Java never felt "new" to me: rather unnecessarily bloated and slow and not offering enough high-level language features to justify the performance penalty over C. Managing memory in C never felt "hard" and when I needed something for quick hacking I could use VB and/or Delphi. Funny, but in many ways Visual Basic has been far superior RAD tool for enterprise SQL pushing. Yes, it was garbage-collected, ran in its own VM and produced fast-launching executables.
In mid/late 90s words like "Oracle", "Sun", "Solaris" and "Java" were used predominantly by older fat-faced salesmen types, while cool kids were coding C on Windows or playing with Linux. Yes, Microsoft was fun back then: it felt young, modern and innovative. Gates and Allen were Brin and Page of the 90s.
I was doing some cool real-time data acquisition/analytics on Windows NT with C++ and MFC. Yes, it felt bad-ass. In my geek circles Borland was considered the most exciting software company at the time. Sun/Oracle and IBM with OS/2? Those were for banks and GM.
Sun hasn't changed since then. I have nothing to thank this company for, except that with Java, they dumbed down our profession to nearly data-entry status and introduced an intolerable boredom to the world of software hacking, something it was missing before 1995. I also believe Java suppressed the development and widespread adoption of Lisp, Smalltalk, ML and Haskell: before Java lots of folks couldn't use them because computers weren't powerful enough, but now they aren't using them because schools stopped teaching and switched to vocational java training.
I wasn't surprised Oracle bought them. Their "Oracle Forms" BS always sucked. With Java and armies of school-trained drones supplemented by Indian outsourcers they finally have a wonderful distribution channel to keep pushing their overpriced stuff through. I guess SAP could buy them too. I've been fortunate enough to avoid this market, so I won't be missing anything.
I think you've got some causality messed up. I don't see how Java could have "dumbed down our profession." I think instead, schools turning vocational reduced the caliber of graduates while simultaneously pumping out large quantities of them.
It could have happened with just about any language. Java just happened to have a lot of things going for it that made it fit perfectly into the mold those vocational schools were looking for: simple memory management, platform abstraction, included libraries, and a large marketing effort creating jobs.
I remember when I was excited about Java. I was writing software for multiple platforms, and the thought of a write once, run everywhere was awesome. I felt so raped when it turned out to be a lie. I was so young and naive. :)
Now, I wish Java had never belched from the bowels of that leviathan. Even worse, I'm working at a game company, building an MMO, and using Java. Egads!
Mostly tangential, but I have to code in MFC at the day job (legacy code base). God, the winapi guys must have had a sadistic sense of humor when they designed it. My guess is they were so used to low level programming that they saw object orientation as a crutch for the infirm. I imagine them shoving random subroutines into objects while laughing maniacally to themselves: "let's see these pansies make sense of DDX controls! muahaaha!"
>before Java lots of folks couldn't use them because computers weren't powerful enough, but now they aren't using them because schools stopped teaching and switched to vocational java training.
Talk about self-denial. Java is definitely not the one to blame for lack of adoption of Lisp, Smalltalk, etc. Many other languages showed that it is possible to thrive despite perceived Java domination (I think that Java is not dominant, neither on desktop nor in web space, maybe in enterprise, but I wouldn't bet my money on it). And, after all, who cares about Java - use the tools you are proficient with, I know I do. Don't bitch about stuff you don't use.
In mid/late 90s I was working at MIT's university network (Athena) and my memory is similar, but different from yours. Perhaps because we weren't a Windows shop. I did actually work on a dual-boot x86 box sometimes, but the dual OSes were Linux and BSD.
We were still doing everything in C, but the cool kids I listened too were saying skip C++ if you want something OO but C-like and go straight to Java. So when I saw Per Bothner's Scheme-to-JVM compiler (Kawa) I knew my next implementation of BRL should use it.
What surprised me was the number of people at MIT who said, "Yuck, Scheme" when I was pushing it as the best syntax for web development (this after 12 years of me ignoring it in favor of C). I think a lot of people did not have a good experience going through 6.001 (SICP), even though other people, like me, loved it. I hate to say it, but what really suppressed the widespread adoption of Lisps may have been SICP.
You forgot Open Office which they bought and opened up and which has changed the competitive landscape in many ways. I'm struck by the number of non-geeks I know who use it daily.
SPARC was a great architecture, and the Sun workstations of yesteryear were where a lot of cutting edge research was done... It'll be interesting to see how many more OpenSolaris features Darwin picks up (ZFS?). It has dtrace already.
Darwin already pickup up ZFS. You can mount zfs volumes in OS X 10.5. You can patch the kernel module of OS X desktop to write to them (but don't, its really slow).
I use Open Office EVERY DAY. Let me say this.... succinctly:
OpenOffice the single worst piece of open source software available in any official repository. It is buggy, unfriendly, bloated, unreliable, and generally behaves in manners that do not make any sense. Crash and burn one boot, load fine the next.
This was the real success of Linux. It didn't take over the consumer desktop market, but it sure decimated the need for specialty servers and proprietary server operating systems.
Once Sun was forced to compete in the commodity hardware market, it was over.
Yes, Sun and Digital (DEC) got killed by the same thing - Clayton Christensen on Digital (not Sun):
Every disruption has three components to it: a technological enabler, a business model innovation and a new commercial ecosystem. In computing, the technological enabler of disruption in computing was the microprocessor. It so simplified the design of a computer that Steve Wozniak and Steve Jobs could just slap one together in a garage. It transformed the industry’s fundamental technological problem—the design of a computer—from a problem that took hundreds of people several years to solve into one that was much simpler.
Then that simplifying technology had to be married with a business model that could take the technology into the market in a cost-effective and convenient way. Digital Equipment Corp. had microprocessor technology, but its business model could not profitably sell a computer for less than $50,000. The technology trapped in a high-cost business model had no impact on the world, and in fact, the world ultimately killed Digital. But IBM Corp., with the very same processors at its disposal, set up a different business model in Florida that could make money at a $2,000 price point and 20% gross margins—and changed the world. It’s a combination of the technology and business model that makes formerly complicated, expensive, inaccessible things affordable and accessible.
Indeed, while Dell, et. al. just spent their time pumping out ever-cheaper products, Sun actually spent cash developing the aformentionted technologies.
Nice "real success": killing the innovators in the UNIX space and letting the vultures come in and eat up all the market share.
SPARC with Solaris is pretty inferior to Linux on x86 for many purposes, including the entire web-stack. This competition forced Sun to innovate, giving us the UltraSPARC T1 and T2, which is simply amazing technology. Suns UNIX innovation (i.e. Solaris 10 with dtrace and ZFS) came as a response to the success of Linux, not in spite of it.
Sun paused to sleep on their .com laurels, and it killed them.
Java was the reason for Oracle to take over Sun (OK, also the hardware). Since Oracles commitment to open source is, "yeah, sure, Oracle will run on Linux. ONE version of ONE distro, but, yeah, Linux", this is more likely than ever to trigger a community-fork of Java, backed by IBM.
There is no way Java is going anywhere. We can actually hope that Oracle loses some interest in Java, thereby loosening their grip and letting the language evolve.
By the way, I think JRockIt is the best JVM out there for server-side deployments. I've used both JRockIt and Sun's JVM on EC2, and JRockIt has much better memory management.
I'll miss Sun and the amazing people I knew there.