I reluctantly switched to Firefox because it still has add-ons and since Chrome's web tools are so good. With Mozilla's Rust adoption, Firefox got fast. This means my web products work a little better on Firefox, intentional or not. When enough people make that choice, a tipping point forms in the future. Paul Graham wrote about this in "The Return of the Mac" [^1].
Don't underestimate the power of your choice at the frontier, even if it takes a while to reverberate through time.
Decades later, I think we are still at a point where following his ideas come at a very steep price in performance and day to day usage.
For instance any dev that touches an iOs app in any way or form (even if it’s just to run in on test devices) is better off with a mac.
There’s ton of prevalent android apps that won’t work without the Play Store, and even rooting the phone is already seen as an hostile act from many vendors.
The list goes on and on, keeping hardware or software free is still an insane move that needs sizeable sacrifices. And it’s scary there’s no indication of the situation to change for the better.
 Y Combinator is (we hope) visited mostly by hackers. The proportions of OSes are: Windows 66.4%, Macintosh 18.8%, Linux 11.4%, and FreeBSD 1.5%. The Mac number is a big change from what it would have been five years ago.
Linux with 11% is quite high, I like that. What seems lower then expected are the Mac numbers to me.
Thanks for sharing.
We are developing an unfortunate dichotomy between commercially supported, closed ecosystems with lock-in and rapid update cycles that provide superior functionality and more community-driven, open ecosystems with standards and future-proofing but inferior functionality.
If the open versions aren't too far behind the closed ones in functionality and performance, that's just another form of competition and perhaps a healthy one. But if we start to get too much lock-in, which is inherently a one-way process favouring the closed systems in this situation, and in particular if important data or external systems become accessible only from the closed systems, then we have a more serious problem, as we're seeing ever more clearly with the worlds of mobile devices, IoT and "evergreen" software.
There is a spectrum from the cathedral to the bazaar. And there are only so many hours in the day. You could become a serf to principles, spending time others have for social activities or relaxing instead on maintaining a purely libre work flow.
So I'd much rather be a serf to principles and have control over my life than be a serf to a bunch of companies with their own goals and agendas. That's just me though.
That is true, but if the more free/open alternatives were never at the same starting point in the first place, you were probably screwed that way too. Neither option let you start in a good place and then move in a good direction, and this is the unfortunate reality I was commenting on above.
I've used Linux on the desktop previously and it left me wanting, so the fact that I haven't gone back tells me that things are indeed improving.
Meanwhile, OS X hasn't added a single feature I use since 10.6. All I've experienced over the years is that it kept getting more bloated, slower, and flakier.
Yes it is, but the reality is that if almost everyone is making the other choice, we stand to lose a huge amount by not following the crowd. That's not a price that most people are willing to pay, which limits the interest in and contributions to the open alternatives, and thus the vicious circle is closed.
Those are unrelated.
The third time, they used Rust. And it worked.
Mozilla research said rush enabled them doing safe multuthreading.
Tried again about six months ago, and I haven't looked back. Firefox is great again. Not sure if the Rust adoption happened between those two points in time.
I still have to use Chrome for Hangouts for work. And still trying to figure out a way around this.
- People preferring the compiler to yell at them to fix their Type mistakes before they hit the "run tests" button.
- Getting a better IR for better error messages.
In fact, it may be easier for developers to write code that runs faster with rust than C++, thanks to much error and exception handling code that you simply do not need in Rust.
Most importantly, however, is finding Rust developers who are experienced enough to write high-performance Rust. I would imagine that it's a lot easier to find C++ devs.
It may be, but there's also the concurrency aspect which is hard to get right in C++. That's where the performance gains come from, partially, because 'fearless concurrency' is one of the Rust's tenets.
That's a bold claim to make without any justification!