Hacker News new | past | comments | ask | show | jobs | submit login

>This brings back memories. It's simplicity and elegance was awesome. I really view Linux as the ultimate successor to SunOS4 (as opposed to Solaris / SunOS 5).

I take your point, but I'd say that the various BSDs are more the successors to Sunos(1-4) than is GNU/Linux.

Since (and it pissed me off at the time) SunOS5+ (Solaris) has a sysV admin/userland (as does GNU/Linux), whereas the BSDs are much closer to what SunOS 4 and its predecessors (based on various Unix/BSD codebases. IIRC, SunOS 4.4 -- the last version that wasn't "Solaris" was based on BSD4.3[0])

All that said, Solaris had some pretty impressive features that Linux is still catching up with. E.g., Zones[1], ZFS[2], etc.

At the same time, there was much more of the "hacker" dynamic with SunOS4 (and its predecessors) than Solaris. Which has also been the case with GNU/Linux.

The former because before Linux (not to mention 386BSD[3]), Unix OS licenses (let alone the hardware it ran on) were way too expensive for widespread use. As such, it was mostly college students using their Sun Boxen to hack on.

But once there was free (as in libre and --well, mostly-- as in beer) Unix (386BSD) and Unix-like (Linux) available for commodity hardware, many, many more folks could access *nix systems to hack on.

From a technical standpoint the BSDs are the real successors to SunOS, but from a Dev/hacker culture standpoint, Linux is (as you point out) a successor as well.

[0] https://en.wikipedia.org/wiki/Berkeley_Software_Distribution

[1] https://en.wikipedia.org/wiki/Solaris_Containers

[2] https://en.wikipedia.org/wiki/ZFS

[3] https://en.wikipedia.org/wiki/386BSD [4]

[4] As an aside, I'm enormously grateful to Lynne and Bill Jolitz for 386BSD. It was a joy to be able to own and use it on my own hardware back then. Especially since $job at the time was mostly on Sun/SPARC. With an additional shout out to Yggdrasil Linux[5], the first Linux distro I ever installed and enjoyed that a lot too!

[5] https://en.wikipedia.org/wiki/Yggdrasil_Linux/GNU/X




Sun first big mistake was doing the deal with AT&T to do Solaris. Sun should have continued to champion BSD derived software.

This hurt them in a few ways, literally every other player in the ecosystem was unified against Sun/AT&T and where there could have been a lot more cooperation, now there was the opposite.

It also massively hurt them in that Solaris was basically shit for a very long time. Many costumers held on to SunOS 4.x for a very long time. Solaris was not popular but that where all the semantic multiprocessor stuff was done. There were people inside of Sun that even in the early 90s still wanted to switch back to SunOS.

Had Sun simply continued to improve SunOS, added multi-processor stuff there, open-sourced things like NEWS (something literally everybody in the company expect a few lawyers wanted) and continued to distribute their compiler for free, it would have become the standard and preferred Unix. They also should have continued with the i386 series of machines.

They put themselves on the path of incredibly proprietary OS and incredibly proprietary CPU that were very high priced as PC compatibles got cheaper and cheaper and x86 got faster and faster.


Yes, I worked at an early ISP and SunOS was popular for "shell" servers. The users did not want Solaris 2.x, mainly because open source software was tough to compile on it. SunOS 4.x was basically the standard open source target at that point (early to mid 90's.)

Solaris was pretty bad up until the >2.5 days. 2.5.1 and above were pretty solid releases. I ran 2.5.1 and 2.6 on my personal desktop (Sparc 10) for a while.


>and incredibly proprietary CPU

This is the opposite of true. sparcv8 is an IEEE standard. You might have to pay a license fee to Sparc International for certification and logo compliance, but that's it. Sun couldn't and didn't stop you from making them.


That only true for SPARCv8 and the first SPARCv9 was interceded in 1995. The IEEE standard was from 1994.

So practically speaking, Sun did move to proprietary CPUs. First when they interceded SPARC and then again when they switched to SPARCv9.

Granted a few years in between when there was a slight amount of competition but that was over by the mid-90s before x86 ever took off fully.


Please look up what "interceded" means.

Even if v9 isn't an IEEE standard, you can still pay Sparc Intn'l for a license. Try going to intel with a request to make x86 chips.


> Please look up what "interceded" means.

That was an Autocorrect issue ...

One can say that 'incredibly proprietary' was wrong, it was less proprietary then many other CPU at the time. It doesn't really change the analysis. Practically speaking that gave most Sun costumers very few option, and increasingly less as the 90s went on. There were just to many RISC ISAs competing for to little market. Sun tried to play the vertical integration walled garden game and in that game its really, really damn hard to win.


I was introduced to UNIX via Xenix, the system was so expensive at the time that the teacher would bring 486 PC tower into class, that we timeshared in groups of 15 minute slots.

We were supposed to prepare our samples as much as we could in Turbo C 2.0 for MS-DOS. When doing UNIX stuff not available in Turbo C, prepare the code, hopefully the right way.

Then at the university, the system we had on the campus was DG/UX, and Solaris.

They eventually got replaced by Red-Hat Linux around 1997.

During the .com wave until 2005, most of the UNIX stuff I used in production was Solaris, HP-UX and Aix, outside my stay at CERN (which was moving away from Solaris around 2002), GNU/Linux was mostly for private stuff at home.

From all the UNIXes, NeXTSTEP derivates, and Solaris are my favourites, as they are/were more than just yet another UNIX clone (Irix as well, but I only know it from articles and legacy docs).




Consider applying for YC's Spring batch! Applications are open till Feb 11.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: