4 ms·

I remember being assigned to look into Solaris when working as a volunteer sysadmin in grad school, where we were a SunOS shop. I took a sparcstation, wiped it

I remember being assigned to look into Solaris when working as a volunteer sysadmin in grad school, where we were a SunOS shop. I took a sparcstation, wiped it, and installed Solaris. This was 1992 or so, so it must have been 5.0 or 5.1. I hated it, but I don't remember very many specifics about why I didn't like it. I think it was partially the unbundled compilers, combined with everything just being "different", combined with perceived slowness. That was the last place I worked with Suns, as my first job was sysadmin'ing DEC Ultrix boxes, and DEC Alphas. Ultrix & OSF/1 were much closer to SunOS than Solaris, ironically. I do wish that Sun would have evolved the BSD kernel rather than jumped to System V.

pjmlp 5y ago
That was around the time that GCC finally started to get some wind, due to the unbundling of UNIX SDK.
- drewg123 5y ago
  I think it also got wind because it was just so much easier to compile stuff with gcc than it was to use the vendor compilers, with all of their incompatible flags and extensions. This was in the days before package managers, when everybody compiled open source stuff themselves, a lot of things didn't use autoconf, etc. I remember compiling almost all open source stuff (emacs, tex, postscript, file utils, etc) with GCC, and reserving the vendor compiler for situations where performance actually matters (math / linear algebra packages, professors' code). EDIT: I remember a few years where people tended to assume all the world ran SunOS 4.1, just like people assume all the world runs some flavor of debian/ubuntu now.
- DonHopkins 5y ago
  The unbundling of the free C compiler and the high price of the unbundled C compiler and AT&T's shitty bloated C++ compiler was emblematic of what was so bad about Sun abandoning their Berkeley BSD roots and getting into bed with AT&T System V with Solaris. And that provided an opportunity for Cygnus Solutions. Not coincidentally, after he founded Cygnus Solutions (which Red Hat later bought), Michael Tiemann worked closely with Sun to support GCC on their platform. https://web.archive.org/web/20160310075610/http://www.toad.com/gnu/cygnus/index.html https://web.archive.org/web/20160310075610/http://www.toad.c... >We had the grandiose idea that major computer companies like Sun, SGI, and DEC would fire their compiler departments and use our free compilers and debuggers instead, paying us a million dollars a year for support and development. That wasn't quite right, but before we starved, we stumbled into the embedded systems market, doing jobs for Intel (the i960, a now-forgotten RISC chip), AMD (their now-forgotten but nice 29000 RISC), and various companies like 3Com and Adobe who had to port major pieces of code to these chips. In that market, once we fixed the tools to support cross-compiling, we had major advantages over the existing competitors, and we swarmed right through the market for 32-bit embedded system programming tools. And ultimately, we did get million-dollar contracts, such as one from Sony for building Playstation compilers and emulators. This allowed game developers to start working a year before the Playstation hardware was available. This enabled the Playstation to come to market sooner, with more and better games. https://web.archive.org/web/20150701032848/http://www.toad.com/gnu/cygnus/company-overview.txt https://web.archive.org/web/20150701032848/http://www.toad.c... >Michael Tiemann, President, has been writing free software since 1987. He wrote the code for GNU C's function inlining. He wrote a portable instruction scheduler which boosted GNU C's performance by 30\% on the SPARC. He is the author of GNU C++, the first available native code C++ compiler. Mr. Tiemann has ported the GNU compiler to the SPARC, Motorola 88000, and National 32032 architectures, as well as adding support for Sun's FPA board on Sun 3s. He ported the GNU debugger to the SPARC and Intel 80386 architectures, extended the debugger and linker to handle C++ features, and ported the linker to SPARC. https://www.oreilly.com/openbook/opensources/book/tiemans.html https://www.oreilly.com/openbook/opensources/book/tiemans.ht... >The real bombshell came in June of 1987, when Stallman released the GNU C Compiler (GCC) Version 1.0. I downloaded it immediately, and I used all the tricks I'd read about in the Emacs and GDB manuals to quickly learn its 110,000 lines of code. Stallman's compiler supported two platforms in its first release: the venerable VAX and the new Sun3 workstation. It handily generated better code on these platforms than the respective vendors' compilers could muster. In two weeks, I had ported GCC to a new microprocessor (the 32032 from National Semiconductor), and the resulting port was 20% faster than the proprietary compiler supplied by National. With another two weeks of hacking, I had raised the delta to 40%. (It was often said that the reason the National chip faded from existence was because it was supposed to be a 1 MIPS chip, to compete with Motorola's 68020, but when it was released, it only clocked .75 MIPS on application benchmarks. Note that 140% * 0.75 MIPS = 1.05 MIPS. How much did poor compiler technology cost National?) Compilers, Debuggers, and Editors are the Big 3 tools that programmers use on a day-to-day basis. GCC, GDB, and Emacs were so profoundly better than the proprietary alternatives, I could not help but think about how much money (not to mention economic benefit) there would be in replacing proprietary technology with technology that was not only better, but also getting better faster.

pjmlp 5y ago

That was around the time that GCC finally started to get some wind, due to the unbundling of UNIX SDK.

drewg123 5y ago
I think it also got wind because it was just so much easier to compile stuff with gcc than it was to use the vendor compilers, with all of their incompatible flags and extensions. This was in the days before package managers, when everybody compiled open source stuff themselves, a lot of things didn't use autoconf, etc. I remember compiling almost all open source stuff (emacs, tex, postscript, file utils, etc) with GCC, and reserving the vendor compiler for situations where performance actually matters (math / linear algebra packages, professors' code). EDIT: I remember a few years where people tended to assume all the world ran SunOS 4.1, just like people assume all the world runs some flavor of debian/ubuntu now.
DonHopkins 5y ago
The unbundling of the free C compiler and the high price of the unbundled C compiler and AT&T's shitty bloated C++ compiler was emblematic of what was so bad about Sun abandoning their Berkeley BSD roots and getting into bed with AT&T System V with Solaris. And that provided an opportunity for Cygnus Solutions. Not coincidentally, after he founded Cygnus Solutions (which Red Hat later bought), Michael Tiemann worked closely with Sun to support GCC on their platform. https://web.archive.org/web/20160310075610/http://www.toad.com/gnu/cygnus/index.html https://web.archive.org/web/20160310075610/http://www.toad.c... >We had the grandiose idea that major computer companies like Sun, SGI, and DEC would fire their compiler departments and use our free compilers and debuggers instead, paying us a million dollars a year for support and development. That wasn't quite right, but before we starved, we stumbled into the embedded systems market, doing jobs for Intel (the i960, a now-forgotten RISC chip), AMD (their now-forgotten but nice 29000 RISC), and various companies like 3Com and Adobe who had to port major pieces of code to these chips. In that market, once we fixed the tools to support cross-compiling, we had major advantages over the existing competitors, and we swarmed right through the market for 32-bit embedded system programming tools. And ultimately, we did get million-dollar contracts, such as one from Sony for building Playstation compilers and emulators. This allowed game developers to start working a year before the Playstation hardware was available. This enabled the Playstation to come to market sooner, with more and better games. https://web.archive.org/web/20150701032848/http://www.toad.com/gnu/cygnus/company-overview.txt https://web.archive.org/web/20150701032848/http://www.toad.c... >Michael Tiemann, President, has been writing free software since 1987. He wrote the code for GNU C's function inlining. He wrote a portable instruction scheduler which boosted GNU C's performance by 30\% on the SPARC. He is the author of GNU C++, the first available native code C++ compiler. Mr. Tiemann has ported the GNU compiler to the SPARC, Motorola 88000, and National 32032 architectures, as well as adding support for Sun's FPA board on Sun 3s. He ported the GNU debugger to the SPARC and Intel 80386 architectures, extended the debugger and linker to handle C++ features, and ported the linker to SPARC. https://www.oreilly.com/openbook/opensources/book/tiemans.html https://www.oreilly.com/openbook/opensources/book/tiemans.ht... >The real bombshell came in June of 1987, when Stallman released the GNU C Compiler (GCC) Version 1.0. I downloaded it immediately, and I used all the tricks I'd read about in the Emacs and GDB manuals to quickly learn its 110,000 lines of code. Stallman's compiler supported two platforms in its first release: the venerable VAX and the new Sun3 workstation. It handily generated better code on these platforms than the respective vendors' compilers could muster. In two weeks, I had ported GCC to a new microprocessor (the 32032 from National Semiconductor), and the resulting port was 20% faster than the proprietary compiler supplied by National. With another two weeks of hacking, I had raised the delta to 40%. (It was often said that the reason the National chip faded from existence was because it was supposed to be a 1 MIPS chip, to compete with Motorola's 68020, but when it was released, it only clocked .75 MIPS on application benchmarks. Note that 140% * 0.75 MIPS = 1.05 MIPS. How much did poor compiler technology cost National?) Compilers, Debuggers, and Editors are the Big 3 tools that programmers use on a day-to-day basis. GCC, GDB, and Emacs were so profoundly better than the proprietary alternatives, I could not help but think about how much money (not to mention economic benefit) there would be in replacing proprietary technology with technology that was not only better, but also getting better faster.