Slashdot Mirror

Actually, containers don't require Sun hardware to run on - they're available on any system that Solaris 10 runs on. From a networking performance standpoint, a server having multiple 1Gb or 10Gb network interfaces is going to need a highly multithreaded stack in order to drive those interfaces. Solaris 10's TCP/IP stack is optimized for those sorts of situations.

Actually Fortress has probably more mechanisms to support high-performance, programmer-controllable parallel computation than any other current language. Unfortunately they haven't published any good papers about the details yet, but slides 21 and thereafter of this Guy Steele lecture http://research.sun.com/projects/plrg/GSPx-Lecture 2006public.pdf make it very clear that Fortress does at least the following:

- Provide an abstract definition of data structures.
- Define reductions and computations over those data structures in an abstract manner.
- Allow specification of concrete mappings to particular processor and network topologies, such that those computations can be efficiently scaled to the local system.

In some ways it reminds me of Google's MapReduce system, but applied at the language level, in a way that allows library authors to create their own data structures and computational mappings in a fully extensible way. Super cool stuff, not only groundbreaking in the computer language world, but also potentially revolutionary in the scientific computing world.

And so what if their interpreter is written in Java? That's like writing off Java because the very first Java engines were bytecode interpreters! You walk before you run, and you interpret before you compile, when building up a new language from absolute scratch.

Also, they put a lot of emphasis on interoperability with Java, so there should be no problem using existing Java codebases to build UIs, etc. for Fortress apps.

Here's the actual summary from the Steele presentation:

- Regions describe machine resources.
- Distributions map aggregates onto regions.
- Aggregates used as generators drive parallelism.
- Algebraic properties drive implementation strategies.
- Algebraic properties are described by traits.
- Properties are verified by automated unit testing.
- Traits allow sharing of code, properties, and test data.
- Reducers and generators negotiate through overloaded method dispatch keyed by traits to achieve mix-and-match parallel code selection.

It's worth noting that Sun's Fortress project was not selected for Phase III of DARPA's HPCS project. (And for good measure, a link to a blog at Sun and an FAQ on how Fortress relates to the other HPCS languages/projects.)

It's worth noting that Sun's Fortress project was not selected for Phase III of DARPA's HPCS project. (And for good measure, a link to a blog at Sun and an FAQ on how Fortress relates to the other HPCS languages/projects.)

I heard his talk at OOPSLA 2006 on the language design decisions they made for Fortress, and although my Fortran (and math) experience is too shallow to fully appreciate it, I found it fascinating nonetheless.

From The Fortress Language Specification, version 1.0alpha:

component HelloWorld
export Executable
run(args) = print "Hello, world!"
end

Personally, I find functional notation and names much easier to understand than mathematical notation and symbols. Of course, I'm not a mathematician, so I guess I'm not the target audience for this project. However, I still think this is a really bad idea.

Sorry, but this loser doesn't even have a beard.

http://research.sun.com/projects/plrg/PLDITutorial Slides9Jun2006.pdf

Fortress uses a lot of unicode mathematical operators, which slashdot will quite pitifully fail to display.

This one looks like a winner.

This one looks like a winner.

(I'm referring here to the current release of the supported Solaris 10 product. As others have noted, there's also OpenSolaris, which available from http://www.opensolaris.org/os, is well ahead of the mainstream release and provides source code, but is a somewhat less complete product bundle. Also, Solaris Express, which is a snapshot of what will become the Solaris 11 release, which like the supported product can be downloaded from Sun's website.)

x64/x86 DVD ---- non sun hardware install set

They are calling it "free and open source" but I guess they mean free as in beer because the free software foundation does not give its stamp of approval for Solaris being GPL.

http://www.sun.com/software/solaris/

I order this just before Christmas and the version that was shipped to me was Solaris 10 06/06. 11/06 was released in November, so I wonder if this is just a way of clearing out the old media they'd otherwise be disposing of?

It's perhaps worth noting that to upgrade from a previous release of Solaris 10 requires the full media if some of the features of 11/06 grab you (http://docs.sun.com/app/docs/doc/817-0547/6mgbdbs mb?a=view)

Is there anything in particular that makes Solaris 10 stand out from say FreeBSD or various flavors of Linux?

I'm not seeing the "news" angle since Solaris 10 and Sun Studio 11 have been available as free downloads for quite some time.

He moved the sun. Impressive.

Impressive? Not necessarily. It depends on which sun he's talking about. They range from 0.37kg to 1122kg.

He moved the sun. Impressive.

Impressive? Not necessarily. It depends on which sun he's talking about. They range from 0.37kg to 1122kg.

Appearance modes are "a hint to the platform". IBM are free to ignore it, but I'm quite surprised if they do with BUTTON. It's used as an example in the documentation:

http://java.sun.com/javame/reference/apis/jsr118/j avax/microedition/lcdui/StringItem.html

Java ME has a tough task (especially the UI parts) being "write once run anywhere" since it runs on top of a lot of proprietary platforms. I think it has done an excellent work with that in mind.

(I've spent several years implementing MIDP on top of such a platform)

Why not get a Sun Certified Business Component Developer Certification instead?

No, that'd avoid confusion with Sun's KVM: http://java.sun.com/products/cldc/wp/

Sure. And the very "design" of AJAX encourages such poor development to occur. The fact that virtually every AJAX application is problematic shows that the problem is not with the developers, but with the technologies those developers are trying to use.

Err... what? Are you saying that when I go to this-magical-php-script.com and download a bunch of PHP apps made by $0M3.Kr@zY.GuY and they don't work that PHP is a design flawed language?

What's your point? That's exactly how Java applications work: poorly. Java applications are often just as problematic as AJAX applications, in most cases.

Are you saying Java apps work poorly? I can understand you commenting on how AJAX apps are, on the whole, not working out very well as you have the initial "Holy crap new hype" bloom if junky software developers trying to make a few bucks while people don't understand the language. What I can't understand is how you can say the Java apps work poorly. Have you run any Java apps on your systems - from real vendors? I'm not talking about "Pizza Joes Amazing Taxes Pro Java Edition" but something more... interesting: http://www.sun.com/software/index.jsp

I just got to the part of your post where you say that you're 16

Just because someone's young doesn't mean that they're stupid. When I was 19, coming out of College, I did placement where I had to roll out a Windows 2000 upgrade in an existing RedHat + Windows NT + Solaris setup. Was I young and I did the job - reading does wonders to expand on what you don't know. Now they have a lot more resources online you can pick through for answers.

In the end, Mr. Anonymous, you should be giving helpful remarks to the rest of us as to where we shouldn't be going for our apps. Giving vague examples of this-and-that app didn't work, this-and-that old Perl app was amazing doesn't help anyone. Give us some meat to chew on, not empty angry remarks.

"Why ZFS for home": - http://uadmin.blogspot.com/2006/05/why-zfs-for-hom e.html

"Here are ten reasons why you'll want to reformat all of your systems and use ZFS.": http://www.tech-recipes.com/rx/1446/zfs_ten_reason s_to_reformat_your_...

And some more technical explanations from Sun's Chief Engineer: - http://blogs.sun.com/bonwick/entry/zfs_end_to_end_ data - http://blogs.sun.com/bonwick/entry/smokin_mirrors

"Why ZFS for home": - http://uadmin.blogspot.com/2006/05/why-zfs-for-hom e.html

"Here are ten reasons why you'll want to reformat all of your systems and use ZFS.": http://www.tech-recipes.com/rx/1446/zfs_ten_reason s_to_reformat_your_...

And some more technical explanations from Sun's Chief Engineer: - http://blogs.sun.com/bonwick/entry/zfs_end_to_end_ data - http://blogs.sun.com/bonwick/entry/smokin_mirrors

"Why ZFS for home": - http://uadmin.blogspot.com/2006/05/why-zfs-for-hom e.html

"Here are ten reasons why you'll want to reformat all of your systems and use ZFS.": http://www.tech-recipes.com/rx/1446/zfs_ten_reason s_to_reformat_your_...

And some more technical explanations from Sun's Chief Engineer: - http://blogs.sun.com/bonwick/entry/zfs_end_to_end_ data - http://blogs.sun.com/bonwick/entry/smokin_mirrors

"Why ZFS for home": - http://uadmin.blogspot.com/2006/05/why-zfs-for-hom e.html

"Here are ten reasons why you'll want to reformat all of your systems and use ZFS.": http://www.tech-recipes.com/rx/1446/zfs_ten_reason s_to_reformat_your_...

And some more technical explanations from Sun's Chief Engineer: - http://blogs.sun.com/bonwick/entry/zfs_end_to_end_ data - http://blogs.sun.com/bonwick/entry/smokin_mirrors

Here you go.

http://sunsolve.sun.com/handbook_pub/Systems/C2/C2 .html
http://sunsolve.sun.com/handbook_pub/Systems/L8/L8 .html

Now, whether or not the home user will be able to afford one of the damned things is another issue :) but one of these bad boys fully loaded will back up that drive.

Although you were being a smart ass -- and I can appreciate that :) -- you do bring up an interesting question. With drives increasing so rapidly and for such inexpensive prices, you'd think that the tape drive manufacturers would be scrambling to keep up and make appropriate backup solutions more affordable for the home user. I don't mind using a mirror to keep the data redundant, but I'd still feel more comfortable having a mirror and a tape backup.

Here you go.

http://sunsolve.sun.com/handbook_pub/Systems/C2/C2 .html
http://sunsolve.sun.com/handbook_pub/Systems/L8/L8 .html

Now, whether or not the home user will be able to afford one of the damned things is another issue :) but one of these bad boys fully loaded will back up that drive.

Although you were being a smart ass -- and I can appreciate that :) -- you do bring up an interesting question. With drives increasing so rapidly and for such inexpensive prices, you'd think that the tape drive manufacturers would be scrambling to keep up and make appropriate backup solutions more affordable for the home user. I don't mind using a mirror to keep the data redundant, but I'd still feel more comfortable having a mirror and a tape backup.

http://www.sun.com/software/star/staroffice/index. jsp

Listen, some people see free software in a negative light: when something goes wrong, who are they going to complain to, a mailing list? OpenOffice may become the standard office tool someday, but not everyone (and I would think the majority: I use linux 24/7, but if I was responsible for a base of OTHER people, I'm not sure I would want to abandon a company funded $3000 existing Office suite for OpenOffice, a tool which I have NOT had years of experience in). Perhaps switchting to StarOffice would have people in less of a panic, especially if they needed help.

Strangers will adopt to OO when more people are using it. This will probably happen quicker in places like France, where the Government is pushing Open Source migration first (it leads to a perception of reliability - that matters to people more than cost, and that takes time). Fortunately, Star Office already has a decent industry reputation. I've used it on campus - it's decent.

Additional details regarding minimum JVM versions required to address the timezone changes:

- Solaris/Windows/Linux: 1.4.2_11
- HP-UX: 1.4.2.11
- AIX: 1.4.2 SR 5 20060420

- Solaris/Windows/Linux: 1.5.0_06
- HP-UX: 5.0.3
- AIX: 5.0 SR 1 20060310

References:
HP: http://www.hp.com/products1/unix/java/DST-US.html
Solaris/Windows/Linux: http://java.sun.com/developer/technicalArticles/In tl/USDST/
AIX : http://www-1.ibm.com/support/docview.wss?uid=swg21 232128

I have a bunch of stupid cobbled together setups to talk about. It all comes from a combination of poor IT staff at university wages, infintessimal budgets and the overbearing institutional and faculty pressures.

1. A "server room" that was essentially the most worthless room in the entire building, a long skinny room with four windows (perfect for keeping an uneven temperature!). Rather than buy 19" racks or even wire racks, they found a bunch of tables and put one server on each all the way around the edge of the room.
1.a. All of the servers were in fact desktop systems; an Ultra 1 was the mail server, a SPARCstation 5 the print server, a Gateway Pentium Pro 200 desktop the web server, etc.

2. A lab had to be moved one room over, because its current location was deemed too valuable. The original room was designed for a lab, it had 20+ fiber optic networking ports, twist-lock power connections in the ceiling, that sort of thing. The new room had two electrical outlets, no dropped ceiling, and one fiber optic networking port. It had previously been used as a copy room/storage closet. The cost to move the fiber optic wiring (just one room over mind you!) was over $25,000.

So instead, I had the great idea to cut a hole in the common wall (above the drop ceiling line), purchase additional ceiling tiles and cut up 2x4's into wooden supports. The original ceiling boxes containing the networking were put on top of the blocks above the new tiles, and extension cables run through the wall into the new room. In the original room, which was turned into a lounge, you couldn't tell that there was anything funny going on.

The best part is that the lab manager, who insisted they needed every single network port, never used a single one of them in the new room. All of those cables now reside in a box marked "Giant waste of money".

3. The main Windows file server was purchased in 2002 and has an internal RAID (bad idea in my opinion). What was huge then is worthless now; 5 disks that total 135GB. To get more space, the administration begged for a single external 250GB USB drive to host all user data. Nevermind that there is no redundancy, that an external drive is more suspectible to theft or failure, and that USB is unnecessarily slowing things down.

4. A system administrator got it into his head that rackmounting was the way to go (I agree). So he begged for a 19" rack to be ordered, and placed all of his servers into it. Except he doesn't have a single rack mountable server, and he didn't get the rails for any of the cases either. So now he has one $500 rack, and 8 $100 shelves to go in it. Same guy also switched the KVM monitor to a 15" LCD that doesn't support the resolutions of 9 out of 10 systems connected to it.

5. A consultant was brought in to tell us what needed to be done with the computing infrastructure (what DOESN'T need to be done is more the question). His main suggestion was to set up a central backup service just for this college, so as to avoid paying the central university IT group fees to use their central service. OK, thats an idea I guess... except that he wanted us to buy this: http://www.sun.com/storagetek/tape_storage/tape_li braries/sl8500/ (its $200,000). Luckily this one didn't actually come to pass.

Basically every day is a new adventure in ridiculous IT methodology.

With respect (this time) I submit that due to the long and consistent history of applet sandbox bugs in all vendors JVM's for the last many years, that the architecture for CLIENT side APPLET Java IS fatally flawed. You think corporate america is using Applets for highly secure enterprise applets? No way! Its fundamentally flawed when one call to System.setSecurityManager(null) totally wipes the entire sandbox for all applets running in a JVM.

Heck, they called setSecurityManager(null) a BUG for NOT WORKING back in 1997 and they FIXED it: http://bugs.sun.com/bugdatabase/view_bug.do?bug_id =4034420

With respect, keep away from Applets if you care about security.

OMG you are smoking Java crack there boy. Client side Java has more vulnerabilities than... Javascript. I love Java, but keep it on the server where it belongs. MySpace is getting ready to consider migrating from .NET to Java, it's solid on the server. But on the client... nope.

Take this from the LAST sunsolve weekly report:

Newly Released Sun Alert Notifications

Sun Alert ID: 102729 (RESOLVED)
Synopsis: Security Vulnerabilities in the Java Runtime
                              Environment may Allow Untrusted Applets to Elevate
                              Privileges and Execute Arbitrary Code
Product: Java 2 Platform, Standard Edition
Category: Security
Date Released: 19-Dec-2006
Date Closed: 19-Dec-2006

To view this Sun Alert document please go to the following URL:
http://sunsolve.sun.com/search/document.do?assetke y=1-26-102729-1

Sun Alert ID: 102731 (RESOLVED)
Synopsis: Security Vulnerabilities Related to Serialization
                              in the Java Runtime Environment may Allow Untrusted
                              Applets to Elevate Privileges
Product: Java 2 Platform, Standard Edition
Category: Security
Date Released: 19-Dec-2006
Date Closed: 19-Dec-2006

To view this Sun Alert document please go to the following URL:
http://sunsolve.sun.com/search/document.do?assetke y=1-26-102731-1

Sun Alert ID: 102732 (RESOLVED)
Synopsis: Security Vulnerabilities in the Java Runtime
                              Environment may Allow an Untrusted Applet to Access
                              Data in Other Applets
Product: Java 2 Platform, Standard Edition
Category: Security
Date Released: 19-Dec-2006
Date Closed: 19-Dec-2006

To view this Sun Alert document please go to the following URL:
http://sunsolve.sun.com/search/document.do?assetke y=1-26-102732-1

OMG you are smoking Java crack there boy. Client side Java has more vulnerabilities than... Javascript. I love Java, but keep it on the server where it belongs. MySpace is getting ready to consider migrating from .NET to Java, it's solid on the server. But on the client... nope.

Take this from the LAST sunsolve weekly report:

Newly Released Sun Alert Notifications

Sun Alert ID: 102729 (RESOLVED)
Synopsis: Security Vulnerabilities in the Java Runtime
                              Environment may Allow Untrusted Applets to Elevate
                              Privileges and Execute Arbitrary Code
Product: Java 2 Platform, Standard Edition
Category: Security
Date Released: 19-Dec-2006
Date Closed: 19-Dec-2006

To view this Sun Alert document please go to the following URL:
http://sunsolve.sun.com/search/document.do?assetke y=1-26-102729-1

Sun Alert ID: 102731 (RESOLVED)
Synopsis: Security Vulnerabilities Related to Serialization
                              in the Java Runtime Environment may Allow Untrusted
                              Applets to Elevate Privileges
Product: Java 2 Platform, Standard Edition
Category: Security
Date Released: 19-Dec-2006
Date Closed: 19-Dec-2006

To view this Sun Alert document please go to the following URL:
http://sunsolve.sun.com/search/document.do?assetke y=1-26-102731-1

Sun Alert ID: 102732 (RESOLVED)
Synopsis: Security Vulnerabilities in the Java Runtime
                              Environment may Allow an Untrusted Applet to Access
                              Data in Other Applets
Product: Java 2 Platform, Standard Edition
Category: Security
Date Released: 19-Dec-2006
Date Closed: 19-Dec-2006

To view this Sun Alert document please go to the following URL:
http://sunsolve.sun.com/search/document.do?assetke y=1-26-102732-1

OMG you are smoking Java crack there boy. Client side Java has more vulnerabilities than... Javascript. I love Java, but keep it on the server where it belongs. MySpace is getting ready to consider migrating from .NET to Java, it's solid on the server. But on the client... nope.

Take this from the LAST sunsolve weekly report:

Newly Released Sun Alert Notifications

Sun Alert ID: 102729 (RESOLVED)
Synopsis: Security Vulnerabilities in the Java Runtime
                              Environment may Allow Untrusted Applets to Elevate
                              Privileges and Execute Arbitrary Code
Product: Java 2 Platform, Standard Edition
Category: Security
Date Released: 19-Dec-2006
Date Closed: 19-Dec-2006

To view this Sun Alert document please go to the following URL:
http://sunsolve.sun.com/search/document.do?assetke y=1-26-102729-1

Sun Alert ID: 102731 (RESOLVED)
Synopsis: Security Vulnerabilities Related to Serialization
                              in the Java Runtime Environment may Allow Untrusted
                              Applets to Elevate Privileges
Product: Java 2 Platform, Standard Edition
Category: Security
Date Released: 19-Dec-2006
Date Closed: 19-Dec-2006

To view this Sun Alert document please go to the following URL:
http://sunsolve.sun.com/search/document.do?assetke y=1-26-102731-1

Sun Alert ID: 102732 (RESOLVED)
Synopsis: Security Vulnerabilities in the Java Runtime
                              Environment may Allow an Untrusted Applet to Access
                              Data in Other Applets
Product: Java 2 Platform, Standard Edition
Category: Security
Date Released: 19-Dec-2006
Date Closed: 19-Dec-2006

To view this Sun Alert document please go to the following URL:
http://sunsolve.sun.com/search/document.do?assetke y=1-26-102732-1

this is the year that sun's "whatever the hell we are calling thin clients this year" breaks the MS stranglehold on the corporate desktop

Basically, OOP languages like C++ and Java use this methodology, but it's obscured through friendly syntax. What we can expose from the above is:
1. When class members change, structs change. This causes binary incompatibility between different versions of libraries. The exception is Objective-C, which looks up members based on their name in a hash table generated at run-time

2. The addresses of branches (specifically CALL to call a function) are indirect (yes, in Obj-C too); this means that you can replace the class with another class that has the same structure but different functions being pointed at. It also means that the CPU can't do branch prediction, which hurts pipelining and intelligent CPU caching, causing pretty big slow-downs.

Secondly, if you have a Java method whose performance is critical, you can declare it as final. This gives the JIT compiler the option to either inline the method or create a direct non-virtual link to the method.

http://shop.sun.com/is-bin/INTERSHOP.enfinity/WFS/ Sun_NorthAmerica-Sun_Store_US-Site/en_US/-/USD/Vie wStandardCatalog-Browse?CategoryName=Sun_StorageTe k_6540&CategoryDomainName=Sun_NorthAmerica-Sun_Sto re_US-SunCatalog

32 terabytes (ish) for $160k. That's list price, and that's Sun hardware. You could definitely get it much much cheaper (I just couldn't be bothered to seek it out).

Storage is pretty cheap these days.

I thought for sure that the D language was at least 25 years old and built way back when. Then about 2 years ago Solaris 10 came out with dtrace and all of the "scripts" have .d extensions. Yeah I know it is unix and the extensions are meaningless.

Dtrace user guide

But the disambiguation page on wikipedia seems to bear out that calling a programming language D will only be interesting. Anyone know if it is trademarked yet?

Some websites offer users the choice of more than one format. It doesn't have to be limited to one choice or the other. Here is one example of a web page that allows users to choose which format they want to use when viewing a video clip. In this case it happens to be a choice between Flash and Ogg.

Several video clips in Flash and Ogg format

It is not unreasonable to expect an official government website to make an extra effort to make public records available to all voters. Offering the content in two alternative formats would be a reasonable solution. At least one of the formats should be an open standard such a Ogg, the other could be a proprietary closed standard that would require using Windows Media Player. Flash might possibly be acceptable too, because most Linux computers can play Flash (although the AMD-64 version of Macromedia Flash for Linux is not yet available).

A couple of years ago TurboLinux 10F was being sold for $69.00 in the U.S. and came with the Cyberlink's PowerDVD software for Linux with support for most Windows Media codecs up to version 9. According to this review, it does it does so in way that was legal and completely licensed.

TurboLinux 10F Review

Codeweavers sells a slightly enhanced version of WINE called Crossover Linux which, among other things, allows Linux users to run various Windows plugins. It allows Linux users to run Windows Media Player 6.4, although I am not sure just how well it does that. If a Linux user is using Windows Media Player 6.4 under Crossover Linux, they should be allowed to view the EU's streaming service.

To be equally fair to all voters, they should also offer their streaming audios or videos in an alternative format such as Ogg Theora. Here is an example of a web page that offers the choice of viewing some videos in either Flash or Ogg. I am using Linux and when I clicked on one of the Ogg links the video began to play perfectly. On most Linux computers the Flash version would also probably work, although the 64-bit version of Macromedia Flash for Linux has not yet been released. I am using the AMD-64 version of Kubunutu Linux without Flash, so I watched the Ogg video instead. If the EU included an Ogg version of their videos, they would then definitely be able to support Linux in a legal way. In a democracy, all voters should be given equal access to public government information. To achieve equal access for all voters, they should make the slight extra effort to include a version of their streaming audios or videos in some other format such as something like Ogg.

OpenOffice is just one of many contenders. Try Star Office, OpenOffice's commerically supported brother. Or KOffice, the office suite from the KDE project. Or for a less mature option, there is Gnome Office which is just a wrapper for AbiWord and Gnumeric for the most part.

not to mention it's 3 times faster on some workloads, and at least 60% faster on others (mysql), all the while being more reliable and less full of security holes ("hello race condition, nice to meet you")

Linux is the Windows of UNIX... about time we all just jump off that crap kernel

I tried this once. It worked well under Win2K server, and the old licensing regime. For the new one, I paid almost as much in client access licenses as I would have for Windows XP licenses. As an academic shop, our costs were low enough for that to make sense, but for those of you in the real world, I'd be more careful.

We ran into a couple of issues. The easily solved one was multiple copies of Office and Matlab are resource hogs. Get a large application server for those, and look into some sort of clustering. The harder issue was that some software actually sniffs to see if you're running on a remote session, and refuses to start if you are.

To an extent, this worked great, as the desktops were more secure and easier to manage, while the Windows boxes transitioned from whatever Dell had cheap that month to multi-proc systems with hardware RAID, high-speed SCSI drives, and redundant power supplies. The only real lossage occurred for those users who needed high-performance 3d rendering on Windows, and most of them were easily transitioned to Linux equivalents. You may wish to consider getting WinTerms if you go this route, as those have fewer parts to maintain and fewer ways for employees to tinker with them.

Quite seriously, you may wish to give Sun Sunrays http://www.sun.com/software/index.jsp?cat=Desktop& tab=3&subcat=Sun%20Ray%20Clients/url and Secure Global desktop http://www.sun.com/software/products/sgd/index.jsp /url a look. It will run from a Linux server, so it's on the right path from your perspective. One past job used an earlier version of this technology for student kiosks in the library, and it cut our maintenance headaches versus real PCs. Just a thought.

I tried this once. It worked well under Win2K server, and the old licensing regime. For the new one, I paid almost as much in client access licenses as I would have for Windows XP licenses. As an academic shop, our costs were low enough for that to make sense, but for those of you in the real world, I'd be more careful.

We ran into a couple of issues. The easily solved one was multiple copies of Office and Matlab are resource hogs. Get a large application server for those, and look into some sort of clustering. The harder issue was that some software actually sniffs to see if you're running on a remote session, and refuses to start if you are.

To an extent, this worked great, as the desktops were more secure and easier to manage, while the Windows boxes transitioned from whatever Dell had cheap that month to multi-proc systems with hardware RAID, high-speed SCSI drives, and redundant power supplies. The only real lossage occurred for those users who needed high-performance 3d rendering on Windows, and most of them were easily transitioned to Linux equivalents. You may wish to consider getting WinTerms if you go this route, as those have fewer parts to maintain and fewer ways for employees to tinker with them.

Quite seriously, you may wish to give Sun Sunrays http://www.sun.com/software/index.jsp?cat=Desktop& tab=3&subcat=Sun%20Ray%20Clients/url and Secure Global desktop http://www.sun.com/software/products/sgd/index.jsp /url a look. It will run from a Linux server, so it's on the right path from your perspective. One past job used an earlier version of this technology for student kiosks in the library, and it cut our maintenance headaches versus real PCs. Just a thought.

Also keep in mind that creating a C++ object in the stack is 25 times faster than creating a Java object (even with the server VM) according to my benchmarks...

If there's one thing we should all know by now, it's that benchmarking the performance of Java bytecode running on a JVM with a Just in Time compiler is very, very tricky. Modern JVMs may apply optimizations to your code while it's running, based on the way the code is actually being used.

Take a look at the slides from Dr. Clifford Click's presentation Java Technology Performance Myths Exposed (PDF) and at Performance Considerations for Run-Time Technologies in the .NET Framework by Emmanuel Schanzer to get an idea of how your code is manipulated at runtime to increase performance. Also, Performance of Java versus C++ by J.P.Lewis and Ulrich Neumann discusses some of the difficulties of benchmarking Java.

The only problem being it ignores TTL and the whole thing has to be reloaded to refresh that cache.

You're right that choosing infinite TTL for DNS lookups by default was moronic. However first Google hit for +DNS +TTL +Java gives you how to set TTL for DNS lookups in JVM startup script:
http://java.sun.com/j2se/1.5.0/docs/guide/net/prop erties.html.

See, there was no need to restart JVM. Lazy developers, lazy admins.

JVM under GPL is still a great thing.

Cheers

You're thinking of windows. phoneME is the latest name for J2ME or Java 2 Micro Edition, the version of Java that is put on phones and PDAs.

Or Looking Glass, perhaps? Looks pretty impressive to me, although I haven't tested it yet.