Will I install it again? No. Like most everyone else, I've moved on to Linux and BSD.
But I'll spin up Solaris 2.5.1 on my SPARCbook or Solaris 10 on my Blade 1500 whenever I need a nostalgia kick.
Point is: I remember writing scripts for then main environment and every script had to work on all terminals. The three variations meant constantly armor-plating things with HPUX/AIX/SUNOS variable checks. I hated that. I was glad when RedHat came along and we stopped doing that.
(Also, back then it was a sign of being cool to have an 8514/A monitor, and not the B&W sun terminals.)
The speed with which Linux killed Solaris however was astonishing.
Perhaps that was just an isolated incident confined to the university recruiters but Sun at that time seemed overly focused on fighting Microsoft instead of improving their product offerings. Perhaps some of anger ended up turning potential users and employees away from Microsoft but I suspect lots of those ended up going with Linux or some other alternative rather than going with Sun.
But they'd been like that for a long time, since before I started there in 1990, long before Java. They DEFINED themselves in terms of Microsoft, to the extreme extent that when Sun Microsystems fell apart into separate divisions, they actually named one of them "SunSoft" to directly position it against Microsoft. As if.
The management at Sun didn't consider Java to be a programming language or software platform, they considered it to be first and foremost their primary weapon of mass destruction in their apocalyptic war against Microsoft, and they didn't consider Java developers to be loyal cherished customers, they considered them to be disposable brainwashed mercenaries in their World Wide War against Microsoft.
It was funny when Sun proudly and unilaterally proclaimed that Sun put the "dot" into "dot com", leaving it wide open for Microsoft to slyly counter that oh yeah, well Microsoft put the "COM" into "dot com" -- i.e. ActiveX, IE, MSJVM, IIS, OLE, Visual Basic, Excel, Word, etc!
And then IBM mocked "When they put the dot into dot-com, they forgot how they were going to connect the dots," after sassily rolling out Eclipse just to cast a dark shadow on Java. Badoom psssh!
https://www.itbusiness.ca/news/ibm-brings-on-demand-computin...
Sun totally dropped the ball fighting their true original enemy AT&T, and they should have put all that effort and energy into improving SunOS and railing against AT&T after SunOS finally beat System V in the Unix market, instead of capitulating to AT&T AFTER SunOS won the Unix war against System V, and then rolling over, giving up, selling out to their mortal enemy, and becoming Solaris.
To port my favorite cross platform Apple/IBM joke:
Q: What do you get when you cross Sun and AT&T?
A: AT&T.
Sun completely forgot they were a hardware company. They made great hardware, but while Microsoft made it easy to develop software for Windows, Sun didn't make it easy to do the same for Solaris (maybe out of fear it'd benefit their remaining hardware enemies IBM and HP). They didn't even try very hard to build low-end machines that could do PC tasks at competitive prices.
(Still used a Sun workstation for the next five years at my first job, though.)
Atlanta is little? (Not aware of whether or not Mellow mushroom is in other states or not)
Several times in my computing career -- say, from 1990 on -- I've seen something start to happen that was clearly going to be huge, only to be told over and over that the obviously-inevitable development was Impossible.
The first example was the ascendancy of the microcomputer. "OH, there will ALWAYS be a space for mainframes! Corporations aren't going to do their accounting on some crappy PC! You're just young and don't see how these things really work."
The second was the way Linux completely destroyed the commercial Unix market (and, with it, the "fancy Unix workstation" market). The power of FOSS is, in my view, very very often overstated by idealists, but it's hard not to see the wholesale corporate/academic/research shift from Solaris, et. al., to Linux on commodity hardware as anything but a HUGE win in terms of computing power.
When Oracle bought them, they pulled the rug and demanded $1K per CPU per year for the privilege if it was non-Sun/Oracle hardware.
If they had just copied the RedHat model they might have been onto a winner.
Free unstable, paid stable.
Yes, that includes the Red Hat model (remember OpenSolaris?)
It was not a lack of persistence
They were a hardware company that designed their own CPUs. If a company could pull an Apple and do an M1, it would be Sun.
Unfortunately, their low-cost machines performed just like PCs and cost the same, without PC software to run on them.
And building x86 boxes that ran with Solaris ensured they'd be no better than PCs, without any compelling software.
Back then I was on Windows and everything I saw on Sun systems looked weird, laggy and just super obscure to me. The whole UI felt like a mix of enterprise and neckbeard charme despite being non-open. It did not seem to run none any of the things I cared about back then. Later on I remembered Sun mainly in context of their stewardship of Java which also didn't help connecting any good feelings to it.
So maybe in 2022 someone could enlighten me about what exactly was so cool about Sun?
The Sun hardware was materially better, or so I recall. It was also more attractive. I actually still have an Ultra5 on a shelf in my office that, for a time, ran Oakley's online store (which my old company built). It's still a good looking machine.
Solaris, too, was widely viewed as better/more workable/easier to use than its competitors. It had more polish.
If you're looking at Solaris with the latest build of Ubuntu or whatever in mind, yeah, it's gonna look neckbeardy and hokey. But that's not the world Solaris was the king of. Back then, using a commercial Unix meant you got a LOT of shit that Just Worked out of the box, and generally an easier time getting your shit running.
I probably spent most of a week building stuff from source and customizing my system. If you wanted a "Unix workstation", Sun was the best. SGI looked cooler but often getting open source stuff built on it was a PITA.
This was during the Windows 95 / NT 4.0 era, so it was definitely a step up on Windows.
Workstations typically came with fast SCSI disks. Once I wanted to find a file after a failed install and used `find`. The search came back immediately, empty. I concluded I was doing something wrong, did `man find` to confirm there wasn't any Sun-specific oddity. Then I tried with `find / -name passwd` because I knew that file existed. Result came back immediately. It was probably caching the whole directory structure in RAM, but the instantaneous response really impressed me back then.
Mouse was laggy, because, IIRC, it communicated at 9600 bps or something like that. Macs and SGIs were much smoother.
> So maybe in 2022 someone could enlighten me about what exactly was so cool about Sun?
The one thing I loved was X and its network transparency. I often run NetScape on a beefier box and it was just like having it running on my workstation. Same with software development and engineering tools. I could reach out to multiple computers (and not only Solaris ones, but any Unix) from my desktop easily.
The same could be said about any other Unix-based workstation with X (NeXT didn't come with it).
The PCs of now-a-days are light-years ahead of what PCs were available in the Workstation era, so much beyond that time, that nobody can think back to that time. This comparision kind of smears some time frames together..
So, in the early 90's , Sun was starting to ship 4-way MP, and 8-way MP. In 95, they introduced 64-bit CPUs just as PCs were really starting to use 32-bit computing. Typical speeds were from 50MHz - 400MHz in this time frame. At the high end you could get up to 256MB of RAM.
Sun Frame buffers at the time were doing 1920x1080 resolutions. (or slightly higher, but granted, a lot of these high-res were black & white in the earlier time frames).
They came with multiple fast/wide SCSI drives, tape drives, etc.
They all came with the vendor's unix. System uptime was usually quite good, months to over a year.
You did have some vendor lock with what software was available, but generally Sun was top of the heap, every vendor had something available for Sun.
Contrast this with PCs at the time.
State of the art PCs were 486 running 50MHz-75MHz, with 8MB of RAM. Maybe 32MB if you wanted to get Windows 95 going.
Video was VGA at 640x480, again not much need for more until after Win95 launched.
Hard drives were IDE, a whopping two if you were lucky. While SCSI existed on PC, it wasn't very common.
Linux and *BSD didn't come out until 1992-93 Even then, uptime of DOS systems or PC unix at the earliest time frame could be measured in days. You had to get _just the right_ hardware to make things work together.
By the end of the '90s. Workstations hadn't progressed very far. PC's grew by leaps and bounds, catching up and surpassing workstations handily.
PCs were actually fairly stable and the PC unix could give you really decent uptime.
But in the before the explosion of PCs, if you wanted power or uptime, or graphics, or disk speed, etc. etc. compared to what else was there, Workstations were the way to go.
The other thing is that OSS databases mostly weren’t a thing, and commercial database platforms like Oracle or Informix were released for Solaris/SPARC first. I worked for a .gov client that was all about RS/6000 and AIX, and it was painful - you’d wait 3-6 months for some patches.
Linux really broke the workstation market circa 1998, and the cracks in the server platform were sort of hidden after the dotcom implosion put lots of gently used gear on the market.
To me this describes what it’s like trying to develop software for Windows vs Unix
Also being different ?
Most of the deployments were naturally Solaris based.
(I still remember when Sun introduced color systems, and they were such status symbols that they went mostly to managers, including one who was color-blind. A warning sign in retrospect, I suppose.)
Good enough and free will beat out better and expensive almost every time (the key part being "good enough", which, granted, makes that rule someone circular ;) ).
It makes the snobs look a bit ridiculous: "Oh, it only took you five years to get to where we were... but we were on expensive hardware you couldn't buy without a major contract, running fewer applications, and invested in a system that would be eaten by the low end so comprehensively now only snobs know how good it was."
I'd give it a chance if it came with CDE or OpenWindows.