Not to disrespect this, but it used to be entirely normal to have a GUI environment on a machine with 2MB of RAM and a 40MB disk.

Or 128K of ram and 400 kb disk for that matter.

A single 1920x1080 framebuffer (which is a low resolution monitor in 2025 IMO) is 2MB. Add any compositing into the mix for multi window displays and it literally doesn’t fit in memory.

I had a 386 PC with 4MB of RAM when I was a kid, and it ran Windows 3.1 with a GUI, but that also had a VGA display at 640x480, and only 16-bit color (4 bits per pixel). So 153,600 bytes for the frame buffer.

640 * 480 / 2 = 150KB for a classic 16-color VGA screen.

The Amiga 500 had high res graphics (or high color graphics … but not on the same scanline), multitasking, 15 bit sound (with a lot of work - the hardware had 4 channels of 8 bit DACs but a 6-bit volume, so …)

In 1985, and with 512K of RAM. It was very usable for work.

a 320x200 6bit color depth wasn't exactly a pleasure to use. I think the games could double the res in certain mode (was it called 13h?)

For OCS/ECS hardware 2bit HiRes - 640x256 or 640x200 depending on region - was default resolution for OS, and you could add interlacing or up color depth to 3 and 4 bit at cost of response lag; starting with OS2.0 the resolution setting was basically limited by chip memory and what your output device could actually display. I got my 1200 to display crisp 1440x550 on my LCD by just sliding screen parameters to max on default display driver.

Games used either 320h or 640h resolutions, 4 bit or fake 5 bit known as HalfBrite, because it was basically 4 bit with the other 16 colors being same but half brightness. The fabled 12-bit HAM mode was also used, even in some games, even for interactive content, but it wasn't too often.

It's so much fun working with systems with more pixels than ram though. Manually interleaving interrupts. What joy.

Do you really need the framebuffer in RAM? Wouldn't that be entirely in the GPU RAM?

To put it in GPU RAM, you need GPU drivers.

For example, NVIDIA GPU drivers are typically around 800M-1.5G.

That math actually goes wildly in the opposite direction for an optimization argument.

Doesn't the UEFI firmware map a GPU framebuffer into the main address space "for free" so you can easily poke raw pixels over the bus? Then again the UEFI FB is only single-buffered, so if you rely on that in lieu of full-fat GPU drivers then you'd probably want to layer some CPU framebuffers on top anyway.

Yes if you have UEFI.

well, if you poke framebuffer pixels directly you might as well do scanline racing.

Alas, I don't think UEFI exposes vblank/hblank interrupts so you'd just have to YOLO the timing.

> NVIDIA GPU drivers are typically around 800M-1.5G.

They also pack in a lot of game-specific optimizations for whatever reason. Could likely be a lot smaller without those.

Even the open source drivers without those hacks are massive. Each type of card has its own almost 100MB of firmware that runs on the card on Nvidia.

That's 100MB of RISC-V code, believe it or not, despite Nvidias ARM fixation.

Someone last winter was asking for help with large docker images and it came about that it was for AI pipelines. The vast majority of the image was Nvidia binaries. That was wild. Horrifying, really. WTF is going on over there?

You’re assuming a discrete GPU with separate VRAM, and only supporting hardware accelerated rendering. If you have that you almost certainly have more than 2MB of ram

Aren’t you cheating by having additional ram dedicated for gpu use exclusively? :)

VGA standard supports up to 256k

Computers didn't used to have GPUs back then when 150kB was a significant amount of graphics memory.

The Acorn Archimedes had the whole OS on a 512KB ROM.

That said, OSs came with a lot less stuff then.

That's only RISC OS 2 though. RISC OS 3 was 2MB, and even 3.7 didn't have everything in ROM as Acorn had introduced the !Boot directory for softloading a large amount of 'stuff' at boot time.

If that is a lot less of things not needed for the specific use case, that is still a big plus.

It was GUI defined manually by pixel coordinates, having more flexible guis that could autoscale and other snazy things made things really "slow" back then..

Sure we could go back... Maybe we should. But there are lots of stuff we take for granted to day that were not available back then.

RISC OS has the concept of "OS units" which don't map directly onto pixels 1:1, and it was possible to fiddle with the ratio on the RiscPC from 1994 onwards, giving reasonably-scaled windows and icons in high-resolution modes such as 1080p.

It's hinted at in this tutorial, but you'd have to go through the Programmer's Reference Manual for the full details: https://www.stevefryatt.org.uk/risc-os/wimp-prog/window-theo...

RISC OS 3.5 (1994) was still 2MB in size, supplied on ROM.

The OS did ship with bezier vector font support. AFAIK it was the first GUI to do so.

P.S. I should probably mention that there wasn't room in the ROM for the vector fonts; these needed to be loaded from some other medium.

Yea, but those platforms were not 64bit

64 bit generally adds about 20% to the size of the executables and programs as t to last on x86, so it's not that big of a change.

When I first started using QNX back in 1987/88 it was distributed on a couple of 1.4MB floppy diskettes! And you could install a graphical desktop that was a 40KB distribution!

I would like to have this again

I prefer to use additional RAM and disk for data not code

There’s an installation option to run apps off disk. It’s called “The Mount Mode of Operation: TCE/Install”.

To think that the entire distro would fit in a reasonable LLC (last level cache)..

I've been wondering if I could pull the DIMM from a running machine if everything was cached.

Probably not due to DMA buffers. Maybe a headless machine.

But would be funny to see.

Like the k language!

With 320x240 pixels and 256 colors

"640k ought to be enough for everyone!"

> Or 128K of ram and 400 kb disk for that matter.

Or 32K of RAM and 64KB disk for that matter.

What's your point? That the industry and what's commonly available gets bigger?

Typical people I know today think it’s normal to spend $300 a month on a family phone plan ($3600 a year).

Back then, you had 1 phone which cost around $50 a month in inflation adjusted dollars.

Also STRYKER.

My parents saved up for years and then kept the same computer for years more. It was normal to have a machine for 10 years, and just one per household.

Upgrading with a hard disk, a second floppy drive, or upgrading the graphics card was common.

PC-MAN did the same thing - except for the LOOP $ delay loops used for the opening screen.

Alley Cat has a timing issue with the “kiss” noise used at the end of the Felicia screen and the “conflict” noise when you run into the dog.

Great to see a discussion of Greg Kuperburg’s games. I love playing PC-MAN for 3 minutes or so with my oldest son. Perfect amount of time and attention.

I’m currently busy disassembling PC-MAN and trying to cram it into 16kB so it could be run on an original PC loaded from cassette. The game was distributed as a DOS booter and boo floppies required 32k, but PC-MAN is about 17k in size.

I got my last Epi-Pen for free, since in my state Medicaid has no copay for prescriptions or else it's $2 or $3.

Medicaid is available in all 50 states.

It is not available to all poor people in all states

https://www.cbpp.org/research/health/the-medicaid-coverage-g...

All the non-expansion states make it available too; just at a lower income threshold or other requirements.

Making a statement that most poor people in America can’t get epi-pens is false. Most poor people can for free.

> Making a statement that most poor people in America can’t get epi-pens is false. Most poor people can for free.

There are 26 million uninsured in the US and deductibles also exist for everyone so this is just blatantly false

I would consider it extremely likely they have some kind of visibility into your data at Meta via Palantir.

You could say you use Y Combinator's news comment board.