[dev] Completeness suckless

[Sagar Acharya]

I recently wrote this article

https://designman.org/sagaracharya/blog/trusting_no_one

being absolutely unaware about suckless and this was brought to my attention.

Suckless's philosophy is hands down amazing and crucial wrt computer security. Although I'd like to point out 1 aspect. Why does suckless target very sophisticated users? If it shuns trying to go after elitist users, it can improve computer security of people all around the world and also themselves, since if others are secure, you yourself will become even more secure!

For it, there would be few requirements. Free software, minimal, easy to use, beautiful to look at (by default). I guess the latter 2 are lagging a bit.

Thanking you
--
Sagar Acharya
https://designman.org

P.S. Shifted completely to dwm this week. Can't even think of anything theoretically better than this!

[Страхиња Радић]

You can build software with the goal of user-friendliness in mind, but
this always comes with concessions. As long as you provide consistent
and easy to understand user-interfaces (APIs, Synopses, etc.), it is, I
think, the user's turn to try to understand them to the fullest, even
if it means learning a bit of a programming language.
How do the others see it?

I think that "user friendliness" carries complexity with it that is just being
concealed from the user. There are prejudices about what is "friendlier": having
a complex program that does a lot of things with little to no input, which also
means less control, or having a simple program that does one thing well, is
simple enough to understand and works well with other such simple programs. It
becomes apparent when there is some error or something unpredicted happens: it
is often difficult to diagnose the error and equally difficult, if not
impossible, to present a working solution for a "user friendly" (=complex)
program compared to a simple program. The solution in the case of a simple
program is exact and textual, easily transmitted in digital form or pronounced,
compared to the difficulty of having to describe where exactly, and how to click
in a GUI.

[Hadrien Lacour]

Post by Sagar Acharya
I recently wrote this article
https://designman.org/sagaracharya/blog/trusting_no_one
being absolutely unaware about suckless and this was brought to my attention.
Suckless's philosophy is hands down amazing and crucial wrt computer security. Although I'd like to point out 1 aspect. Why does suckless target very sophisticated users? If it shuns trying to go after elitist users, it can improve computer security of people all around the world and also themselves, since if others are secure, you yourself will become even more secure!
For it, there would be few requirements. Free software, minimal, easy to use, beautiful to look at (by default). I guess the latter 2 are lagging a bit.
Thanking you
--
Sagar Acharya
https://designman.org
P.S. Shifted completely to dwm this week. Can't even think of anything theoretically better than this!

Where do we stop, though? For me, sh (even with all its braindamage) is exactly
what suckless for the masses is, allowing for almost Lego-like ease of building
software as you want it. Actually, it is only the concept of pipes and line
oriented programs that provide that.

Security isn't the main point of suckless, it's only a consequence of clean and
simple code.
The main point is to empower people, but not through the spoon-feeding and
shoestring-tying the FSF likes so much, but by helping those who help
themselves.
That's how you get actual "computer freedom", by being able to program and
fullfill more and more of your needs/wants. suckless programs are made to
fit this mindset: simple enough to be modified and built around the UNIX
philosophy of use with other simple programs.

The people who want that simplicity but with more power (very hard thing to
create, actually), turn to Scheme or Tcl.

[Greg Reagle]

Post by Hadrien Lacour
Where do we stop, though? For me, sh (even with all its braindamage)

Speaking of the brain damage of sh, I highly recommend rc [1] which is available
in 9base [2] and other sources like Plan 9 Port.

[1] http://doc.cat-v.org/plan_9/4th_edition/papers/rc
[2] https://tools.suckless.org/9base/

[Hadrien Lacour]

Post by Greg Reagle

Post by Hadrien Lacour
Where do we stop, though? For me, sh (even with all its braindamage)

Speaking of the brain damage of sh, I highly recommend rc [1] which is available
in 9base [2] and other sources like Plan 9 Port.

I've found my calling with Tcl, personally. It's really what the shell should
have been, in my opinion, though it lacks the easy '|', which does live as
pipethread if it's really wanted.

[Sagar Acharya]

interesting article!

Thanks

I think, the user's turn to try to understand them to the fullest, even

if it means learning a bit of a programming language. - Laslo
Majority of users in this world would never learn a low level programming language like C. They are incapable.
It doesn't need to be complex. Few things make enormous impact like choosing good font, size of bars, color, connecting icons with programs, good battery and datetime displays. A simple 1 line script can run in the background. One would just take few important elements of User experience and add them here. I'm just talking wrt dwm here but I mean this at all levels. Few important UI elements should be taken care of.

Security isn't the main point of suckless, it's only a consequence of clean and simple code. - Hadrien

I guess this is where I diverge, user centric things always work better and have more power. One can always add a few more simple things, keeping minimalism of suckless intact. One can create dwmd (dwm for dumb) with few more features.

The main point is to empower people, but not through the spoon-feeding and shoestring-tying the FSF likes so much, but by helping those who help themselves. That's how you get actual "computer freedom", by being able to program and fullfill more and more of your needs/wants. suckless programs are made to fit this mindset: simple enough to be modified and built around the UNIX philosophy of use with other simple programs. - Hadrien

Majority of the people in this world are never going to bother with creating their own code. I have studied Engineering Physics at IIT Delhi. There, Computer Science guys are insane geniuses. They don't bother to set their system up. Very few do. Most use dual boot Ubuntu and Windows. Now that I have completed my Bachelor degree, I do not know a single person in my surroundings who uses GNU/Linux or BSD OSes. Using wm on this is way off! How would someone with no basic logical skills would do this? They won't.

I agree with Sagar - Edward

Thanks

Thanking you
Sagar Acharya
https://designman.org

[Hadrien Lacour]

Post by Sagar Acharya
I guess this is where I diverge, user centric things always work better and have more power.

More features != More power.

Post by Sagar Acharya
One can always add a few more simple things, keeping minimalism of suckless intact. One can create dwmd (dwm for dumb) with few more features.

Minimalism/power and catering to the lowest common denominator are simply and
completely incompatible. You've got to understand that interfaces for grandma
or Joe down the street and the ones for those who use dwm/bspwm can't be made
one. You have to be best at what you do, not a ridiculous jack-of-all-trades.

Post by Sagar Acharya
Majority of the people in this world are never going to bother with creating their own code.

Then they'll never be "free" while using something they will never bother
understanding.

Post by Sagar Acharya
I have studied Engineering Physics at IIT Delhi. There, Computer Science guys are insane geniuses. They don't bother to set their system up. Very few do. Most use dual boot Ubuntu and Windows. Now that I have completed my Bachelor degree, I do not know a single person in my surroundings who uses GNU/Linux or BSD OSes.

Which goes on to show that Computer Science has nothing to do with computers and
that scientists aren't necessarily the same as engineers/programmers,
nor are they wiser than average.
(I say that coming from one of the best universities in France; I didn't have
very much respect left for the word "scientist" when I left).

[Jeremy]

Post by Hadrien Lacour

Post by Sagar Acharya
I have studied Engineering Physics at IIT Delhi. There, Computer Science guys are insane geniuses. They don't bother to set their system up. Very few do. Most use dual boot Ubuntu and Windows. Now that I have completed my Bachelor degree, I do not know a single person in my surroundings who uses GNU/Linux or BSD OSes.

Which goes on to show that Computer Science has nothing to do with computers and
that scientists aren't necessarily the same as engineers/programmers,
nor are they wiser than average.
(I say that coming from one of the best universities in France; I didn't have
very much respect left for the word "scientist" when I left).

Do architects lay brick?

Jeremy

[Hadrien Lacour]

Post by Jeremy

Post by Hadrien Lacour

Post by Sagar Acharya
I have studied Engineering Physics at IIT Delhi. There, Computer Science guys are insane geniuses. They don't bother to set their system up. Very few do. Most use dual boot Ubuntu and Windows. Now that I have completed my Bachelor degree, I do not know a single person in my surroundings who uses GNU/Linux or BSD OSes.

Which goes on to show that Computer Science has nothing to do with computers and
that scientists aren't necessarily the same as engineers/programmers,
nor are they wiser than average.
(I say that coming from one of the best universities in France; I didn't have
very much respect left for the word "scientist" when I left).

Do architects lay brick?

No, but they better understand how bricks work. Unless "architect" just means
"designer", nowadays, and I'm just being out of touch with the times.

[Jeremy]

Post by Hadrien Lacour

Post by Jeremy

Post by Hadrien Lacour

Post by Sagar Acharya
I have studied Engineering Physics at IIT Delhi. There, Computer Science guys are insane geniuses. They don't bother to set their system up. Very few do. Most use dual boot Ubuntu and Windows. Now that I have completed my Bachelor degree, I do not know a single person in my surroundings who uses GNU/Linux or BSD OSes.

Which goes on to show that Computer Science has nothing to do with computers and
that scientists aren't necessarily the same as engineers/programmers,
nor are they wiser than average.
(I say that coming from one of the best universities in France; I didn't have
very much respect left for the word "scientist" when I left).

Do architects lay brick?

No, but they better understand how bricks work. Unless "architect" just means
"designer", nowadays, and I'm just being out of touch with the times.

I would think that the difference between a designer and an architect
is the level of detail involved.

The architect may consider power & water lines, the durability of the
structure, cost of the materials.

Offloading the work of configuring the desktop doesn't negate the
theoretical knowledge.

The practical knowledge of computer science comes from experience in a
consumer-driven, market environment. Acedamia is not that environment.

[Jeremy]

Post by Sagar Acharya

I think, the user's turn to try to understand them to the fullest, even

if it means learning a bit of a programming language. - Laslo
Majority of users in this world would never learn a low level programming language like C. They are incapable.

Where would you say their capacities are lacking?

Jeremy

[Mattias Andrée]

On Fri, 9 Apr 2021 12:13:23 -0700

Post by Jeremy

Post by Sagar Acharya

I think, the user's turn to try to understand them to the fullest, even

if it means learning a bit of a programming language. - Laslo
Majority of users in this world would never learn a low level programming language like C. They are incapable.

Where would you say their capacities are lacking?

I would say just determination or interest. Most people are capable of
learning programming (that is not to say most people are capable of
being good at it, but probably they can become good enough), and if
you can learn a high level language, you can definitely learn a low
level language.

Post by Jeremy
Jeremy

[Teodoro Santoni]

Post by Sagar Acharya
It doesn't need to be complex. Few things make enormous impact like choosing
good font, size of bars, color, connecting icons with programs, good battery
and datetime displays. A simple 1 line script can run in the background. One
would just take few important elements of User experience and add them here.
I'm just talking wrt dwm here but I mean this at all levels. Few important
UI elements should be taken care of.

dwm is just a window manager.

Post by Sagar Acharya
I guess this is where I diverge, user centric things always work better and
have more power.

Users come in many shapes and sizes, for example "hackers",
"programmers", and "clients". Nobody who at least a bit care for
his/her mental hygiene should start coding anything to build a project
aimed for the general public or "to succeed" in some software arms
race, bringing his/her software to fit any present or future use-case.
I think that wouldn't be a good idea even in exchange for a really
bountiful paycheck.
It's because you end with dealing with power dynamics between people,
concede, fight, persuade, befriend, you may still bring out of this
diplomacy game the perfect software, but most of those who partake end
up poisoned, producing worse software and/or disliking more and more
the production of software.

Post by Sagar Acharya
One can always add a few more simple things, keeping
minimalism of suckless intact. One can create dwmd (dwm for dumb) with few
more features.

A lot of people had a take for a user-friendlier dwm, you should try
those, or similar wms with "battery pack included" like i3 or awesome.
After that, if "the market" hasn't anything for you, fire up the
editor and hack on dwm.
Don't take this as a "go away": what i'm suggesting for you is to
understand that this graphical interface is twenty-something years old
(counting wmii), so it would be really hard to come up with a really
novel and miraculous concept upon dwm.

Post by Sagar Acharya
Majority of the people in this world are never going to bother with creating
their own code. I have studied Engineering Physics at IIT Delhi. There,
Computer Science guys are insane geniuses. They don't bother to set their
system up. Very few do. Most use dual boot Ubuntu and Windows. Now that I
have completed my Bachelor degree, I do not know a single person in my
surroundings who uses GNU/Linux or BSD OSes. Using wm on this is way off!
How would someone with no basic logical skills would do this? They won't.

That's what windows, mac os, or a FLOSS OS with
GNOME/Enlightenment/Elementary are for. Conquer any heart in this
world for dwm + shell scripts is of no use.

[Sagar Acharya]

I say that coming from one of the best universities in France; I didn't have very much respect left for the word "scientist" when I left

Same here. Exactly same! But they for sure are intelligent. One of the key things in capitalism is getting tamed by the powerful. Majority of these geniuses don't know how their work impacts the world! They get tamed by people who work against their own principles just because knowing their tamer's principles is difficult!

Now try bspwm

Damn that's good! Very simple logic. But I'll stick to dwm for now. Check it after a while.

Where would you say their capacities are lacking?

Most don't have the logical abilities. Amongst ones who do, many don't want to read in times of videos. They want instant gratification. For few it is also about power. You suggest them to use dwm. They'll think they're above you socially and won't use it. This maybe a bit biased but 1 thing I know for sure, that people are far, very far away from using minimal systems.

The practical knowledge of computer science comes from experience in a consumer-driven, market environment. Academia is not that environment.

+1.

A lot of people had a take for a user-friendlier dwm, you should try those, or similar wms with "battery pack included" like i3 or awesome.

Again, I'm not talking about dwm specifically. i3wm is a single wm and while it's good, I think suckless can target COMPLETE systems. It's philosophy is great.

We believe this should become the mainstream philosophy in the IT sector. - On suckless site.

This IT sector is constantly making things easier for people or atleast making people believe it is. If suckless wants this to become mainstream philosophy, it has to do itself. People don't wanna change especially when computers are just the tools to achieve some end.

If suckless makes it easy for people to use minimalistic systems, they will. I shifted from complete windows to dual boot to complete GNU/Linux (Parrot Linux) to Parabola (Plasma) to Parabola (dwm) in about 3 years. I can't imagine shifting directly to current state. I would get frustrated and go back to Windows if someone would ask me to get to this point from complete Windows. But if there were tutorials explaning why would you use wm instead of de, wm were a bit more beautiful, things would be much simpler. Plasma is hands down beautiful! But it's very complex!

Thanking you
Sagar Acharya
https://designman.org

P.S. I'd love to contribute.

[Hadrien Lacour]

Post by Sagar Acharya
If suckless wants this to become mainstream philosophy

That's where the fallacy is. I suppose suckless devs aren't naive and simply
know that people aren't the same (dare I say "not equal"), so they simply make
software for themselves and people like themselves.

[Hiltjo Posthuma]

Post by Sagar Acharya

I say that coming from one of the best universities in France; I didn't have very much respect left for the word "scientist" when I left

Same here. Exactly same! But they for sure are intelligent. One of the key things in capitalism is getting tamed by the powerful. Majority of these geniuses don't know how their work impacts the world! They get tamed by people who work against their own principles just because knowing their tamer's principles is difficult!

Now try bspwm

Damn that's good! Very simple logic. But I'll stick to dwm for now. Check it after a while.

Where would you say their capacities are lacking?

Most don't have the logical abilities. Amongst ones who do, many don't want to read in times of videos. They want instant gratification. For few it is also about power. You suggest them to use dwm. They'll think they're above you socially and won't use it. This maybe a bit biased but 1 thing I know for sure, that people are far, very far away from using minimal systems.

The practical knowledge of computer science comes from experience in a consumer-driven, market environment. Academia is not that environment.

+1.

A lot of people had a take for a user-friendlier dwm, you should try those, or similar wms with "battery pack included" like i3 or awesome.

Again, I'm not talking about dwm specifically. i3wm is a single wm and while it's good, I think suckless can target COMPLETE systems. It's philosophy is great.

We believe this should become the mainstream philosophy in the IT sector. - On suckless site.

This IT sector is constantly making things easier for people or atleast making people believe it is. If suckless wants this to become mainstream philosophy, it has to do itself. People don't wanna change especially when computers are just the tools to achieve some end.
If suckless makes it easy for people to use minimalistic systems, they will. I shifted from complete windows to dual boot to complete GNU/Linux (Parrot Linux) to Parabola (Plasma) to Parabola (dwm) in about 3 years. I can't imagine shifting directly to current state. I would get frustrated and go back to Windows if someone would ask me to get to this point from complete Windows. But if there were tutorials explaning why would you use wm instead of de, wm were a bit more beautiful, things would be much simpler. Plasma is hands down beautiful! But it's very complex!
Thanking you
Sagar Acharya
https://designman.org
P.S. I'd love to contribute.

Whats stopping you from doing it?

[Sagar Acharya]

Whats stopping you from doing it? - Hiltjo

Can't do it alone! It's a huge task. I can contribute somewhat. I'm currently working on a thing which is much more important wrt computer security! But I gotta do it. Point taken.

Sagar

[Edward Willis]

Hello. I am new also to this mailing list, although I have read suckless.org and cat-v.org for a long time.

I agree with Sagar. Simple software allows for people to use their computers for longer, which saves on waste, saves money, and is more inclusive of users who don’t have the latest and greatest technology. Libraries and other public institutions that provide computer access are especially hard pressed to keep up with increasing hardware requirements.

Cheers,
Edward

PS: As this is my first message on this list please inform me if I am doing anything wrong in the formatting or sending of this message.

On Fri, 9 Apr 2021 14:54:31 +0200 (CEST)
Dear Sagar,

Post by Sagar Acharya
I recently wrote this article
https://designman.org/sagaracharya/blog/trusting_no_one
being absolutely unaware about suckless and this was brought to my attention.

interesting article!

Post by Sagar Acharya
Suckless's philosophy is hands down amazing and crucial wrt computer
security. Although I'd like to point out 1 aspect. Why does suckless
target very sophisticated users? If it shuns trying to go after
elitist users, it can improve computer security of people all around
the world and also themselves, since if others are secure, you
yourself will become even more secure!
For it, there would be few requirements. Free software, minimal, easy
to use, beautiful to look at (by default). I guess the latter 2 are
lagging a bit.

I understand your point completely and it has been part of many debates
in the last few years. I used to be more open towards "more
user-friendliness" a few years ago, but realized that if you look at
modern users, most of them are unwilling to invest time and effort into
getting something to work and expect too much from others in terms of
support and guidance.
Initiatives to "open up" such things are noble in their motivations,
but actually make things worse.
You can build software with the goal of user-friendliness in mind, but
this always comes with concessions. As long as you provide consistent
and easy to understand user-interfaces (APIs, Synopses, etc.), it is, I
think, the user's turn to try to understand them to the fullest, even
if it means learning a bit of a programming language.
How do the others see it?
With best regards
Laslo

[Hadrien Lacour]

Post by Sagar Acharya
P.S. Shifted completely to dwm this week. Can't even think of anything theoretically better than this!

Now, try bspwm :)

[Daniel Cegiełka]

Choose an OS which has a small kernel and other minimal software
that you need. (OpenBSD is the only one I can see here today)

suckless is strongly Linux-oriented - musl, ubase, smdev, nldev, nlmon

I wonder if it would make sense to make a little Linux kernel.
Everything that is unnecessary could be deleted. And yes, this
sounds controversial, but for example:

https://github.com/michaelforney/linux-headers/tree/master/scripts

Daniel

I recently wrote this article
https://designman.org/sagaracharya/blog/trusting_no_one
being absolutely unaware about suckless and this was brought to my attention.
Suckless's philosophy is hands down amazing and crucial wrt computer security. Although I'd like to point out 1 aspect. Why does suckless target very sophisticated users? If it shuns trying to go after elitist users, it can improve computer security of people all around the world and also themselves, since if others are secure, you yourself will become even more secure!
For it, there would be few requirements. Free software, minimal, easy to use, beautiful to look at (by default). I guess the latter 2 are lagging a bit.
Thanking you
--
Sagar Acharya
https://designman.org
P.S. Shifted completely to dwm this week. Can't even think of anything theoretically better than this!

[Hiltjo Posthuma]

Post by Daniel Cegiełka

Choose an OS which has a small kernel and other minimal software
that you need. (OpenBSD is the only one I can see here today)

suckless is strongly Linux-oriented - musl, ubase, smdev, nldev, nlmon

No, there are also many other tools (dwm, dmenu, st) etc which are not
Linux-oriented. The userbase is probably mostly Linux-oriented, but that's
something different.

musl is a nice project (imho) but not part of suckless.org itself.

Post by Daniel Cegiełka
I wonder if it would make sense to make a little Linux kernel.
Everything that is unnecessary could be deleted. And yes, this
https://github.com/michaelforney/linux-headers/tree/master/scripts
Daniel

I recently wrote this article
https://designman.org/sagaracharya/blog/trusting_no_one
being absolutely unaware about suckless and this was brought to my attention.
Suckless's philosophy is hands down amazing and crucial wrt computer security. Although I'd like to point out 1 aspect. Why does suckless target very sophisticated users? If it shuns trying to go after elitist users, it can improve computer security of people all around the world and also themselves, since if others are secure, you yourself will become even more secure!
For it, there would be few requirements. Free software, minimal, easy to use, beautiful to look at (by default). I guess the latter 2 are lagging a bit.
Thanking you
--
Sagar Acharya
https://designman.org
P.S. Shifted completely to dwm this week. Can't even think of anything theoretically better than this!

[Sagar Acharya]

I too agree with Hiltjo. I don't think suckless is Linux oriented.

At the time of writing that article, I thought Linux and BSD are the only 2 usable kernels. Practically it seems so, but recently I saw microkernel vs monolithic kernel debate and got interested in seL4. I'm not sure whether it's usable yet or not.

wrt unnecessary things being deleted, Linux has more than 1.6M lines so think about for how much time you'd have to think what's necessary and what not? (writing the script which you pointed)

Thanking you
Sagar Acharya
https://designman.org

[Edward Willis]

Post by Daniel Cegiełka

Choose an OS which has a small kernel and other minimal software
that you need. (OpenBSD is the only one I can see here today)

suckless is strongly Linux-oriented - musl, ubase, smdev, nldev, nlmon
I wonder if it would make sense to make a little Linux kernel.
Everything that is unnecessary could be deleted. And yes, this
https://github.com/michaelforney/linux-headers/tree/master/scripts
Daniel

I recently wrote this article
https://designman.org/sagaracharya/blog/trusting_no_one
being absolutely unaware about suckless and this was brought to my attention.
Suckless's philosophy is hands down amazing and crucial wrt computer security. Although I'd like to point out 1 aspect. Why does suckless target very sophisticated users? If it shuns trying to go after elitist users, it can improve computer security of people all around the world and also themselves, since if others are secure, you yourself will become even more secure!
For it, there would be few requirements. Free software, minimal, easy to use, beautiful to look at (by default). I guess the latter 2 are lagging a bit.
Thanking you
--
Sagar Acharya
https://designman.org
P.S. Shifted completely to dwm this week. Can't even think of anything theoretically better than this!

I don’t think it makes sense for the suckless guys to try trimming down that bloated mess (Linux kernel).

To be honest I’m wondering if the love they give OpenBSD, as a desktop OS, is misplaced. OpenBSD is 22M lines if you include the entirety of the files, and 16M counting just the lines with code on them. Granted I think that number is across the whole base distribution which makes it smaller than the Linux kernel alone. But OpenBSD is still REALLY fat compared to operating systems of the past. FreeDOS is ~50k lines, and then there are things like KolibriOS to consider.

I don’t know if you guys have seen this, but I’d like to share a video from YT that I think is appropriate for this group:

watch?v=kZRE7HIO3vk

Cheers,
Edward

[Sagar Acharya]

Post by Edward Willis
I don’t think it makes sense for the suckless guys to try trimming down that bloated mess (Linux kernel).
To be honest I’m wondering if the love they give OpenBSD, as a desktop OS, is misplaced. OpenBSD is 22M lines if you include the entirety of the files, and 16M counting just the lines with code on them. Granted I think that number is across the whole base distribution which makes it smaller than the Linux kernel alone. But OpenBSD is still REALLY fat compared to operating systems of the past. FreeDOS is ~50k lines, and then there are things like KolibriOS to consider.
watch?v=kZRE7HIO3vk

That video is pure gold!

I'll revise my article wrt OpenBSD but it's the least evil when seen wrt usability. I doubt it's that gigantic though! How do you visualize the solution? Writing from scratch targeting an embedded board?

Thanks
Sagar Acharya

[Daniel Cegiełka]

Post by Edward Willis
I don’t think it makes sense for the suckless guys to try trimming down that bloated mess (Linux kernel).
To be honest I’m wondering if the love they give OpenBSD, as a desktop OS, is misplaced. OpenBSD is 22M lines if you include the entirety of the files, and 16M counting just the lines with code on them. Granted I think that number is across the whole base distribution which makes it smaller than the Linux kernel alone.

Linux:

drivers: > 700 MB
arch: > 135 MB
Docs: > 50 MB
tools etc: > 42 MB
fs: > 42 MB
kernel? 11 MB!

What is your arch? x86? It's <15 MB. arm64? 13 MB.

32 MB its 32bit arm.. Android? Can be deleted.

The only archs we may be interested in are x86(_64), arm64 and
power64. The rest is for cross-compilations.

Docs? Can be deleted. fs? ext4 is ~1.7 MB. Do you need heavy file
systems (xfs, btrfs)? If not, can be deleted.

Drivers. net > 100 MB. But do you need drivers for expensive
low-latency server network cards on your laptop or home workstation?
You can always compile such a driver outside of the kernel.

gpu >250 MB. In many cases it can be deleted as well.

So if you only choose what a typical user may actually be interested
in, such a striped-down kernel may be less than 100-150 MB or ~10-15
MB as gz or xz.

Updates? It is not that difficult. The rule is simple: if it doesn't
touch the remaining code, skip it.

But OpenBSD is still REALLY fat

If I remember correctly, version 3.8 was the last one with the minimal
option. Later it was only fatter and fatter. OpenBSD is no longer a
lightweight solution. It's probably easier to configure Linux kernel
than to do a lightweight OpenBSD installation (perl for pkg,
clang/llvm etc). See sta.li/morpheus or Oasis Linux.

compared to operating systems of the past. FreeDOS is ~50k lines, and then there are things like KolibriOS to consider.

...and Haiku, Plan 9/9front etc. etc. Does it have drivers for modern hardware?

Daniel

Post by Edward Willis
watch?v=kZRE7HIO3vk

That video is pure gold!
I'll revise my article wrt OpenBSD but it's the least evil when seen wrt usability. I doubt it's that gigantic though! How do you visualize the solution? Writing from scratch targeting an embedded board?
Thanks
Sagar Acharya

[Cág]

I used to be more open towards "more user-friendliness" a few years ago,
but realized that if you look at modern users, most of them are
unwilling to invest time and effort into getting something to work and
expect too much from others in terms of support and guidance.
Initiatives to "open up" such things are noble in their motivations,
but actually make things worse.

Spot on. For me "user-friendly" means "straightforward once learned".
Unix is, as such, user-friendly. suckless too.