Bill Joy On Sun, Microsoft, Open Source, and Creativity

maitas writes "In this interview, Bill Joy talks about green energy and technology. His main point is: 'I'm all for sharing, but I recognize the truly great things may not come from that environment.'" The interview really runs the spectrum from the iPad to Microsoft, and from green tech to nanotech.

Comments

Sun software

[yyxx]

I was using Sun workstations for a long time. Their hardware was decent and cheap. As for the software, the best thing about it was that you could remove most of the Sun crap and replace it with GNU software. And when the Linux kernel was reasonably stable and we got cheap PC hardware, it was time to ditch the Sun hardware too. That's the history of Sun and Sun software R&D in a nutshell (except for Java, which is another sad story).

Re:Sun software

[MoxFulder]

I was using Sun workstations for a long time. Their hardware was decent and cheap. As for the software, the best thing about it was that you could remove most of the Sun crap and replace it with GNU software. And when the Linux kernel was reasonably stable and we got cheap PC hardware, it was time to ditch the Sun hardware too. That's the history of Sun and Sun software R&D in a nutshell (except for Java, which is another sad story).

I agree with this assessment, other than Sun hardware being cheap... perhaps it was a bargain compared to other commercial Unix vendors back in the 90s, but by the time it became plausible to choose between Linux-on-x86 and Solaris-on-Sun, Sun was really way more expensive.

Here's my historical perspective...

In 2001-2002, I worked at a small company making speech synthesis software. Our products had been developed on Sun workstations, and most of us developers used them still. They were very reliable once set up correctly, and they had nice, big, clear CRT monitors, nice optical mice, nice keyboards with extra programmable function keys, and fast SCSI hard drives. They ran the CDE GUI desktop, which was ugly and clunky, but worked out-of-the-box. We relied on the proprietary XWave software for audio waveform analysis, but otherwise used GNU tools almost exclusively.

Developers, especially the young-uns like myself, were rapidly acquiring enthusiasm for Linux. I was 19 and had been using Linux for years and got a lot of my older coworkers enthused, although I liked Solaris too.

Solaris still had a few key advantages:

• Audio "just worked." Getting OSS audio (/dev/dsp) to work under Linux was a chipset-dependent pain in the ass and it the device I/O semantics differed from Solaris.
• GUI desktop: Solaris's CDE desktop sucked, but GNOME was pretty awful in those days too.

Linux was building up a lot of advantages though, and fast:

• Any old programmer could slap it on any old Windows box lying around. Solaris hardware was expensive as hell, and no one knew how to upgrade it besides our one in-house guru. There was no plug-and-play... even replacing a Sun keyboard could have incomprehensibly weird side effects. Linux was bending over backwards to get plug-and-play support for all kinds of hardware.
• Package management. Debian had APT already, which rocked. You could just apt-get GCC/GDB/Emacs/CVS/MySQL and be up and running. Under Solaris, we had to rebuild everything from source on new/reinstalled systems. An annoying bottleneck, and Sun was slow to recognize and embrace this software distribution model. The community-run Sunfreeware.com was in an embryonic stage at that point.
• Way faster compile times. x86 processors (P3? P4?) were killing Sparc. I remember that Solaris was very reliable for multitasking, whereas Linux at that time would bog down when you ran too many CPU- or I/O-intensive tasks at once. But if you were running one big compile and needed to finish it ASAP, x86 was superior.

Basically, Linux was fixing its deficiencies (audio, reliability, GUI) a lot faster than Sun was fixing theirs. Performance comparison was exacerbated by Sun's hardware: it was expensive and hard to upgrade, so we resisted upgrading it, so it started to seem slower and slower and even less appealing.

Sun had built its business on reliable hardware coupled with a highly-regarded, reliable UNIX OS that only had to support a small range of hardware (not unlike Apple's Mac model). They seem to have been completely blindsided by Linux's ability to support an incredible range of commodity hardware, and they seemed utterly ignorant of the fact that their proprietary development tools sucked, and everyone wanted to use GNU tools.

Re:Even Windows for free would have replaced Solar

[Mr. Slippery]

Remember - most(almost all?) open source contributions come from people who have software jobs, quite often jobs which directly compete with the open source initiative they are contributing to.

More like, "quite often jobs which make use of the open source initiative they are contributing to."

For example, a while ago I contributed some code to WebInject. It was code I got paid to write at my day job; I found WebInject, said "This would be useful to us if it had X, Y, and Z", added X, Y, and Z -- getting paid to do so, same as if I was writing our own bespoke test tool -- and contributed the code back.

I suspect that this sort of scenario is at least an order of magnitude more common than people contributing to free software projects that directly compete with their day jobs.