Comment by TacticalCoder
8 hours ago
> What everybody except Steve Jobs got wrong was not taking user interface seriously enough.
But Linux took over the world and it had a crappy UI. I was there, back when Xorg didn't even exist, running X and messing with modelines on Linux. All the window manager sucked big times.
Sun, just like SGI, got attacked by cheap beige PCs running either Windows or Linux.
I don't think their UI had much to do with their downfall and, arguably, SGI's UIs were better than the competition and they still disappeared too.
Also weren't most machines sold by Sun servers anyway? Their famous "We the dot in .com" tagline.
> But Linux took over the world and it had a crappy UI. I was there, back when Xorg didn't even exist, running X and messing with modelines on Linux. All the window manager sucked big times.
My first exposure to Solaris was in 2003 when Linux had Gnome 2.6 and Solaris had CDE. The UI on my Debian box seemed decades ahead of the Solaris machines I managed, which looked like Windows 3.1 and performed about as well.
Combined with Linux package managers being infinitely better, the GNU core utilities having so many more conveniences and options than the Solaris equivalents, I felt like Sun was this dinosaur that couldn’t die fast enough. I’ll still never understand why so many people praise Sun for its engineering culture when free software seemed so superior to me at the time.
My favorite anecdote exemplifying my opinion about Solaris at the time: you log into the console of your Sun workstation, running SunOS, with your Sun keyboard plugged in, and press backspace. It echoes ^H back to you. They couldn’t even figure out how to configure their own keyboards correctly out of the box.
Maybe in the 90s the standards were different. Maybe their tech was better. But by the time I ever used their software is was the butt of jokes.
This was my experience as well, from the same time period 2003ish. My University's computer lab was all Sun. I thought CDE was shocking, I was immediately repelled by it when I started using it and had a similar experience with those Sun keyboards.
One of the other students in that class gave me a Mandrake Linux cd (It ran KDE) he taught me how to use SSH. After that I worked from home and submitted my assignments remotely. I have not looked back from Linux since.
I don't think it was limited to Sun. The first job I had out of university there were computers running HPUX also had CDE gave me flashbacks straight back to university. It definitely gave me the impression that all the commercial Unix companies were ancient Dinosaurs that had zero taste.
> My favorite anecdote exemplifying my opinion about Solaris at the time: you log into the console of your Sun workstation, running SunOS, with your Sun keyboard plugged in, and press backspace. It echoes ^H back to you. They couldn’t even figure out how to configure their own keyboards correctly out of the box.
That's really an old-school UNIX convention that should have died with the move from teletypes to CRT terminals. The default behavior was that ERASE was # and KILL was @ (rather than backspace and Ctrl-C). It was typical for these to get set properly on logging in.
Ahh, ^H brings back memories. We had Solaris x86 machines in the computer lab at UC Santa Barbara when I was there from 97 to 02. They were deathly slow never mind that CDE was miserable too. I made it slightly less painful by loading the BSD versions of the utilities and using those. Pretty much gave up and ran FreeBSD on the desktop with Window Maker.
Sun got by because they had faster and more reliable hardware than what was available in the x86 space at the time. eBay's services ran on a handful of 64 CPU Starfire servers. And proprietary UNIX workstations had a bunch of specialized apps that weren't available on Linux at that point.
The vast majority of Linux installations don’t have a UI, they’re headless, and it was the same with Sun. I.E. the UI is irrelevant to the success or failure of Linux and Sun.
Linux took over the world because any licensing model that requires a per CPU license simply can’t work in a massively scalable situation like server farms and cloud.
At the end of the day, the UI on a Unix server was always shell based. That was just the best way to configure a server. (And, in some way Infrastructure as Code is an evolution of the shell scripts a unix admin would use.)
Linux won on the server side because they got just as right as Unix. The crappy UI (windows) was the one that didn't work with shell scripting, didn't work with ssh, etc.
IaC is really just declarative shell scripts. It's a different language, but it's often just kicked off by some bash in a git action.
Once you have git and actions, I don't think that the language really matters all that much.
Linux was free, and people will put up with a lot of shit for that.
I also recall that Linux on a PC often beat the performance of Unix equipment I could access in contemporary times. There was a lore around those machines, but my actual experience doing basic FOSS-adjacent dev work (bash, make, gcc, autoconf) was that the Linux PC was faster as well as inexpensive.
The exceptions were mostly hardware graphics accelerators in some labs, when my Linux PC had to do it in software. And I got myself a surplus DEC Multia (alpha) that was really fast with Linux for natively compiled code, beating my contemporary 486DX3-100. But, the PC was still faster for web browsing, because running Netscape on alpha meant doing x86 emulation!
Likewise, my ~90 MHz Pentium MMX laptop with Linux was faster at regular dev work on my work-supplied Sun Ultra workstation in the late 90s. (Edit: Hmm, I am not sure about this laptop anymore, it might have been 166 MHz??)
Due to my positive experiences, I was an aggressive early adopter (and frequent beta tester) of nearly every Linux ecosystem frontier for decades. E.g. 2D accelerators, 3D accelerators, SMP, large file support, gigabit ethernet, wireless ethernet, power management, virtualization, 64 bit architectures, SCSI/SATA/Firewire/e-SATA/NVME, hardware and MD-RAID, ext2/ext3/xfs/btrfs, ...
Exactly. Free vs. $8k (in 1990 dollars) for a NeXT computer was a pretty easy choice for most people.
"Free", until mid 90s I guess most people bought Linux distros on CDs, later DVD (no downloads). For me the big, big plus on Linux (vs. Dos / Windows), had been all the free compilers.
2 replies →
to be fair, nobody in 1990 decided between NeXT and Linux, because (A) Linux didn’t exist yet and (B) they were / would have been completely different markets and totally different capabilities.
3 replies →
> and it had a crappy UI.
That wasn't even true back then.
> But Linux took over the world and it had a crappy UI.
Linux took over in all the places where a good GUI didn't matter. In ran most of the Unix server software on hardware orders of magnitude cheaper than what was needed to run a commercial Unix. Windows ran on that same cheap hardware but was much more expensive for the same capability.
Despite Sun's marketing push Linux was really the dot in most of the dotcom darlings.
> Linux took over in all the places where a good GUI didn't matter.
like smartphones ;-)
There's a vast gulf between Android and a Linux desktop environment. Android is a Linux kernel with the Android environment on top and is not just some Linux distro like anyone uses on a PC. ChromeOS is very similar.
I remember a web page put up by an SGI employee, saying that x86 wasn't going to destroy the Unix workstation business. The title was "Pecked To Death By Ducks".
Then Intel came out with dual-core processors, which was the new hotness. So he wrote a followup page, literally titled "Pecked To Death By Ducks With Two Bills".
At the time, I thought they were funny, and made reasonable arguments. But in retrospect, they were just trying to distract the customers from the fact that Intel was overtaking them.