP (M1Pro) only has 16 GB RAM which is the same as its successor (MBP 2015 with 16 GB RAM). Do I use all of that? Not really, but the main culprit is browser(s) (which includes apps these days). Curious if you can play Steam games well on FreeBSD. FreeBSD has a couple of neat things (tho ZFS is now better on Linux). I've always preferred PF to IPT.
I used to run Gnome on 8 MB in the late 90s
I still have some screenshots in my GitHub repository of what my ArchLinux with AwesomeWM (X11) looked like in 2009.
Those screenshots also contain the RSS, as luck would have it.
34MiB when on the desktop (clean), running X.org, AwesomeWM and xcompmgr (for compositing). Screenshot: https://github.com/aktau/awesome/blob/master/screenshots/200...
57MiB with a couple of applications open. From memory: urxvt running htop, thunar (XFCE file manager) and the Mirage image viewer (which is Python, not otherwise known for efficiency). Screenshot: https://github.com/aktau/awesome/blob/master/screenshots/200...
Nowadays, even with a tiling WM that's supposed to be lightweight (say: Sway), the minimum appears to be well over 300MiB (see https://www.reddit.com/r/linux/comments/1njecy5/wayland_comp...). GNOME 49 takes up around 1GiB last time I tried it (NixOS). Interestingly https://www.reddit.com/r/swaywm/comments/oghner/how_does_the... from 5 years ago mentions Sway only using 115MiB. What happened?
Theories I have:
- 32-bit to 64-bit means all points are double the size. That would account for something.
- Wayland vs X11. I should compare Sway versus X.org+i3.
- General library and system daemon bloat.When I was watching that Lunduke's video a couple of days ago initially I was thinking he's just making a joke of that Vendefoul Wolf distro on 200MB box. I recalled using FreeBSD as access server with lots of modems (PPP/SLIP), Apache, Samba and QuakeWorld server running on a box with just 32MB of RAM. That was also my daily working machine with XF86 and Enlightenment desktop manager, circa 2000. So, 200MB is a whole lot of memory!
While this is cool, it all goes out the window the minute you run any app
Exactly. The issue today is that even if you optimize your OS and DE to be very memory efficient, it matters very little as soon as you open a modern web browser. And without a modern web browser a big part of the online experience is broken.
Aren't you optimizing your OS/DE to be memory efficient in order to use that memory for other things (like more web browser tabs)?
Linux is already pretty decent on memory even with a DE like gnome/cinnamon running. I have an 15 year old laptop with Cinnamon / Arch and it uses 700-800MB once booted.
Frequently the issue isn't memory (I've maxed the memory out at 8GB).
Older systems have issues with composing the web pages itself. A lot of people blame JS, but a lot of the time JS isn't the issue e.g VSCode & Discord runs nicely on the laptop.
e.g. YouTube has a feature called ambient mode, this adds like a white gradient around the video element. On old intel iGPUs this absolutely wrecks video playback performance and you get about 15FPS video. Turning this off the video plays smoothly. The backdrop IIRC is a CSS gradient, the GPU/CPU doesn't have the grunt to render the backdrop gradient with the video on top at a smooth FPS.
So while you might have the memory for the extra tabs, browsing can be a sluggish experience because the GPU/CPU just can't execute the code on the page well.
Eh, kinda. Work forces me to have Jira, Confluence, Gitlab, Copilot, the other Copilot formerly known as Outlook, the other other Copilot formerly known as Teams, as well as Slack of course, and a dozen other webslop apps open… and it still all fits in <8GB RAM.
Which is a lot worse than the <1GB you'd get with well-optimized native tools, but try running Win11 with "only" 8GB RAM.
I'm convinced the next windows GUI will just be an electron app that runs copilot as the desktop, forcing you to argue with it to open a file or run a program. Doesn't even have titlebars or window buttons or a task bar, just one big copilot bar at the bottom that you can ask whats already running or to close an app. All of this written in JavaScript of course
Beautifully, blissfully broken
Running apps is what RAM should be used for, not wasted on the base system.
Depends if it is Electron crap, or properly native coded.
Unused RAM is wasted. But used RAM is also wasted, sometimes. If I can accomplish the same thing with less RAM, that's better, because it lets me do other things at the same time. It doesn't mean I'm not going to use that RAM, that would be pointless. My desktop running dwm typically idles at ~50GiB RAM usage from random crap I've got running. But I can prove that the desktop is using no more than like 300MiB.
Unused RAM is usually used by the page cache in modern OSs.
I remember booting up Debian into an X11 session on a laptop with only 8 MB of RAM.
(This would have been circa 2000, and I think I had to try a few different distros before finding one that worked. Also I don't think I did anything with it beyond Xterm and Xeyes.)
Ran linux in an 8 mb 486 in the 90s. X ran in 256 color mode and twm or mwm were the window managers. It was so hard to use though. Had to setup modelines settings for your monitor in a textfile and theoretically could damage it with wrong iputs. Programming X fuggedabout it - I was from turbo borland msdos land where everything was neatly documented and designed with clear examples to make programming easy. I was lucky to get an x program to even compile. Hard to find books back then. Pre Amazon. Xv image viewer probably the only thing i used X for. Actually used the machine most of the time in the text mode terminals using alt function keys and used lynx as a browser (before javascript… but gopher was becoming obsolete at that point… ftp still popular though ) with random assortment of svgalib programs for any graphical stuff. Still there was something magical about seeing that black and white check pattern come up and the little X mouse cursor appear.. like there were… possibilities.
Yeah, it was a different world. I worked at a company using X + Motif on SCO Unix back in the early 90s. I had a 386sx with 8mb ram + 6 MB on an ISA expansion card! When you changed a header file constant (like a label string) and had to recompile the ~1 MB(!) executable, it really was coffee break time - compile time was ~1 hour for a full rebuild. Strangely enough, our current project on a 16-core VM also takes nearly an hour for a full rebuild - but we have parallel build options that go much faster.
I also ran Linux+X11 on my 486 (for some grad work) with 32 MB, IIRC. ATI Mach32 graphics card, Nec 5FGe monitor (loved that one!), etc..
Yes, I remember making my 12" IBM monitor scream as I put the wrong mode information in the config file for X. I think I was on RedHat 5.0 from a cover CD, on a 486 DX2 with 64 MB of RAM (I was poor; everyone else was on Pentium IIs or IIIs and I was using computers the school threw out, scraping together motherboards and RAM).
I am amazed to discover that Xfce of that era was so CDEsque: https://www.linux.co.cr/desktops/review/2000/xfce-3.3/help.h...
It was originally created as a CDE clone (thus the original name "XForms Common Environment")
I'm always wanted to know what XFCE means
That would have been then already some kind of anachronism. 8MiB RAM was workable (but only barely so with X11) in the early nineties. Late nineties 64MiB or more were common.
Yes, it was an old laptop at the time.
Back in 1993, I remember booting SLS Linux on a 386 laptop with 3 megs of RAM (1 meg on the motherboard, 2 meg expansion.) I could barely get it to startx and open an xterm, so I mostly used it in from the console!
Before Linux, I was experimenting with Coherent.
I doesn’t feel like that long ago when I built a swarm of Arch Linux based thin clients which PXE booted from a SLES DHCP & NFS host.
That was probably around 2010 or 2015.
Those images had to run on a thin client with 512 MB RAM.
I think I chose XFCE as the DE.
In college we had a network of Sun workstations and some of the machines had only 8MB of RAM, IIRC. This was in the 90s.
Then again, the X desktop was really minimal and I would use them mostly to code in C using a terminal.
My first PC had 16 MB of RAM, which later obviously became too slow to be usable. I remember I had to wait around a minute for Fallout to load a level, which you had to do fairly frequently.
I remember buying a bulky external 2MB RAM extension (I think I bought another 2MB) before that for my Amiga 500 running a full desktop OS already on 512k 'Chipmemory' using it mostly to actually as a TempFS to accelerate loading. That was beginning to mid 90s, I guess. But running netbsd on the Amiga meant that you would already at that time need 16MB of RAM and a CPU with an MMU as well as an HDD (my friend across the street did that with his A1200 I think I remember). You would only do it if you wanted more networking beyond BBS I guess.
4 minutes to load Enclave level save game with pentium 200mhz with 32mb ram
32 MB? The opulence!
Me too, but I was able to do it around 1995-1996 :) Also remember Windows95 can boot with 4MB of RAM, and was decent with 12MB.
Windows95 was decent even with 8 MB, on a 66 MHz or 100 MHz 486 CPU.
With either 4 MB or only a 386 CPU, it was definitely crippled, making an upgrade not worthwhile.
Windows 95 on a 386 CPU with enough RAM was alright. Not fast but very useable.
https://youtu.be/Pw2610paPYM?t=72
But most 386 didn't have 8+ megabytes, and some 386 had a 286 like data bus, making it even slower. (386SX)
> Windows 95 on a 386 CPU with enough RAM was alright.
I benchmarked it for PC Pro Magazine when it came out.
We had to borrow a 4MB 386SX from a friend of the editor's, as we had nothing that low-end left in the labs.
In our standard benchmarks, which used MS Word, MS Excel, PowerPoint, Illustrator, Photoshop, WinZip, and a few other real apps, Win95 1.0, not 95A or OSR2, was measurable faster than Windows for Workgroups 3.11 on MS-DOS 6.22, hand-optimised.
When it needed the RAM, 95 could shrink the disk cache to essentially nothing. (Maybe 4 kB or something.) Win3 could not do that.
It was SLOW but under heavy load it was quicker than Win3 on the lowest-end supported hardware.
Under light load, Win3 was quicker, but Win95 scaled down very impressively indeed.
On paper a 386sx is slower than a 386dx, and certainly is in terms of RAM access. But in practice you'd need some expensive hardware to fully take advantage of that speed, like EISA cards and a motherboard that supported them (or, MCA cards on one of the higher end IBM PS/2 models). The typical ISA cards of the era were limited to 8 MHz and 16 bits no matter what processor or motherboard you used.
The 386dx could also use a full 32-bit address space, whereas the 386sx had 24 address lines like the 286. But again, having more than 16 MB would have been expensive at the time.
A few years back, I had fun setting up an old X11 terminal I had in my rather eccentric retro computing collection.
But I don’t think I had much memory in it. I had ordered a fair bit more, but maybe only 4-8M.
I did get it to work with only minor difficulties, but man only the simplest of applications could run. The barebones basic GUI text editor that came with Ubuntu couldn’t even start up.
I don’t know how resolution maps to ram in x11 but I assume at least one byte per pixel. Based on that assumption, there’s no chance you’d even be able to power a 4k monitor with 8mb of ram, let alone the rest of the system.
This was the main driver of VGA memory size for a time - if you spent money on 2MB card instead of a 1MB, you could have higher resolution or bit depth.
if you had a big enough framebuffer in your display adapter, though, X11 could display more than your main ram could support - the design, when using "classic way", allowed X server to draw directly on framebuffer memory (just like GDI did)
Correct, 4k is very modern by these standards. But then I'm old, so perhaps it's all about perspective.
Back in the days when computers had 8MB of RAM to handle all that MS-DOS and Windows 3.1 goodness, we were still in the territory of VGA [0], and SVGA [1] territory, and the graphics cards (sorry, integrated graphics on the motherboard?! You're living in the future there, that's years away!), had their own RAM to support those resolutions and colour depths.
Of course, this is all for PCs. By the mid-1990s you could get a SPARCstation 5 [2] with a 24" Sun-branded Sony Trinitron monitor that was rather more capable.
[0] Maxed out at 640 x 480 in 16-colour from an 18-bit colour gamut
[1] The "S" is for Super: 1280 x 1024 with 256 colours!
X11 was designed to support bit depths down to 1 bit per pixel.
X11, yes, but the modern graphics cards, no.
graphics cards at the time X11 was designed though, yes
Presumably every pixel is 32 bits rather than just 8. So the count starts at 33.2MB just for the display.
It is now, but back then it was 1 byte, with typical resolutions being 800x600. There were high-color modes but for a period it was rare to have good enough hardware for it.
I have run x11 in 16-color and 256-color mode, but it was not fun. The palette would get swapped when changing windows, which was quite disorienting. Hardware that could do 16-bit color was common by the late 90s.
Fun thing - SGI specifically used 256 color mode a lot, to reduce memory usage even if you used 24bit outputs. So long as you used defaults of their Motif fork, everything you didn't specifically request to use more colors would use 256 color visuals which then were composited in hardware.
Much better to stick to 1 bit per pixel. :-)
Like in Sun SPARCStation ELC. No confusing colors or shades.
1bpp (at low resolution) is still relevant today on epaper screens, though some of them now allow for shades of grey or even color.
Most aren't all that low res either... 300dpi is standard.
But what if it's a UTF8 bit? Then it'd be 2 bits.
Which proves time travel exists, all those "two bits" references in old Westerns.
Damn pixel bit-depth bloat!
Good thing 4k monitors didn't exist in 2000
My comment was tongue in cheek while simultaneously highlighting that at least some increased ram consumption is required for modern computing, and highlighting how incredibly far technology has come in 2.5 decades.
Consider the Sharp Zaurus (SL-C860, for example): - Intel Xscale PXA255 400 MHz - 64 MB SDRAM - Linux/QTopia desktop environment
At the end of the post there is a comparison of ram usage of different desktop environments and the used ram is reported differently by every tool. So what exactly is being here measured as the used ram?
It used to be like that, computer had limited resources and desktop environments were light. Then at some point RAM became less and less of an issue, and everything started to get bigger and less efficient.
Coyuld anyone summarize why a desktop Windows/MacOs now needs so much more RAM than in the past? is it the UI animations, color themes, shades etc etc or is it the underlying operating system that has more and more features, services etc etc ?
I believe it's the desktop environment that is greedy, because one can easily run a linux server on a raspberry pi with very limited RAM, but is it really the case?
> Coyuld anyone summarize why a desktop Windows/MacOs now needs so much more RAM than in the past
Just a single retina screen buffer, assuming something like 2500 by 2500 pixels, 4 byte per pixel is already 25MB for a single buffer. Then you want double buffering, but also a per-window buffer since you don't want to force rewrites 60x per second and we want to drag windows around while showing contents not a wireframe. As you can see just that adds up quickly. And that's just the draw buffers. Not mentioning all the different fonts that are simultaneously used, images that are shown, etc.
(Of course, screen bufferes are typically stored in VRAM once drawn. But you need to drawn first, which is at least in part on the CPU)
Per window double buffering is actively harmful - as it means you're triple buffering, as the render goes window buffer->composite buffer->screen, and that's with perfect timing, and even this kind of latency is actively unpleasant when typing or moving the mouse.
If you get the timing right, there should be no need for double-buffering individual windows.
You don't need to do all of this, though. You could just do arbitrary rendering using GPU compute, and only store a highly-compressed representation on the CPU.
Yes, but then the GPU needs that amount of ram, so it's fairer to look at the sum of RAM + VRAM requirements. With compressed representations you trade CPU cycles for RAM. To save laptop battery better required copious amounts of RAM (since it's cheap).
The web browser is the biggest RAM hog these days as far as low-end usage goes. The browsing UI/chrome itself can take in the many hundred megs to render, and that's before even loading any website. It's becoming hard to browse even very "light" sites like Wikipedia on less than a 4GB system at a bare minimum.
> is it the UI animations, color themes, shades etc etc or is it the underlying operating system that has more and more features, services etc etc ?
...all of those and more? New software is only optimized until it is not outright annoying to use on current hardware, it's always been like that and that's why there are old jokes like:
"What Andy giveth, Bill taketh away."
"Software is like a gas, it expands to consume all available hardware resources."
"Software gets slower faster than hardware gets faster"
Sometimes there are short periods where the hardware pulls a little bit ahead for a few short years of bliss (for instance the ARM Macs), but the software quickly catches up and soon everything feels as slow as always (or worse).
That also means that the easiest way to a slick computing experience is to run old software on new hardware ;)
Indeed. Much of a modern Linux desktop e.g. runs inside one of multiple not very well optimized JS engines: Gnome uses JS for various desktop interactions, and all major desktops run a different JS engine as a different user to evaluate polkit authorizations (so exactly zero RAM could be shared between those engines, even if they were identical, which they aren't), and then half your interactions with GUI tools happens inside browser engines, either directly in a browser, or indirectly with Electron. (And typically, each Electron tool bundles their own slightly different version of Electron, so even if they all run under the same user, each is fully independent.)
Or you can ignore all that nonsense and run openbox and native tools.
> Much of a modern Linux desktop e.g. runs inside one of multiple not very well optimized JS engines
A couple of years ago I saw a talk by Sophie Wilson, the designer of the ARM chip. She had been amused by someone saying there was an ARM inside every iPhone: she pointed out that there was 6-8 assymetric ARM cores in the CPU section of the SOC, some big and fast, some small and power-frugal, an ARM chip in the Bluetooth controller, another in the Wifi controller, several in the GSM/mobile controller, at least one in the memory controller, several in the flash memory controller...
It wasn't "an ARM chip". It was half a dozen ARMs in early iPhones, and then maybe dozens in modern ones. More in anything with an SD card slot, as SD card typically contain an Arm or a few of them to manage the blocks of storage, and other ARMs in the interface are talking to those ARMs.
Wheels within wheels: multiple very similar cores, running different OSes and RTOSes and chunks of embedded firmware, all cooperatively running user-facing OSes with a load of duplication, like a shell in one Javascript launching Firefox which contains a copy of a different version of the same Javascript engine, plus another in Thunderbird, plus another embedded in Slack and another copy embedded in VSCode.
Insanity. Make a resource cheap and it is human nature to squander it.
Which is baffling as to why they chose it - I remember there being memory leaks because GObject uses a reference counted model - cycles from GObject to JS then back were impossible to collect.
They did hack around this with heuristics, but they never did solve the issue.
They should've stuck with a reference counted scripting language like Lua, which has strong support for embedding.
A month with CrunchBang Plus Plus (which is a really nice distribution based on Openbox) and you'll appreciate how quick and well put together Openbox and text based config files are.
I've found that Gnome works about as well as other "lighter" desktop environments on some hardware I have that is about 15 years old. I don't think it using a JS engine really impacts performance as much as people claim. Memory usage might be a bit higher, but the main memory hog on a machine these days is your web browser.
I have plenty of complaints about gnome (not being able to set a solid colour as a background colour is really dumb IMO), but it seems to work quite well IME.
> Or you can ignore all that nonsense and run openbox and native tools.
I remember mucking about with OpenBox and similar WMs back in the early 2000s and I wouldn't want to go back to using them. I find Gnome tends to expose me to less nonsense.
There is nothing specifically wrong with Wayland either. I am running it on Debian 13 and I am running a triple monitor setup without. Display scaling works properly on Wayland (it doesn't on X11).
> I find Gnome tends to expose me to less nonsense.
IMHO, I find the reverse. It feels like a phone/tablet interface. It's bigger and uses way more disk and memory, but it gives me less UI, less control, less customisation, than Xfce which takes about a quarter of the resources.
Example: I have 2 screens. One landscape on the left, one portrait on the right. That big mirrored L-shape is my desktop. I wanted the virtual-desktop switcher on the right of the right screen, and the dock thing on the left of the left screen.
GNOME can't do that. They must be on your primary display, and if that's a little laptop screen but there is a nice big spacious 2nd screen, I want to move some things there -- but I am not allowed to.
If I have 1 screen, keep them on 1 screen. If I have 2, that pair is my desktop, so put one panel on the left of my desktop and one on the right, even if those are different screens -- and remember this so it happens automatically when I connect that screen.
This is the logic I'd expect. It is not how GNOME folks think, though, so I can't have it. I do not understand how they think.
> IMHO, I find the reverse. It feels like a phone/tablet interface. It's bigger and uses way more disk and memory, but it gives me less UI, less control, less customisation, than Xfce which takes about a quarter of the resources.
I've used Xfce quite a lot in the past and quite honestly most of the "customisation" in it is confusing to use and poorly thought out.
I've also found these "light DEs" to be less snappy than Gnome. I believe this is because it takes advantage of the GPU acceleration better, but I am not sure tbh. The extra memory usage I don't really care about. My slowest laptop I use regularly has 8GB ram and it is fine. Would I want to use this on a sub 4GB machine, no. But realistically you can't do much with that anyway.
Also Gnome (with Wayland) does a lot of stuff that Xfce can't do properly. This is normally to do with HiDPI scaling, different refreshrates. It all works properly.
With Xfce, I had to mess about with DPI hacks and other things.
> Example: I have 2 screens. One landscape on the left, one portrait on the right. That big mirrored L-shape is my desktop. I wanted the virtual-desktop switcher on the right of the right screen, and the dock thing on the left of the left screen.
> If I have 1 screen, keep them on 1 screen. If I have 2, that pair is my desktop, so put one panel on the left of my desktop and one on the right, even if those are different screens -- and remember this so it happens automatically when I connect that screen.
I just tried the workspace switcher. I can switch virtual desktops with Super + Scroll on any desktop. I can also choose virtual desktops on both screens by using the Super + A and then there is virtual desktop switcher on each screen.
I just tried it on Gnome 48 on Debian 13 right now. It is pretty close to what you are describing.
> This is the logic I'd expect. It is not how GNOME folks think, though, so I can't have it. I do not understand how they think
I think people just want to complain about Gnome because it is opinionated. I also don't like KDE.
I install two extensions on desktop. Dash to Dock and Appindicators plugins. On the light DEs and Window Managers, I was always messing about with settings and thing always felt off.
COSMIC is gaining ground as a JS-free alternative to current desktops, so hopefully you won't be limited to openbox and such.
Openbox isn't limiting me, Wayland still has no advantages for what I do with desktops.
As written on a sibling comment, maybe RAM being hard to get will bring some of that back.
I really needed to save to buy RAM sticks back in the day.
I am wondering if, with memory and storage prices skyrocketing, there will be more effort on making computing use less resources?
Unlikely. If you can't afford RAM, how can you afford the SaaS contracts that keep devs employed?
They typically also need GPU acceleration, these days, and that can be an even bigger bottleneck, with the drivers often not supporting older cards.
Running Alpine Linux with a minimal window manager gives me similar RAM usage, about 150MB
This is quite good !
My Alpine Desktop (Root on ZFS, Wayland/Sway) starts with about 550MB
Ye gods, that's a lot. I run Xfce on Alpine on plain old ext4 and get 200MB or so used.
I remember, in 2007, running FreeBSD on a desktop with 512MB RAM and only using 64MB of it running full GNOME 2 and a running instance of Firefox with a couple tabs. A totally standard desktop experience.
Even better, my laptop at the time had only 128MB of RAM and ran Windows XP - a supported, albeit minimal, configuration. XP was bloatier than FreeBSD of course, and ran correspondingly less well, but replacing explorer.exe with a shell called "blackbox" - an openbox-alike - and carefully curating applications (e.g. K-Meleon instead of Firefox) rendered it a perfectly viable multitasking desktop. I have a screenshot from that machine showing an AIM window, an mp3 player, an IDE for an embedded system, and a web browser with the documentation open for that IDE, all running comfortably (on one of its several desktops - yes you could have multiple desktops on XP with alternative shells such as blackbox).
Computers now require approximately 30x the RAM to achieve similar levels of "barely viable" performance - 4GB is considered the absolute minimum for general purpose desktop viability. And qualitatively speaking, what do they do now, that my 2007 fleet did not do? It is difficult to say. One is led to the conclusion that something has gone terribly awry with resource consumption.
> It is difficult to say.
It isn't: you can still download the 2007-vintage FreeBSD desktop and run it in a VM today if you'd like. The CD image-files are quick downloads with modern broadband speeds. Prepare to be disappointed though.
The computer I installed Slackware 2.0 was a P75 with 16 MB of RAM!
Naturally this cannot work when every application is an instance of Chrome.
I am glad for the RAM prices, maybe this will teach a new generation on how to care about their data structures again.
> I am glad for the RAM prices, maybe this will teach a new generation on how to care about their data structures again.
I have been thinking that exact same thought recently.
I hope prices stay elevated for a few years, so people learn to be a bit more frugal with resources.
It's the web browser and electron based apps that are the primary consumers of ram on my desktops with the DE and OS ram usage being minimal by comparison.
I have an ancient laptop from 2008 with 4GB of ram that runs a modern KDE desktop and related applications just fine that I use for troubleshooting stuff. However, the moment I open a web browser it basically falls to pieces.
I hate everything about this. :-/
> I have an ancient laptop from 2008 with 4GB of ram that runs a modern KDE desktop
Try Alpine Linux, with Xfce which can do most of the same things. Then enable swap compression -- add this to the end of the kernel line in your bootloader:
zswap.enabled=1
This compresses everything going to swap, and decompresses it coming back: less disk reads and writes, and less space used.
Everything gets quicker.
That's easy to fix:
Step 1:
sudo tee /etc/tmpfiles.d/mglru.conf <<EOF
w- /sys/kernel/mm/lru_gen/enabled - - - - y
w- /sys/kernel/mm/lru_gen/min_ttl_ms - - - - 300
EOF
Step 2:
apt install zram-tools
sed -i 's/#PERCENT=.*/PERCENT=130/' /etc/default/zramswapI'm on an older LTS kernel so no support for lru_gen but I will definitely check out zramswap - thanks.
Then also do this:
echo vm.page-cluster=0 >> /etc/sysctl.d/85-swappiness.confPeople are saying that 8GB for Android phone is too small. All of those applications are not electron either.
There is bloat everywhere.
4GB still seems excessive, by at least one and probably several orders of magnitude, for what vanilla KDE actually does: browse files, manage windows, and edit text. And KDE is one of the best modern options.
KDE doesn't need 4 GB of ram - that's just what my laptop happens to have which is more than ample to run the OS + KDE + native applications.
200MB for a desktop sounds massive to some of us :D
Back in the day I used to have a desktop running, with applications, in just 512KB. Getting that memory upgrade to a full 1MB was amazing.
Yup, fond memories of my Amiga 500+ (full meg, woo!)
Or the Atari ST! I have one at home with 1 MB of RAM in it and it still flies. Boots up in less than a few seconds, which is faster than any of my modern PCs.
And for many scenarios people use their computers for, it would still be enough today.
Of chip RAM, too!
I used to run Acorn's RISC OS in just 2MB of RAM.
I ran it in 1MB. :-)
And it was fast and responsive, too.
Soon afterwards I bought a Psion 3 which ran a multitasking GUI OS on an 8086 in 256 kB of RAM.
That space was shared with file storage in a RAMdisc.
https://phonedb.net/index.php?m=device&id=826&c=psion_series...
It was perfectly viable to have multiple apps open and flip between them. It ran for weeks on a pair of AA batteries.
A long time ago the power supply blew out in the machine I played Counter Strike: Source on and I was a teenager just barely 16 with no money so I couldn't replace it.
I was able to keep in touch with my drug dealers and my girlfriend's friends (who were also all super hot) which was very important to me at that age, in an environment where you really needed a car or people who had cars to do anything with anyone worth doing anything with.
I got OpenSolaris booted on a Pentium II box that had 384mb of RAM then ran Openbox and a communications suite of SILC, IRC, Pidgin, Finch (a text frontend to libpurple), and some XMPP+OTR clients -- all in Solaris Zones to not get my shit wrecked by the same RCE exploits I was using against other Pidgin users (which seemed to be as numerous as exploits for the official AIM client). This was before Facebook.
Solaris Zones gave me that feeling of power over software that Qubes enthusiasts like to talk about, similar dopamine+endorphin flow to being a military dictator of a 3rd world country. Shit was so cash.
Thanks to Unix' elegance, I still had a life until moved enough herb to assemble another box I could run Counter Strike: Source (on FreeBSD, Cedega for the win) on.
I’m surprised that OpenSolaris had hardware support for random Pentium II boxes, but I guess if you had a supported Ethernet card that everything else could work…
Thanks for letting all these nerds on HN know how important it was to maintain contact with a drug dealer and super hot girls when you were a hip teenager, I mean... i totally get it because I was also a really cool hip teenager. Did we just become best friends?
Think of the gravity that Instagram/Facebook has today, or maybe things are different today, so had for millennials. Try to take away a young adult's phone today, you'll risk being eliminated. We had some neat handhelds with PCMCIA slots that OpenBSD ran on in those days but it was only the kids in "rich" neighborhoods that also had them and I was a year behind in getting those. The critical mass of the network effect at that time was on desktops and iBooks.
> super hot girls
Yeah a San Francisco 7 was like an 8 in Los Angeles and easily a 10 in most towns (in those days).
They were prowling MySpace just as much as anyone else. You know what they're up to.
Cool post. So much could be done on a couple hundred megabytes of ram back in the day, with spinning rust as storage to boot!
Woulda been a nice article if it covered the real reason xlibre’s founder got fired from RH, Enrico’s had a long history of pissing people off and posting cringe on main: https://www.theregister.com/2021/06/11/linus_torvalds_vaccin...
Odd little rant suddenly appears in the middle of it :D
Here is how I set up minimal Desktop, WATCH 4 VIDEOS ABOUT HOW DEI IS KILLING OPEN SOURCE PROJECTS, and here is my loader.conf ...
The issue (I think) is that FreeBSD and other non-Linux, X11-using distributions are being ignored in the path to using Wayland; deprecating X11 has a much broader impact as a result, which leads to supporting XLibre which does support X11 and does support non-Linux Unices that are running X11.
Why should I care though?
Could say the same thing about why it's in the blog post.
You don't have to care at all. It's just an odd blog post that just from technical intro to rant about DEI and censorship and back to technical details. And joecool1029 just provides more context to what was said in the blog post.
The way Xorg is handling this is surreal. Now, they're trying to rewind Xorg 2 years just to get rid of XLibre's main developer's contributions[0].
Because his contributions are bad... Not because anyone has an agenda.
If that was true, then why did the reviewer approve and merge them?
Indeed. Xenocara would be a better bet. But the more these people have to collaborate with different people, the more empathy they will need.
https://git.sr.ht/~rabbits/fashware
About Nemo (Fran J. Ballesteros from plan9/9front) he has half as encuse as he grew up (for sure) under the Francoist regime probably from the loaded family side, and, thus, he had to swallow tons of literal extreme right wing ideology even at school (Franco's regime). But the point on being a conspiranoid about the Covid... I would expect more sanity from the mindset from a guy perfectly abled in algoritmics, math and by proxy, science. Echo chambers create these kinds of idiots even on really smart people (the far right in Spain used cult like mechanics too), and I'm sure Fran changed a bit over time for the better.
On the Cosmopolitan/APE person, I remind you that if you want to get back to Reissanance times, I'm a Spaniard, and thus, your whole ideology pales against the Iberian Humanism from the School of Salamanca, where at the time we were the Enlightened ones and you were just a bunch of WASP uneducated hicks living in filthy villages in the middle of Europe.
Back to 9intro, even if you dislike ~nemo, 9intro it's still worth to learn programming on 9front, it's a great book to share and learn from. If would be a waste to ditch it just because some old fart doesn't get into the times.
EDIT: ok, now I see ~nemo it's not that old, so a plausible indoctrination from the Francoism wouldn't apply there; but I'm pretty sure being a conspiranoid on Covid doesn't look like the normal socialization out there.
The future of computing, now that @sama is gobbling up all the RAM.
It says: that uses 217 MB RAM with Devuan, and Devuan is a fork of Debian13
Yes, 200MB RAM without any non-essential apps (not really useful unless for a specific use case).
Cant wait to boot up my Windows 11 total bloat machine at home and work
I kinda wanna try linux again...
it gets better every day
Just to remind people here: a single uncompressed "4k" picture is 33MB. Have your compositor hold 10 of them and you get 330MB just for the window images.
Across multiple monitors my desktop is 6400x2160, which at 32 bits comes to 55MB.
Considering memory is slow and GPU compute these days is cheap maybe it would make sense to relayout and rerender things each frame directly into screen buffer instead of keeping the window surface buffers resident. That would require rewriting quite a lot of things though.
You need window surface buffers in order to do seamless compositing, scaling etc. It's quite technically possible to achieve pixel-perfect 2D rendering with GPU-side compute, but in many ways it's still an open problem.
Scaling is for legacy apps, right? Modern apps should get the area to render and the desired pixel density.
Not sure what you mean specifically by "need" in "need ... to do compositing". Compositing is just a way (e.g. rerender only on changes, cache results) of running a desktop environment. Strictly speaking you don't need compositing, you can just use immediate mode across the DE and apps.
The tradeoff of course is that if an app is lagging you get a blank rectangle instead of a frozen picture. Well not quite 0 or 1. You can cache lowres and/or compressed frozen picture periodically to improve UX.
> Modern apps should get the area to render and the desired pixel density.
What if you want to smoothly slide an app window over to a second monitor with a different pixel density? That's admittedly a very rare thing, but some people seem to be obsessed with it and insist that it must work. You either have to compose some window surface, or just use clean vector rendering throughout.
Windows doesn't care and neither do I. But still, this can be done in immediate mode if the DE can tell the app it wants it to render the window in multiple rectangles with different pixel densities.
I have a hope for the whole idea, because imo it could significantly improve text rendering in VR by passing or allowing realtime access to the projection matrix along with the areas to render to. Regular VR compositing distorts text and vector graphics due to reprojection.
Plus, as noted above in VRAM speed vs GPU compute speeds, it might actually be faster and more power efficient overall if done right. See e.g. the famous Windows Terminal optimization issue with glyph atlas caching and object reuse.
Need vsync. Last thing I want is screentearing desktop.
Not sure what makes you think this precludes vsync. I already outlined a few options on what to do if apps fail to produce frames at refresh rate.
opens blog post
sees lunduke
closes blog post
seriously, what's with people's love of this guy? besides politics, I have not seen anything that suggests engineering prowess from this guy, only "rust bad".
I think he did a good job with a report on Mozilla's spendings. Also in general he shows a lot of cases of hypocrisy in the modern software industry.
People like his technical opinion because they like his politics. That’s the whole grift-influencer economy. If someone is good at one thing (and validates some of my views), then obviously he’s right about everything.
When people feel underrepresented to the point of being bullied they turn to any voice which seems to reflect even a tiny fraction of their frustrations.
There's a real mean spirit in open source lately and a lot of it seems to revolve around political views. There's become this idea that if you and I disagree on politics then it would be impossible for us to write quality software together. It's damaged a lot of good will and cohesion that used to exist within the open source software community.
This used to be about making free software to people so that they weren't abused by corporations. Now it's about pushing agendas and creating exclusion criteria. There's only one group in this scenario that benefits from this outcome.
If you don't like Lunduke then you should recognize the factors that give rise to people like him. Unless your solution is to completely eliminate anyone who disagrees with you then your apparent mindset only furthers the problem.
I wish we could put all this aside and just enjoy open source again.
My existence is not political. If someone doesn't think I should have rights and/or exist and/or thinks I am inferior because of who I am, then no, we cannot write quality software together.
If someone disagrees with me on tax, foreign relations, government services, defense, etc policy, sure, we can disagree and still work together.
What gives rise to people like Lunduke is not a simple thing, and something I don't think society fully understands.
In a way, "someone doesn't think I should have rights and/or exist and/or thinks I am inferior because of who I am" is pretty much the definition of (some kind of) politics. All sides play this game, e.g. many extremists these days argue that the "intolerant" shouldn't have rights or even exist by definition, but then the political football becomes who gets labeled as "intolerant" to begin with.
(And maybe it's true that those on opposite sides cannot work together on good software, but that's easily addressed since all FLOSS licenses include the right to fork and merge changes.)
Not agreeing with a particular description or categorization of you is not the same as thinking that you don't exist and not agreeing that you should have certain non-universal rights based on that categorization or that you should be able to force others in agreeing with your views isn't the same as thinking that you shouldn't have rights period.
When people believe "they are product", bully Open Source developers for not following their demands and got expected response than entities appear that validate their wrongs for views (money).
Lunduke spreads misinformation. That's anti Open Source, anti community.
> Lunduke spreads misinformation
He doesn't. He just reports events as a journalist. He doesn't fight against open source.
Name how it's possible to improve security on X11 without breakig changes.
Lunduke made factually wrong claims for hype. His mob are keen to attack Open Source developers.
You can use Xephyr or Xnest to sandbox an untrusted or insecure application within its own X11 instance. This gives you the exact same kind of security property that Wayland happens to enforce out of the box for its clients, except that it need not apply to basic desktop components such as the window manager or the desktop panel. You don't even need Xlibre or anything, this stuff has been around for ages. It's not rocket surgery!
>Name how it's possible to improve security on X11 without breakig changes.
Namespaces. It's been done already. Look into XLibre.
>Lunduke made factually wrong claims for hype.
Citation needed.
>His mob are keen to attack Open Source developers.
Doesn't own a mob, and never happened. Horrible accusation, by the way.
Dont present our hypothesis as a hard fact. I actually think it is completely false. Not only I was never interested in his political opinions, and followed him because of his humoristic takes "Linux sucks", and not about Rust or whatever; I actually never encountered a single video before joining his "lunduke journal" where his right-wing views would be visible.
He has made funny videos, it was fun to watch. Its kinda hard to enjoy them now after learning he s dumb as a rock and justifies killings if you are of tje wrong nationality
He is an influencer if you will.
Skilled enough but the main use is as a news resource like this. The guy ion the blog would not have found out about this unless Lunduke posted about it.
Do you understand? :)
For us not in the know, why is this bad?
Is he ""bigoted"" ? :(
The maker of the provocative "Linux sucks" series is a bit of a troll. He's made videos on technical projects he doesn't understand (or care about) and just mocks them if they don't gel with him. As far as I can tell he doesn't really care, or if he thinks he does - his actions aren't translating well.
How do I know? As a FOSS developer myself with a decade plus public history I also happen to know a few people running prominent FOSS projects.
He's burned bridges for no good reason. He doesn't care.
I have no idea who he is, never heard of him. You shall not judge a book by its cover but .. he is making it hard. His video titles are:
* Devuan: The Non-Woke Debian Linux Fork (Without Systemd)
* NeoFetch But in Rust and More Gay
* Chimera Linux is "Here to Further Woke Agenda by Turning Free Software Gay"
* Are Jews the Cause of DEI in Big Tech?
Yeah .. I did not watch a single video of his. But just from a short few seconds It's not anything I want to invest time in to see if he has a point or not. Life is too short.
Whatever I might agree or disagree with, this is annoying to look at, but his stuff keeps coming up in my YouTube feed. Even it looks slightly interesting, I know it will be some rant involved about a thing not related to technology, but some developer's personal opinions on non-tech ideas. I get it - people are horrible! Sheesh!
FWIW, probably not much, he said he had a Jewish background ... in, like, the one video I watched and eventually gave up on.
You can just mute/hide a channel from your feed permanently.
what's especially strange to me is that in the more distant past, he was a pretty normal guy - at least as normal as any other linux user. Heck, he had a super great podcast (Linux Action Show).
Something changed in the 2014ish time-frame when it got more and more politically extreme.
what do you think changed culturally around 2014 (I'd say it started a little earlier, maybe 2011)?
[dead]
* Are Jews the Cause of DEI in Big Tech?
...errrrrrrrrrrrrr, plot twist, he is a jew himself, or at least he claimed he is.
> NeoFetch But in Rust and More Gay
apt-install --fuck-yes gay-rust-neofetch
I’ll look to migrate to chimera shortly, but only if it includes gay neofetch.
Now I want to make a Woke Linux to drive this guy insane, the CoC alone will make his face melt
It won't do anything of the sort. It will allow him to make 200 videos complaining about it, get a load of ad-revenue and sell subscribestar memberships.
The best thing to do with people like Lunduke is ignore them.
Bonus points if you can make it non-binary.
you're at least 10 years too late with that idea
He's harassed people, including one of my friends.
And me.
Lunduke is a grifter and just generally a bit of an idiot.
e.g. I remember he once claimed Google was censoring him when he was de-listed from search, this was way back in 2009. His site had a malicious iframe because the PHP CMS he was using had been compromised.
His politics are kinda irrelevant to me. There are people who are Agorist/Libertarian/Conservative tech influencers online that do decent and informative content e.g. Sam Bent.
>Lunduke is a grifter...
And somehow you care so much you've created this account just to attack him.
I'd suggest going out for a walk.
Yes. I created the account because someone asked what the problem was with Lunduke and I had something to say. I've been aware of Lunduke for quite a while and he has always come off a clown.
The fact is that hasn't actually given much to the community and has been a drama, pretty much since his appearance in Linux land. People used to dislike him then and wanted him gone and this was well before the current culture war nonsense that is often seen on YouTube, Twitter and backwaters like Rumble.
> I'd suggest going out for a walk.
I go out for an hour walk in the countryside every lunch time. I am not sure what my exercise routine has got to do with criticising a long time troll and grifter.
nice, now open a web browser and any modern website /s
If you want to stay low on resources you can use Netsurf.
I just started it and loading my https://vermaden.wordpress.com/ page used 86 MB of RAM - a lot less then Firefox or Chrome.
It should run very quickly given that there are few things competing with it for resources like CPU cache space.
It's like "your car is going to get dirty why even wash it?"
I'm sure Lynx would be fine.
Been over a decade since I used a terminal browser, how do they handle modern websites with javascript?
They don't. Javascript is completely unsupported.
frogfind can be quite helpful.
[flagged]
Saying shill implies XLibre is some kind of scam he's trying to trick you into, where as, if you don't like it, you're absolutely free to not use it.
Or, what's the popular line in this scenario, "if you don't like it go make your own?"
I'm sorry.. I just find the reliable cult of software personality on HN to be a little frustrating.
Xlibre is not really a scam but it isn't much of a serious technical project either. I suppose you could call it a low-effort meme of the typical 4chan variety.
> it isn't much of a serious technical project either
How have you reached this qualification?
> I suppose you could call it a low-effort meme of the typical 4chan variety.
Time will tell. I think there's a lot of love for X11 that people offhandedly discount. I'm sure I'll end up using it shortly as I genuinely dislike wayland.
If you often click these types of articles and they all say XLibre, then it is probably the best tool for what you want to achieve.
For us not in the know, why is this bad?
Even ignoring all the politics and crazy stuff from the "maintainer"; most of their contributions were just shuffling code around and causing a lot of breakage. The typical "but this looks nicer so its better" type of programmer, not the type of code I'd rely on personally.
It's not, GNOME-Foundation aka Redhat aka IBM wants to "kill" X11, so they attach labels to discredit a project that wants to maintain X11:
https://en.ubunlog.com/The-XLibre-case-brings-out-the-worst-...
Same old story they try to repeat with DHH or the Framework Laptops because how dare you to support Omarchi:
https://world.hey.com/dhh/calling-someone-a-nazi-is-a-permis...
[flagged]
Please help me, Lunduke (who is a Jew) likes a project that is "nazi"?
Or is it maybe that people who call everyone a "nazi" who they disagree the real one's?
Stop degrading that word (to make common and without any weight) and use it for you crippled worldview.
These days, the label is being used to signal the labeled person is a target for violence[0].
Horrible.
0. https://world.hey.com/dhh/calling-someone-a-nazi-is-a-permis...
[flagged]
What's CoC in this context?
Code of Conduct
[flagged]
"Redskirts" hahaha you are so funny and clever. Let me say something you kinda guys like to hear: "Keep your social bullshit politics out of my tech stories"