Developing for a single platform in 2025 is like developing for a single platform in 2005, if you don’t care about mobile.

The desktop marketshare of the various platforms hasn’t fundamentally shifted since then. Mobile was all additive, and Microsoft lost it. But Mac and Linux remain roughly where they were.

I believe, that the GP comment is too dismissive, but it is true. In 2005 when the dominance of IE started to dissolve it was not the best move to develop for a single browser. Though people still did it.

Today we see a rise of ARM on desktops, developing for x86 excludes Mac users, but the situation moves in a direction when exclusive x86 software will exclude an assortment of users of different OSes who chose to buy ARM desktop/laptop.

But I completely understand the choice made by the author, to use vector extensions on two (or three? RISC-V?) processors would be a much more additional work. The project is FOSS so anyone can jump in and add support for ARM vector extensions. Hopefully it will be easier then to write it from scratch, because you can compare intermediate results bit to bit, and catch mistakes red handed.

This chart indicates lots of growth for OSx since at least 2009 (as their data goes back to).

https://gs.statcounter.com/os-market-share/desktop/worldwide...

Would be interesting if they took out enterprise and/or computers were forced to use rather than chose to purchase.

Not everything is a VC funded thing. This is clearly a research project at a university - notice the ff.cuni.cz links in the images.

It's mff.cuni.cz. ff.cuni.cz is faculty of philosophy, mff is faculty of math and physics :)

Paraphrasing somebody: Win32 is the most stable Linux API.

Unironically, true.

As one of your sibling comments points out: it works perfectly on Wine.

Right, but here's the deal. Those same people tried and failed and learned from there mistakes, built rockets instead and still went to the moon.

Kind of an important last bit of the story, there.

Oh and in the span of human history, this happened in a blip of time...

We've barely finished inventing computers and the internet and we already are developing AI.

Ya'll need perspective.

That's a thing, in Washington.

Where does Washington get elves from?

They came from across the veil to join him in his fight against the English. Obviously.

Canada

Mostly Skamania county.

ASCII has tones of unused chars we could grab for coding.

Mm yes but BEL is not a character that can be typed on a keyboard with any amount of ease.

Is there an actual new source on this, beyond the Bezos Gazette?

in French (Its legalese, so the translation will be weird): https://www.courdecassation.fr/decision/636b6e3e67b11ddcd1c4...

this.

not the lame knock-off that is rimworld.

A single threaded heap of hot garbage-collection.

Except for the garbage-collection, single threaded code is extremely common across all "simulation" games. Too many interdependencies makes it difficult to parallelize.

Dwarf Fortress was looking into offload fluid computation to other cores, not sure if it was done.

I personally have not had a lot of performance issues with the base RimWorld game after a patch that was released targeting performance a year or two ago. Much more often it’s a poorly optimized mod that causes issues.

I don’t often play very far into the insane late endgame though (200+ colonists/raiders on the map at once etc). I could easily see it being a problem at that point.

And this is one of those successful PC games made with the Unity engine.

We use metric in the "American Space Program".

I cannot think of an engineering or scientific endeavor in the US that DOESN'T use the metric system.

The US is officially, metric; ansi converts back for the uneducated.

Here's a colorful video: https://youtu.be/SmSJXC6_qQ8?t=63

Here's boring government text: https://www.nist.gov/pml/owm/si-units-mass

> The US is officially, metric; ansi converts back for the uneducated.

This is the type of flamebait we don't need.

The American Public doesnt use metric. Thats why the article is written with miles.

Yeah why don't we just place an order at the first available orbital facto...oh no such thing exists. Well we can order some aluminum from an asteroid mining smelt...oh that doesn't exist. Well I'm sure the useful payload entirely built in orbit is...oh.

There's no infrastructure in space to do anything. Even if you built a rocket in orbit that doesn't do you any good if the payload in sitting on a pad on the ground.

In my opinion the ground is the best place to make them. It's much harder to do independent inspections of the work when you have to fly the expert who wants to look at the assembly into space and train them how to do EVAs.

The two things I think would be interesting to demonstrate in space is aluminum extrusion production (standardized framing) and fuel production. To do either of those sustainably in space would be an impressive accomplishment.