Minimalism
- Inicie sesión ou rexístrese para enviar comentarios
(this post has been written by someone who was a hardcore minimalist and who used Hyperbola in the past but is now content with "non-minimalist" tools such as Emacs and systemd)
Least minimalist Hyperbola user: has no graphical environment, does everything through command-line, uninstalls the package manager because it's unnecessary bloat, no dbus, no systemd, no background processes running, asynchronous execution of processes is disabled because it is unnecessary bloat, programs are compiled directly as part of the shell because having separate programs is bloat, no multitasking, no zstd because it's corporate Faceb00k bloat, laptop levitates due to lightweightness in operating system,
Most minimalist Trisquel user: computer constantly swapping, installs 300 extra Emacs Lisp packages into Emacs because default Emacs isn't a feature-rich-enough operating system, 99% of the system packages are provided by systemd, user never has to get out of bed because computer has become sentient and does the user's laundry, user unironically uses a Desktop Environment rather than a window manager, switches away from LXDE after learning that KDE is more bloated, has both KDE and GNOME installed so that perse can have all the programs from both desktop environments, uses algorithm to automatically upscale all images viewed in the web browser to 4K, smoke comes out of computer and computer is used for warmth in winter
I still use Hyperbola on a few of my computers. I have four main ones, one of which is not in great shape that uses Hyperbola.
My t430 which kind of feels like its not as sturdy as it used to be.
And my X230 which I still use pretty regularly.
The other two each use a different distro, one is devuan and one debian.
I admit, systemd isn't as shitty as I thought. It still makes me nervous due to the code size and the concept of doing one thing and doing it right, sounding more sane to me. but I am surprised how fast it works on ARM64 devices even in boot.
I guess x86 is more borked than ARM64 with systemd.
Like literally, it booted in 2 seconds. Although, it could also be that I had no full disk encryption minus boot. I prefer to do this usually.
Btw, you don't need to be that minimalist on Hyperbola.
I use JWM, JWMKIT and that's frankly good enough for me.
I'm working on a new version of the Libre-antiX respin. The system boots into tty using just 16.4MB of memory, and loads a DWM instance using just 98MB of memory.
Not quite as low as the 76MB of memory we were seeing with Hyperbola and JWM a couple of years ago. But a lot less than Libre-antiX used to use for DWM, which I think the minimum was about 144MB back then. Might have to do with the fact that Libre-antiX no longer uses elogind, but uses seatd now instead.
To boot Libre-antiX into JWM now takes more - 104MB. Not sure why - JWM used to use less memory than DWM. Has JWM gotten bloated the last few years?
I did a net-install of Trisquel, and with minimal xserver-xorg and dwm I'm using 161mb, so not bad at all. Various systemd processes are using up 89mb of that amount, so with a lighter init manager you could instantly get Trisquel down near 100mb with a window manager. If I could just get rid of systemd-journald and replace it with the functionality of rsyslog I would save 22.5mb and be very happy, but unfortunately no one that I can find has written a full how-to of doing that.
Update: Oh my goodness, I did it! (sort of)
I followed this how-to from RedHat - " Configuring system logging without journald ": https://access.redhat.com/articles/4058681
I've got my minimal Trisquel net-install with dwm using rsyslog instead of systemd-journald. journald is still a process using 1.8mb in the background, but now I'm saving over 20mb! The system booted up using 68mb of memory, and I'm logged into dwm using 138mb.
Progress!!! I might write a how-to today.
Edit: useful background info from the rsyslog project is here: https://rsyslog.adiscon.narkive.com/Xggqr6TJ/can-replace-and-eliminate-journald
I am a translator!
Is your goal to make a GNU/Linux distro work on hardware that has very limited memory? Any specific device in mind? Also, would it be feasible to load the entire file system in memory and then run without any disk?
>"Is your goal to make a GNU/Linux distro work on hardware that has very limited memory?"
Yes, but also to learn how to control systemd in general.
>"Any specific device in mind?"
Not any one in particular, although I do have a rather large collection of old systems. But I would also like to exert a great deal more control over systemd on the newer systems.
>"Also, would it be feasible to load the entire file system in memory and then run without any disk?"
I know how to do this easily with Libre-antiX. I don't know the procedure for Trisquel, although it's probably some simple combination of config hacks or commands.
@Kiki_the_Cyber_Squarrel "... smoke comes out of computer and computer is used for warmth in winter" that sure made me laugh, I appreciate your humour!
(Myself am not good with tech at all, have not figured out yet how to install Hyperbola.)
To you and the OP, this guide worked for me:
There is also this guide, I think, that worked for me.
https://wiki.hyperbola.info/doku.php?id=en:manual:beginner_installation
Also testing installing in a virtual machine before real hardware may help.
Something like
qemu-img create -f qcow2 hyperbola.qcow2 40G
To set how much space to use on your system for the virtual machine install.
qemu-system-x86_64 -hda hyperbola.qcow2 -boot d -cdrom hyperbola-milky-way-v0.4.4-dual.iso -m 2048 -smp 2 -enable-kvm -nic user,hostfwd=tcp::10037-:4571
The qcow2 is where the virtual machine name file may be. It does not need to be named hyperbola.qcow2.
hyperbola-milky-way-v0.4.4-dual.iso is the iso file you are testing, it could also be something like trisquel_11.0_amd64.iso
or any other iso file I think.
man qemu can help you see more about qemu, I think ,hostfwd=tcp::10037-:4571 had something to do with the internet, but it has been some time since I used this or installed hyperbola.
https://wiki.hyperbola.info/doku.php?id=en:system:display_servers:xenocara
and
https://wiki.hyperbola.info/doku.php?id=en:system:userspace:desktop_enviroment:lumina_desktop
can help with setting up a desktop or windows manager.
I think sound and touch pad are partly setup with a program install and config file, but I do not remember the setup as well as when I first did it.
qemu-system-x86_64 -hda hyperbola.qcow2 -boot d -m 2048 -smp 2 -enable-kvm -nic user,hostfwd=tcp::10037-:4571
Can help you boot it after the virtual machine install.
https://wiki.hyperbola.info/doku.php?id=en:system:init_systems:openrc
I think is about an init system so that may also help.
You may be doing some things like differently with Trisquel.
Trisquel has a graphical off-line installer, does not need me to set up a "config" file for sound or touch pad, and by default (I think) has a desktop that is set up by the graphical off-line installer, but setting up these (or close to these) things may not be too hard in Hyperbola. Plus if you did not wish to have those it could be faster a setup.
I think pmount or something like that name may help with mounting things.
So testing both with Qemu could help you get used to Qemu and setting up things in a virtual machine.
The https://wiki.hyperbola.info/
could help with other things.
rm -rf ~/.local/share/Trash/*
Can help delete trash and
cat /dev/null >~/.bash_history && history -c && exit
can help clear bash history.
It may help on both Trisquel and Hyperbola as well as other distros, but I think bleachbit also can be used to do that graphically on Trisquel.
https://www.hyperbola.info/packages/?sort=&q=trash
Also has some command line programs at this time for trash like trash-cli and trashman could help.
I mostly use the above rm command to clean out trash on Hyperbola and the trash icon in the Mate desktop, whatever that is called if it is not part of Mate.
Care should be done when cleaning bash history and the trash in-case you need to remember bash history or deleted something by mistake so you do not lose things you wish to keep or get back to a comman in bash you use frequently.
I am a member!
>smoke comes out of computer
That happened to mine once. While it was running, the place where the power cable goes into the tower started sparking, hissing and smoking when – in another room of my flat! – I plugged in an electric device.
According to an electrician who had a look at this flat's power lines because of an unrelated problem, my lines or outlets have a weird layout, something about half of them being in a single chain instead of in separate branches I think. I guess that's what also caused the smoke event.
I'm glad that quickly unplugging the computer stopped the sparking and it didn't cause a fire in my flat, but after that, the desktop computer would no longer start. I had to use a half-broken laptop as replacement, then another one, and when all of them had become unusable, I tried to buy a new computer at a time when that was difficult because of a pandemic-induced spike in the demand for remote-working hardware, exacerbating any pre-existing supply issues.
Hyperbola has very minimum requirements though so I very much doubt you would have that problem unless you were using ultra ancient hardware.
I love messing about with minimalism. The current cutoff/frontier, as far as hardware goes, I feel is somewhere around the Core2Duo generation of gear. Anything after that, and you can not really tell a huge difference (relatively). For example, if you have an i5-2XXX Sandy Bridge with 8GB RAM and a SSD vs an i5-10XXX with 8GB RAM and a SSD, You are not going to see much difference in daily tasks on Linux Lite vs. Pop_OS.
LXLE is now dead. It used to be the fastest (what I would call "complete & polished") distro for old hardware.
Anyone else mucking about with musl, busybox, libc derivatives?
I am hoping for a Wayland "LXDE-esque" DE one of these days...
Why did LXLE die anyhow?
Part 1 Complexity_vs_simplisity what is blob_complexity/non-free_complexity.
If there is a binary blob distro that is fully licensed under CC0/Public domain or a permissive license that is in a small iso file and another open source (the freely licensed type) distro AGPL3+ or other free license including CC0 that also has a large and high complexity code base what would be more easy to understand? The small freely licensed blob or the freely licensed large complex code base?
Even small things may be hard to reverse engineer, but a large free_as_in_freedom_licensed code base could also take time to understand.
I saw that rnetclient is the outcome of reverse-engineering ReceitaNet, the official and proprietary software that Receita Federal distributes. Also Gnu was started by making freedom replacements of other things, like how the Gnu operating system started than included the Linux kernel some time after.
https://www.gnu.org/gnu/gnu.html
So reverse engineering can be done, but is likely thought of as hard by most people.
Free_as_in_freedom_licensed does not (likely) need to have reverse engineering done but still takes some time to understand.
So if all copyright, patent, trademark, and even licenses were gone (or not taken into consideration), what amount of complexity and amount of free source code becomes the complexity of a small binary blob or becomes non-free because it is as hard to understand or change as a binary blob?
I type about this so that I (and likely others) do not wish to make obscured "source code" by mistake by having too much complexity that makes obscured "source code" and than by doing that break the GPL3+ or/and the AGPL3+.
Though I do not know how much complexity may = "obscured source code is not source code".
Part 2 I'm glad Freedom software exists.
I'm glad that there is fully free as in freedom software, both for simple and complex things, and that reverse engineering does not need to be done.
Finding out when or if "binary_blob_complexity = large_complex_source_code_base" could help show if "simple code base vs complex code base" has anything to do with free software or not.
Also for the purpose of compliance with the GPL3+/AGPL3+ to provide real/not_obscured source code (I think) and for helping with augmented versions as more complexity and/or monolithic and/or augmented code may have more things needing to be linked in a way to GPLed code and its license.
If "binary_blob_complexity = large_complex_source_code_base" would that mean if a binary blob is non-free because it is hard to understand/change than a large_complex_source_code_base is also non-free?
Or if a large_complex_source_code_base is free than would that make some small binary blobs free if under a free license as the small binary blobs may be as "hard to understand/change" as the large code base?
I think that may be how the large_complex_non-simple_code_base_may_be_like_binary_blobs discussions may have come about.
I'm glad there are both simple as TempleOS (even though there are many other problems with TempleOS) free software (as much as I know for TempleOS, I did not check all it's code base)
https://trisquel.info/en/forum/i-believe-copyleft-harmful-free-culture-movement#comment-176233
as well more complex software (I think) like any of the
https://www.gnu.org/distros/free-distros.html
https://www.gnu.org/distros/free-non-gnu-distros.html
So simple fully free licensed free_as_in_freedom_licensed software and complex fully free licensed free_as_in_freedom_licensed software may both be nice for some things.
And with both simple (free_as_in_freedom_licensed blob free) software and complex (free_as_in_freedom_licensed blob free) software, maybe each person can chose what is the right software for themselves.
I think copyleft/share-alike may help both free as in freedom software and free-culture, though if anyone does not like/not_think_is_free/think_is_not_simple_enough
there is the (I hope it is free_as_in_freedom_licensed software even though it has many problems) TempleOS.
https://trisquel.info/en/forum/i-believe-copyleft-harmful-free-culture-movement#comment-176233
I have both Hyperbola and Trisquel on my computer at this time, as I like some things from both. Hyperbola has a relatively small code base and Trisquel has a graphical off-line installer, does not need me to set up a "config" file for sound or touch pad, and by default (I think) has a desktop that is set up by the graphical off-line installer.
The link in the other forum post was also saved.
https://web.archive.org/web/20240310182319/https://forums.hyperbola.info/viewtopic.php?id=995
Thank you all (everyone doing anything free_as_in_freedom_licensed like software or otherwise) for all your help. I hope everyone can find software that works well for them.
https://trisquel.info/files/king-james-bible-pure-cambridge-edition-pdf_archive.torrent
I did not check if all forks of TempleOS are CC0/Public domain, though I think TempleOS is in the Public Domain/Free, though it may have many problems in it.
So too simple may have problems, but maybe it could work well/better for some people.
https://www.gnu.org/philosophy/when-free-software-isnt-practically-superior.html
https://github.com/cia-foundation/TempleOS
https://github.com/Zeal-Operating-System/ZealOS
https://github.com/tinkeros/TinkerOS
https://github.com/minexew/Shrine
Whatever people believe at least none of the freedom distros and their communities
https://www.gnu.org/distros/free-distros.html
https://www.gnu.org/distros/free-non-gnu-distros.html
are like the proprietary/non-free and partly/or_fully closed-source Red Star OS and its development or other non-free software.
https://en.wikipedia.org/wiki/Red_Star_OS
https://www.gnu.org/proprietary/proprietary.html
I did not know if complexity_is_helpful_sometimes vs complexity_is_obscured_sometimes was about free software or another thing so that is why this is in the troll area and not the free software area.
I saw the topic "Minimalism" so I thought to post it here instead of make a "How much complexity is obscurity?" topic. As this may be about "complexity vs minimalism" or "how much complexity or minimalism is still free and helpful."
I think the complexity_is_helpful_sometimes vs complexity_is_obscured_sometimes could be close to the
https://en.wikipedia.org/wiki/Tanenbaum%E2%80%93Torvalds_debate
As that had something like the monolithic_kernels_are_helpful_sometimes vs microkernels_are_helpful_sometimes debate. Though I did not research that debate yet.
The free software definition does not deal with complexity. At all.
You seem to confuse "complex" with "obfuscated". As the free software definition claims:
Obfuscated “source code” is not real source code and does not count as source code.
https://www.gnu.org/philosophy/free-sw.en.html#make-changes
The point of obfuscated software is that it was not written that way. It is the output of a transformation of the actual source code (think of minimized JavaScript), which must be available for the software to possibly be free.
Thank you for that information and clarification about obfuscated code, Magic Banana.
Thank you for that information and clarification about callable and invoking, jxself.
Thank you everyone else for all the help provided before.
I still have a large amount to learn about both programing as well as law, but I'm glad others know much more about both than I do.
Also the four freedoms
https://www.gnu.org/philosophy/free-sw.html#four-freedoms
do show
The freedom to run the program as you wish, for any purpose (freedom 0).
So if someone wished to make a complex or simple program and not an obfuscated program it would block freedom 0 to say only simple or only complex programs can be compiled. Both simple and complex programs can be done to preserve freedom.
Also what may be simple or complex real source code may look complex to someone else.
Like when I did not even know about header files well enough to know about what was in darknet or how header files even work.
https://trisquel.info/en/forum/edge-cases-free-software#comment-177352
Or about like when I did not even know about callable and invoking.
modified versions must still be callable
Or how the Gnu software was started, like invoking gcc as cc I did not know that (invoking or callable) could be done and still be free. I see a little more of how the Gnu system/software was started.
No one's claiming gcc is cc, but it can still be run that way for compatibility with programs that expect the C compiler to be called cc.
I did not know some or any freedom suporting programs expect the C compiler to be called cc ("The original (proprietary) UNIX C compiler") and that those programs could still be called libre software.
I thought all libre software could only "expect the C compiler" to be called as one of the free/libre compilers.
I thought libre always needed to show/call/invoke libre things or those would be suggesting/showing/calling/invoking non-free things.
I also made that invoking mistake before, likely. I did not know about invoking also.
https://trisquel.info/en/forum/edge-cases-free-software#comment-177357
The freedom to study how the program works, and change it so it does your computing as you wish (freedom 1). Access to the source code is a precondition for this.
"As you wish" could also mean making a complex program or distro, or a simple program or distro, so if anyone wishes to have simple or complex program or distro they can do that.
As well as making simple programs have > or | in scripts to send things around to do complex things. I.E. invoke the expected names I think.
The freedom to redistribute copies so you can help others (freedom 2).
People can redistribute both complex and simple programs and operating systems. Freedom 2.
The freedom to distribute copies of your modified versions to others (freedom 3). By doing this you can give the whole community a chance to benefit from your changes. Access to the source code is a precondition for this.
People can redistribute both complex and simple programs and operating systems and edit both to make complex ones more simple and simple ones more complex.
And I'm glad people found ways to make both simple and complex fully free/libre distros, because I did not know about all the technical and legal ways that that could be done.
Like if any name could be used for (invoking and callable) things to make free/libre replacements for non-free software to replace a whole non-free operating system. But libre supporters did know both the legal and technical ways to do so.