Slashdot Mirror

They could try out Solaris for x86 platforms.

Maybe its because Sun is giving away servers. For free. No cost. And each free server would add ... let me think ... ummm ... zero dollars to the total.
http://blogs.sun.com/roller/page/jonathan/20051218 /

Maybe not.

Only a true, dyed-in-the-wool, ignorant Microsoft sycophant could possibly make such a statement.

You've obviously never heard of Sun Solaris.

Even you can try it, for free, legally, but you won't since you can't see past Microsoft.

There are only a few things I'll give solaris credit for, one is stability (If you never want to update your system, patches for a standard solaris install come out 3-5 times a week or more.) and two is binary compatibility: http://www.sun.com/software/solaris/guarantee.jsp

The Network is the Computer(TM)

Doubly-linked lists have been in Java since 1.2.

Don't be confused by the name. It is doubly linked. Check the javadoc.

> That requires a heap-compacting allocator (aka a "handle" allocator).

Modern compacting allocators do not use handles.

Sun's Java implementation switched to handle-less collection in 1.2.x series (Exact VM):

http://research.sun.com/features/tenyears/volcd/pa pers/heller.htm

Sun's current Java runtime (HotSpot) uses handle-less collection:

http://java.sun.com/products/hotspot/docs/whitepap er/Java_HotSpot_WP_Final_4_30_01.html#pgfId=108223 8

> That requires a heap-compacting allocator (aka a "handle" allocator).

Modern compacting allocators do not use handles.

Sun's Java implementation switched to handle-less collection in 1.2.x series (Exact VM):

http://research.sun.com/features/tenyears/volcd/pa pers/heller.htm

Sun's current Java runtime (HotSpot) uses handle-less collection:

http://java.sun.com/products/hotspot/docs/whitepap er/Java_HotSpot_WP_Final_4_30_01.html#pgfId=108223 8

Mildly interesting but in terms of modern games it won't cut the mustard so then you are looking at smaller games and a platform to learn on. IMO there is only one top platform to learn game programming for and that is the Java mobile phone platform J2ME that even comes with a standard (and pretty simple) 3D API. Its also the fastest growing gaming market out there at the moment, and its still an area where mere mortals can break in and create the killer game and make money. Aiming at the PC market means you are against the multi-million dollar budget folks and that just isn't sensible.

If you want to learn game programming... get a decent phone.

How dare you slander my browser of choice

How on Earth did they port Solaris to the 2600?

Full forward binary compatibility across server OSes is very important. Sun has binary compatibility since Solaris 2.3, maybe earlier - meaning you can take a program (binary) that was compiled on Solaris 2.x and copy it to a Solaris 10 SPARC server (Ultra4, Ultra4+, or T1 based CPU) and run it unchanged successfully. See here: http://forum.sun.com/thread.jspa?threadID=27689&ts tart=75

There is a caveat about device drivers, but the vast majority of code **will** run fine. I'm not certain if IBM on RS6000, PowerX chips also is binary compatible, but I'm positive that HP-UX isn't after years of having to recompile old HPUX code to get it to run on the next major OS release.

Sun is currently shipping EIGHT core CPUS, and each core handles 4 threads... so you are talking 32 threads in one RU of space.

http://www.sun.com/servers/coolthreads/overview/in dex.jsp

This is silly. Microsoft made a conscious decision to license software per CPU (or per Slot) rather than core, and they had announced that they are doing so because multi-core looked like the natural way in which CPUs will improve, given that the Mhz war has ended. In fact they were the first major company to do so.

Also, this does not really eat into MS bottom line compared to Oracle or IBM. Most of MS revenue comes from the the desktop, while they are just competing in servers. Sql Server suddenly becomes more attractive, given Oracles complicated multi-core policy. (Remember that Oracle earlier announced that every core is a CPU, its just recently that they realized it will be a disaster and modified their original plans.)

Earlier CPU speeds doubled every 18 months. Multi-core will simply take another approach to achieving the same. I am not sure how this will hurt software companies any more than increasing cycles/sec.

Sun has been shipping 8 core T1 processors for a few months. T1000 and T2000 rack servers based on "CoolThreads" - whatever that is. See here: http://store.sun.com/CMTemplate/CEServlet?process= SunStore&cmdViewProduct_CP&catid=141649. IBM has QuadCore Power5+ chips shipping any day now.

Both Solaris and AIX scale over 100 CPUs already. Good luck AMD and Intel on getting Microsoft to create a standard OS (not their funny datacenter version) that is the same on 1 CPU or 124 CPU systems.

Why wait? Sun already makes processors with 8 cores. For realz.

Only 4 cores and due out in 2007!? If multi-core is the future then sun must be ahead of their time:

http://www.sun.com/processors/UltraSPARC-T1/index. xml/

This Aero Glass gui sounds an awful lot like Sun Microsystem's Looking Glass. Sun wrote a gui in java that's really cool. You can do all sorts of neat stuff w/ it. This Aero Glass sounds like that. They both eat up resources like crazy too. :-)

Couldn't agree more. Sun seem to be back on their game with the new T1 Niagara sparcs - 8 cores, yum!

http://www.sun.com/servers/index.jsp?cat=CoolThrea ds%20Servers&tab=3

My shop has recently taken a dozen T2000s in for testing - as a web farm they're outperforming everything we currently have by a country mile.

Fundamentally, though, as far as I can see, having threads introduces a whole new dimension of nondeterminism. This makes correctness and debugging more difficult. Much more, IMO.

As processors move to 8-way cores, 64-way cores, etc. (which will presumably eventually become available even for desktops), I might end up changing my mind about this. But right now it seems like hardware is cheap, developer time is expensive, and anything that lets me trade "less hardware" for "more developer time" doesn't seem very attractive.

If you've done a lot of non-threaded stuff and a lot of threaded stuff and the threading didn't cause you noticeably more grief, then you may definitely count yourself lucky.

Are you familiar with modern threading tools? These allow you to create, gather and hold your frequently accessed model data in RAM where it belongs, as VM objects without copy/parse requirements, and serve your user state machine changes while using the whole CPU capacity of your servers.

Yes, it's about time. Two years ago Sun tried this, then RedHat tried it and from what I can tell both quit. Novell deserves credit for getting it done.

I work as a system designer leveraging HP-PA-RISC, IBM, Sun, and HP-win32/64 servers. I primarily work in the UNIX space, but occassionally have to purchase Windows servers. Enough background.
Prior to the UltraSPARC 4+ and the T1 processor from Sun, the Power5 core was approximately equiv to 4 UltraSPARC4 CPUs. For example, a P550 http://www-03.ibm.com/systems/p/hardware/entry/550 q/index.html with 4 cores, is roughly equiv to a V1280 with 10 CPUs. When you add up the total cost of ownership, most software packages are licensed either by core or CPU count. 4 CPU licenses are always always less costly than 10 CPU licenses.
Now the Power5+ processors (which I've already purchased almost 30 servers), feel 2x faster than a Power4. The P550Q Sun's T1000/T2000 application servers are also very impressive, but they aren't general purpose and the workload needs to match what these servers are designed to handle or you will be disapointed. OTOH, a $8k list price with appsrv performance like a $50k box is impressive (assuming that is a true number). I haven't gotten any T1-based servers http://store.sun.com/CMTemplate/CEServlet?process= SunStore&cmdViewProduct_CP&catid=141649 yet.
These servers are too new for many benchmarks outside what each vender claims. I've found IBM and Sun to be truthful in their performance for nearly available systems. They have to be since it is too easy to see where they might have lied in just a few months by checking http://www.spec.org/

Sun introduced class data sharing between VMs in Java 1.5.

There's actually quite a bit of work at Sun that is doing exactly what you suggest.

This describes a system for running multiple Java apps in one virtual machine.
http://research.sun.com/projects/barcelona/papers/ oopsla01.pdf

However, there are safety issues with such an approach. If one application manages to crash the JVM, all the applications running on the shared JVM will stop. So other approaches were also investigated such as sharing memory between virtual machines

http://research.sun.com/projects/barcelona/papers/ ecoop02.pdf

and building Java classes into shared libraries that allow non-changing parts of the class representation in memory to be shared across JVMs.

http://research.sun.com/projects/barcelona/papers/ ipdps03.pdf

There's actually quite a bit of work at Sun that is doing exactly what you suggest.

This describes a system for running multiple Java apps in one virtual machine.
http://research.sun.com/projects/barcelona/papers/ oopsla01.pdf

However, there are safety issues with such an approach. If one application manages to crash the JVM, all the applications running on the shared JVM will stop. So other approaches were also investigated such as sharing memory between virtual machines

http://research.sun.com/projects/barcelona/papers/ ecoop02.pdf

and building Java classes into shared libraries that allow non-changing parts of the class representation in memory to be shared across JVMs.

http://research.sun.com/projects/barcelona/papers/ ipdps03.pdf

There's actually quite a bit of work at Sun that is doing exactly what you suggest.

This describes a system for running multiple Java apps in one virtual machine.
http://research.sun.com/projects/barcelona/papers/ oopsla01.pdf

However, there are safety issues with such an approach. If one application manages to crash the JVM, all the applications running on the shared JVM will stop. So other approaches were also investigated such as sharing memory between virtual machines

http://research.sun.com/projects/barcelona/papers/ ecoop02.pdf

and building Java classes into shared libraries that allow non-changing parts of the class representation in memory to be shared across JVMs.

http://research.sun.com/projects/barcelona/papers/ ipdps03.pdf

Sun not to long ago announced all of their server software is free (as in beer for now) and then you can pay for support.

Their messaging server does shared contacts, appointments, to-do lists, has and Outlook connector and even ties in instant messaging.
http://www.sun.com/software/products/messaging_srv r/index.xml

I have no idea if this works on a small scale, but I'm going to try it out once I get some free time. Googling for anyone that has tried it gets me know where.

That might happen. It would be interesting if that substitute kernel turned out to be Solaris.

You can hack it in by making a user for every program, but I don't think that should be necessary.

Why not? It doesn't harm the system.

Besides, who'd want to do that?

For example the install scripts or the OS itself.

Such as, read only access to itself and its own directory, write access to an incoming files directory, read/write to its own registry key and the temp folder, outgoing network access to tcp port 80, and that's it. It shouldn't be allowed to spawn child processes, read system information like hostname, username, usergroup, OS version, and it can't because all those permissions were absent from iexplore.acl (but can of course be assigned).

You might be interested in Java then. By the virtue of the Security Manager, you can in fact define for every Java app you run what it is allowed to, and what it is not allowed to do.

For example, using the FilePermission you can restrict the app's read or write access to certain directories only. Analogously, using the SocketPermission you can restrict network access of that program to certain hosts or certain ports only. With an application-specific policy file, you can control the rights of this app extremely fine-grainedly.

Better yet, you can try it for yourself:

https://mustang.dev.java.net/

The sourcecode is also available under a non-contamination agreement.

The Desktop features of Mustang are documented here.

At least Sun knows about what risks it takes to administrate a Sun Java webserver :-)

Btw, sun have a "correct" java grammar on there site:

http://java.sun.com/docs/books/jls/second_edition/ html/syntax.doc.html

(I however think thats an old version of Java. found it with google.)

/ Balp

Hmmm I wonder why i don't think Java when Im thinking Ajax applications.

HAVE YOU SEEN Suns own tutorials for whats involved in writing simple ajax applications in Java??
http://java.sun.com/developer/technicalArticles/J2 EE/AJAX/

Compare that with a something from Ruby, that does a lot more with a lot less code, configurations and time:
http://media.rubyonrails.org/video/flickr-rails-aj ax.mov

Now I wonder why small Web 2.0 companies aren't flocking to Java to power their Web 2.0 apps.

They exist?

Depends... do they mean "in the Sun(TM)"?

It seems Sun is coming out with java Server Faces as an asnwer to rad and php. Have any of you guys tried it?

I am surprised it has no mention here on slashdot so far but most new young slashdot geeks are not interested in java. What a difference a decade makes?

This is why Java Server Faces are taking over to streamline Servlet development. You can include custom tags right in html like php from what I heard. I have not used them yet but they look promising.

"There is only ONE computer that could possibly handle these calculations."

Just remember, if it asks you if you want to play a game, SAY NO!

> There is only ONE computer that could possibly handle these calculations.

WOPR? A mere abacus. Mention it not.

There is only ONE computer that could possibly handle these calculations.

I have an idea, if your fortune gets to be so large that even the IRS can't figure it all out, you should be required to give some of it away to the poor until they can do the necessary calculations.
That being said, I will accept cash and postal money orders only please.

x86 Solaris is Solaris. It's the same tree.

http://www.sun.com/software/solaris/opensolaris.js p - "both SPARC and x64/x86 (it's a single source base)"

Granted, an 'interesting' use of the word 'both' when listing 3 platforms, but Solaris is Solaris, regardless of platform.

Speaking of zealots, wasn't Jonathan Schwartz ranting about how the GPL sucked at some keynote address last year? Oh wait! He did! He even wrote about on its weblog:

Back in the Solaris 5 timeframe there was a PowerPC version available. The idea was that it was going to be available on the IBM/Apple/Motorolla 'Refrence Platform'.

Although there is no such thing as Solaris 5, I can confirm that a PowerPC port of Solaris existed. It was Solaris 2.5 that was available for SPARC, PowerPC, and x86. Here's some of Sun's documentation about it.

I also remember seeing PowerPC versions of patches on the SunSolve contract support site for Solaris, so I can only assume based on this that they not only made it but actually sold it to customers. (If not, why provide supported patches?)

Of course, at the time, so many industry players were getting behind the PowerPC (such as IBM, Apple, Sun, and Motorola) that lots of people thought it might unseat x86. Especially since x86 performance had stagnated a bit at that time. So lots of people ported to PowerPC as a bit of insurance, just like they ported to Itanium for insurance a few years later.

Read the blog post. The author specifically specifies the GPLv3, and not just "the GPL." The blog title is "Thinking About GPL3..." and he links to a copy of the GPLv3 draft. (Which actually says "THIS IS A DRAFT" right on it, don't know why he linked a copy and not the GPLv3 site, but...)

So, yes - he's talking about using the GPLv3 as opposed to the current GPL.

Which is silly, because the GPLv3 is still in draft form. It's not released. Speculation about how to apply the GPLv3 would make sense, talking about actually releasing something under it is rather early. Unless he intentional means to say "we have no plans on releasing an open source version of Solaris until 2007."

He stores his "home collection" on an A5200?!

Now that's what I call hardcore porn.

Slow low power processors with multiple cores handling multiple processes:

http://www.sun.com/servers/coolthreads/overview/in dex.jsp

CPUs with six or eight cores, with four threads per core. Sun says their new CoolThreads Servers offer significant power, cooling, and space savings.

I believe the servers are too new for anyone to have a solid opinion about, but I know Sun has been actively moving in this direction for a while.

Not only does anecdotal evidence point to that not being a true statement... as I work, programming, at a company of no more than 20 people ( there is one other programmer ), and I know of at least two other less-than-30-people companies that are all-Linux... and several small ( some very small ) businesses that are mostly Apple... actually, from my experience, *most* non-MS stuff happens with smaller companies. Big and mid-size companies have IT departments that tend to standardize on site-licensed MS stuff, and don't want to or don't feel they can support multiple platforms. Smaller companies have the flexibility to adopt 'alternative' systems, and often seek to minimize costs ( including patching and support )... and sometimes they decide that means not being a Microsoft shop.

Clients come to me. They don't need or want non-Microsoft solutions.

Do you have lots of experience with non-Microsoft stuff ? Do you advertise that? Or do you tout Microsoft certifications and get references from clients that are Microsoft sites ? Why would a client come to you for a non-Microsoft solution, as opposed to someone advertising OSS, *nix, or Apple consulting services ? Maybe you don't get non-Microsoft clients because you don't attract them? Maybe your clients would consider a non-MS solution if you offered it?

"Still, there are companies that use Java as it was intended- to write cross-platform apps that can be deployed under Linux, Solaris, Windows, or OS X, all using the same bytecode package and one unified codebase."

Now see, right up until then I thought you were really a developer. No self respecting developer would spout that completely marketing made up crap in the last 2 years. It has been disproven so many time not even sun says it anymore.

As far as Java UI interfaces on commercial products, I was actually thinking directly of one example I'd seen recently , but I know for a fact that there are many, many others. Funny enough, I checked out that company, and they're -get this- a fairly small, nearly all-Microsoft shop.

Of course, some really high-quality Java distributions include ( often minor ) code tweaks that make things work a bit more smoothly on different platforms... but really, they're usually minor and there are packages that aid in implementing those platform-specific things when they're deemed important, so a unified codebase, even in a UI app, isn't just a myth. Funny thing is, right up until you made that comment, I thought you might have some idea what you're talking about... but clearly, you're one of those programmers who thinks they know Java, but really can't hack it. That, or you made up your mind about Java years ago, when JVM bugs were much more common. Today? Sure, you *can* write an app that doesn't work correctly on multiple platforms, but it's not common. Actually, I'd argue that the past 2 years ( especially for OS X/Windows compatibility ) have produced the best "write once, run anywhere" JVMs in history.

True, Sun doesn't use "write once, run anywhere" as the main selling point of Java anymore ( since they learned people beat them about the head with JVM bugs when they say that, and that cross-platform code is just one selling points of Java ), but to say they don't use that phrase anymore just shows how little you know about Sun's marketing and products.

Go ahead, belittle Java as much as you want, tell yourself it's not a viable platform, pretend it's an all-Microsoft world... the number of positions posted on job boards looking for Java programmers ( including those looking specifically for GUI programmers ) as well as Linux programmers seems to indicate that your opinion on the matter might not be based in fact or reflected in reality.

Itanium has been taking share from both IBM power and Sun Sparc.

True but can they compete with the UltraSparc T1 (which has 32 threads compared to Intel Itanium's 2 threads)?