Follow Slashdot stories on Twitter

 



Forgot your password?
typodupeerror
×
Sun Microsystems Entertainment Games Science

Sun Surges Into Research, Virtual Worlds 56

An anonymous reader writes "Sun Microsystems appears to be shifting its focus back to research, after several years of promoting its commodity servers and Java software. Earlier this week, it talked about its new Andy Bechtolsheim-designed video server in the New York Times. Yesterday, it invited reporters in to preview its plans to develop faster switches, new programming languages, and 3-D virtual workplaces. Robert Sproull, director of Sun Labs, made clear that Sun has big ambitions. 'General purpose computers have to be rethought,' he said. Among the projects close to leaving the labs is Project Crossbow, an evolution of the networking stack in Solaris; Project Sedna, a next generation switch for storage-area networks; and MPK20, a virtual workspace built on top of Sun's Darkstar gaming server."
This discussion has been archived. No new comments can be posted.

Sun Surges Into Research, Virtual Worlds

Comments Filter:
  • by Palmyst ( 1065142 ) on Friday April 27, 2007 @01:02PM (#18903921)
    From my experience, the FORTRAN community is the most resistant among programmers to switch languages. Even F99 hasn't got much traction with them. So, best of luck with Fortress.
    • by Timesprout ( 579035 ) on Friday April 27, 2007 @01:12PM (#18904143)
      Its easy. Sun plan to send samples of FORTRAN code to the police saying it scares them and all FORTRAN coders will be arrested for writing such disturbed text. No more FORTRAN community and a greenfield for SUN's new language. Problem solved.
      • The result is a server system capable of permitting hundreds of thousands of computer users to interact in a three-dimensional simulated on-screen world described as a "metaverse."

        Rumor has it Sun's designed this amazing new USB interface for this system: thought-activated, bio-compatible, and self-sustaining, provided you don't mind floating in goo [dictionary-of-matrix.com]

      • This might work in Boston, but what about the rest of the country?
    • Re: (Score:2, Interesting)

      by Framboise ( 521772 )
      On the contrary the Fortran user community has been using an evolving language that allows code reuse over 50 years. Only the outsiders have a frozen and outdated opinion about what is Fortran today. Heh F99 doesn't exist. Fortress seems an attracting language for Fortran users because it allows to express algorithms in a way close to what mathematicians do since over a century. For example by using Unicode Fortress has finally a charatcter set matching the ones used by scientists.
      • by UtucXul ( 658400 ) on Friday April 27, 2007 @01:32PM (#18904503) Homepage

        For example by using Unicode Fortress has finally a charatcter set matching the ones used by scientists.
        I do a lot of my work in FORTRAN 77, but I'm interested in programming languages, and tend to switch my smaller tools around from one language to the next (currently I like Python a lot). So I probably fall right into the class of people who Fortress is interesting to. But the idea of Unicode using programming languages seems like a really bad idea for a language that is trying to replace a numerical workhorse. The last time I looked at Fortress, I seem to remember that if you don't use a unicode aware editor, there was some LaTeX-like way to input math also, but even that seems a little heavy weight for numerical programming.

        I guess what it comes down to is that of all the failing of FORTRAN, the fact that its math is less pretty than LaTeX does not seem like an important one. I know Fortress has some other features, but the whole pretty math character thing seems to be the one that comes up most.

        • Actually, Fortress seems, in many ways, a nice piece of work. Your comment is well taken, but it does the effort a bit of a disservice to help people obsess about what is frankly a minor detail (and honestly I'm not persuaded Unicode support can be classed as a bug). Here, let's tell the /. community something exciting about the language: like the fact that its loops are parallel and distributable by default, and it is the iterator that determines the loop's serialisation pattern, if any.

          Seriously, languag
          • by UtucXul ( 658400 )
            I agree, cosmetics are not the only thing that matters in a language (after all, I use enough FORTRAN 77 that it would be sort of perverse for me to care only about a language being pretty).

            All that parallel loop stuff is potentially very interesting. I work mainly with a parallel MHD code. If we could get rid of all or most of the MPI statements from the code it would dramatically simplify the code. I spent most of Monday hunting down a stupid little MPI bug.

            So, from my perspective, Fortress is main

        • Re: (Score:3, Insightful)

          by Coryoth ( 254751 )
          I think you'll find that the Unicode character set is the one that gets commented on the most because that's the only feature you can glean from a supreficial skim of the quite long and detailed spec. If, on the other hand, you actually read the spec you'll find a lot of other very nice features, good concurrency control, software transactional memory, a very nice component system, an interesting parametric polymorphism system, some good functional programming primitives, and more. It is worth actually read
          • by UtucXul ( 658400 )
            I have to admit, you sort of got me there. I downloaded the spec. I even started reading it. I could tell there was some potentially good stuff in there, but I didn't get through that much of it before moving on to skimming.

            Although in my defense, the fact that when I read it, the closest thing there was to a compiler for it was an implementation of it that runs over the java virtual machine. And while that may be a good thing for language research, it isn't all that attractive from a numerical perfor

            • by Coryoth ( 254751 )
              Fortress won't be hitting prime-time anytime soon. They are still working with a demo implementation for testing purposes (it doesn't even have all the features from the spec working -- contracts weren't available yet last time I played with it). Still, the language design looks pretty solid, and when they eventually come to building a final product I think it will be well wort investing time in if you do much numerical work. That means it may be worth investing a little time now familiarising yourself with
  • Obvious, -1 (Score:4, Funny)

    by bricriu ( 184334 ) on Friday April 27, 2007 @01:11PM (#18904103) Homepage
    Anyone who comments on how their virtual workspaces could be called "DarkSun" will be shot with a railgun. That is all.
    • If Sun is into development and building its brand on it, they should make an open-source video card. For Sun and pc platforms, solaris and linux of course. And most likely a professional and gaming version.
      • Interesting but Sun hasn't been a graphics company for about 10 years. All their
        graphics people now work at nVidia.

        Also, Darkstar is has nothing to do with graphics. It's a transaction server.
        Someone like WoW or Second Life could replace all their backend servers with Darkstar
        to make one contiguous 'world' instead of shards and separate servers for players.
  • Jonathan Schwartz (Score:5, Interesting)

    by alienmole ( 15522 ) on Friday April 27, 2007 @01:13PM (#18904153)
    Sun's new CEO [wsj.com] is the driving force behind this. Quite a change from Scott McNealy.
    • No surprise they are starting to focus on doing something constructive. McNealy was way too obsessed with fighting Microsoft to give SUN any solid direction.
      • by mandelbr0t ( 1015855 ) on Friday April 27, 2007 @01:40PM (#18904665) Journal
        Indeed. My experience with Sun vs. Microsoft politics is that Sun was clearly willing to sink to Microsoft's level. They played heavily with vendor lock-in, trashing of un-(Sun)-certified techies and various other tactics that I had only really associated with Microsoft. I found these politics to be most like something you'd hear on a primary school playground.

        "You don't want to switch to a .NET deployment. Java programmers are only slightly more cultured than cavemen, and Solaris SysAdmins are known to hang out at Furry parties. Besides, where are you going to find parts for all those Ultra-2s in your basement?"

        "Oh yeah? Well James Gosling is a poo-poo head!"

        No one wins in these dealings except Microsoft. Let me jump on the "good to see Sun doing something constructive" bandwagon.
      • Arrgh. I suck at being funny. I got the trash-talking backwards :(
        • It works as a demonstration of just how ineffective they were, though. I think your point got through.
    • I agree (Score:3, Insightful)

      by WindBourne ( 631190 )
      I have been quick to point how much damage that McNealy has done to Sun (while the fan boys defend him and has lousy last 5 years). Schwartz is the guy who might just bring Sun back again. He still has a LONG ways to go, but at least he is no longer lying and playing costly games such as funding SCO against Linux (huge waste of good will for Sun) or going after MS on Java. At one time, McNealy was good, but that was when the industry was much smaller and they had lots of room to maneuver. Now that they are
  • by Anonymous Coward
    I still haven't received my Solaris 10 DVD.
  • The Sun? Appearing in virtual worlds? I knew it! There's no escape from the daystar! We're all doomed! Dooooooooooooomed!
  • Lawsuits... (Score:3, Interesting)

    by ragtop70 ( 1094351 ) on Friday April 27, 2007 @02:29PM (#18905557)
    Sun should watch out by naming anything Darkstar... As the former owner of darkstar-reserarch.com, I was approached by someone (forgot the name, but then again, that's also being polite) from Illinois who claims to own every possible use (patents and copyrights) of the terms "Darkstar" and "Stealth" and provided documentation to the effect that he had forced Fortune 100 companies to stop using either term when referring to any product if they had not licensed the use of the term from him. This guy had nothing to gain by pursuing legal action against me (a hobbyist who simply wanted a domain name for e-mail and personal web hosting), but he has nothing to lose, really, by bringing legal action against Sun.
    • I assume that he had trademarks too, or you were had!
      • I looked it up through the appropriate channels, did my research and realized that if Fortune 100 companies weren't willing to stand up to this guy, that I might as well... so, to that end, I sent him a reply stating that it was not for commercial purposes, that I would in no way, shape or form use the domain for commercial purposes and that if he wanted to pursue legal action that I would see him in court. There was no reply forthcoming from him and I let the domain expire.
  • It's that simple. Please post more news about things they'll never bring to market. They are folding in on themselves because they forgot what servers are supposed to do. And that is to SERVE PEOPLE. Thanks for playing Sun. You were rich while it lasted.
  • In a SAN, network performance is paramount. So who would buy a SAN product named for a rock that takes a long time to orbit the sun?
  • I'm wondering if darkstar is a way for Sun to make it's Sun Grid Utility profitable.
    Darkstar would be one way of showcasing how their Grid could be used to benefit developers
    of server intensive application.
  • "We'll need to do a meeting with Bob in accounting. He's the giant floating eyeball with tentacles."
  • by ChrisGilliard ( 913445 ) <christopher.gill ... m minus language> on Saturday April 28, 2007 @12:07AM (#18909405) Homepage
    MPK stands for Menlo Park (where Sun's headquarter's is located). There are 19 buildings on the Sun Menlo Park campus. Hence, MKP20 is the vitual building.
  • Here comes the second coming of parallel program. This is gonna be cool.

Every nonzero finite dimensional inner product space has an orthonormal basis. It makes sense, when you don't think about it.

Working...