Slashdot Mirror

The only place where I have seen frames used to good effect is Javadoc and similar API documentation. Using frames, you get a convenient index of the classes on the left, and can scroll that list independently of the main API docs.

For example, check out: the JDK API

However, if you are referring to cross-site frames, I completely agree with you (and am grateful for Firefox's 'show only this frame' function).

I seem to remember Java having some sort of hook into Direct3D for graphics. It may have Microsoft's version in Win2K and not the Sun Microsystems version, so take it with a grain of salt. But I know whatever version of Java was used definately interfaced with Direct3D, and did so incorrectly.

I had to find out how to set the flags to explicitly disable it, as the damn thing would actually Blue-Screen my Win2K box.

Looking now, seems it was actually Sun's version; they even state how to disable it, and warn about how it could crash on Win2K (heh). And it was introduced in version 1.4.1_02, per this page. Just search for the "d3d" System Property, a little above the halfway point of the page.

More info at

http://www.sun.com/service/sunmd/
http://en.wikipedia.org/wiki/Sun_Modular_Datacenter

Sun Modular Datacenter (Sun MD, known in the prototype phase as Project Blackbox) is a portable data center built into a standard 20-foot intermodal container (shipping container) manufactured and marketed by Sun Microsystems. An external chiller and power are required for the operation of a Sun MD. A data center of up to 280 servers can be rapidly deployed by shipping the container in a regular way to locations that might not be suitable for a building or another structure, and connecting it to the required infrastructure.[1] Sun Microsystems states that the system can made operational for 1/100th of the cost of building a traditional data center

Beowulf is not going to help here. Beowulf is for clustering computational resources. You want to cluster storage resources, something like RAID-Z.

I know exactly what the difference between the two are. I also know that native machine code for one system can be interpreted on another. That is what non-JIT emulators do. Emulators like VICE and Java. I also know that using the technique of "Just In Time" compilation, chunks of native code for one system can be recompiled to the host platforms native code for dramatic speed ups. This is done in some Jave emulators, as well as many other emulators such as UAE.

Right from suns site a description of the java virtual machine looks exactly like the description of a physical processor. Why? Because the JVM is an emulator of a real machine that may or may not have ever actually been built.

I understand the definition just fine. You are just trying to change the definition to something that boils down to "cross platform is whatever I decide is cross platform". You define one emulator (JVM) as cross platfrom, while other emulators (VICE) you define as not. You define some interpreted instruction sets (Python) as cross platform, while other interpreted instruction sets (6502) are not cross platform even when code written in it can be run across multiple platforms.

If I write a VM like the Python VM, but I dropped the context of white space, and use commands that look like one or two bytes for each command, would that be cross platform or not?

I installed Ubuntu on my MBP too, eventually got sick of little driver annoyances and I'm now running on a Dell netbook that I knew would work perfectly with Ubuntu. Ubuntu 10.04 seems to have sorted out some problems for the MBP (which I just use as a server now) but I already use my netbook 100% of the time so I don't really care anymore.

Right, some of this is just you complaining and not actually finding out how to use your computer. Look at http://blogs.sun.com/arungupta/entry/osx_5_alt_tabbing_in to find out how to switch between apps and between windows in a single app..

As for the dock, yes it integrates the idea of Widnows quick launch, and I think it does it very well (so much so that whenever I install Ubuntu I install a dock instead of the taskbar at the bottom of the screen). I grew to like the concept pretty quickly. I didn't find it hard to notice the arrow/dot below the apps at all, and I had it configured so that there were only a few apps in there so that it wasn't such a mess as when you first install OSX. Extra ones would pop in there too when I was running apps that I didn't usually use, and I thought it was just a great space saving concept overall.

I never really used Expose. I'm an alt-tab kind of guy, though often clicking on the dock is handy in Ubuntu because it also switches to the appropriate desktop and application with one click rather than requiring a few different keystrokes. I'm guessing it would work the same with Spaces on the latest version(s?) of OSX.

As for the right click stuff I can't comment, I don't remember it being a problem. I actually prefer using "touch two fingers to the pad and then left click" for a right click, but I don't have that option on my netbook.

I agree that lack of cut and paste is annoying, yes.

I built a ZFS file server last year using OpenSolaris. I can't speak for FreeBSD, but it sounds like you had a much more fruitful experience than I did.

OpenSolaris is very picky when it comes to hardware and device drivers. It reminds me of trying to install Linux twelve years ago. There is a Hardware Compatibility List but it is behind the times and poorly maintained. The problem is that Solaris was developed for a relatively small pool of certified enterprise grade hardware. Trying to get consumer grade hardware working can sometimes involve buying one product after another until you find a model that works, no matter how well you think you have researched it.

I had a nightmare trying to get my Intel PRO/1000 Gigabit networking card working properly. It worked fine for a year, then suddenly suffered massive performance loss after I upgrading the OS. The whole saga is documented here. It got to the point where I was monitoring network traffic using WireShark and buying new switches. In the end I gave up, installed Ubuntu Server and set up an XFS software RAID. All my problems suddenly vanished - I had a lightning fast file server again!

So I literally wasted several solid weeks fighting OpenSolaris. I also wasted a good amount of money on upgrades that in the end were unnecessary. Even the best minds on their support forums were unable to help (and believe me, I gave them enough traces, WireShark dumps and analysis!) This is a shame, because ZFS itself was wonderful to use - it was just the rest of the system that let it down.

- Andrew

You do not see this sort of API stability from almost any other vendor. API that worked in Windows 95 still works, more or less.

Solaris has always done great in this regard. Sun in fact has maintained binary compatibility up to Solaris 10, the current production release. It's even a guarantee.

You can use completely free and open source tools to create high-quality Flash content. For example, you can use the combo of Sun's JDK, the Free & Open source FlashDevelop IDE (Microsoft .NET 2.0 required) and Adobe's Open Source Flex SDK.

that's because there aren't any high profile web sites written in Java - they're too slow, buggy and useless. Even assuming what you say is 100% true, how come vendors patch security flaws in Java itself and they never get exploited?

We had our corporate system breached due to a flaw in a very big, expensive 'Enterprise' java web system so I know it from experience.

How about a quick google for information. This one has a table of java web framework security features. This one (pdf) describes "Our static analysis found 29 security vulnerabilities in nine large, popular open-source applications, with two of the vulnerabilities residing in widely-used Java libraries. In fact, all but one application inour benchmark suite had at least one vulnerability."

as well as "A recent penetration testing study performed by the Imperva Application Defense Center included more than 250 Web applications from e-commerce, online banking, enterprise collaboration, and supply chain management sites [54]. Their vulnerability assessment concluded that at least 92% of Web applications are vulnerable to some form of hacker attacks"

Ho hum. I guess you think Java is completely secure. What a fool. Hope no-one hires you to write any of them, well, not until you go to java.net and look up all the 'how to make your java app secure' tutorials. Sounds like "fucking morons" like you need them more than most.

Wasn't Sun supposed to revolutionize the world with a similar 3D desktop back in 2004?

http://java.sun.com/developer/technicalArticles/J2SE/Desktop/lookingglass/

The problems with Flash are many and technical. It's so badly designed from a security perspective, that it's almost like a Microsoft product. The giveaway that it is not is that it runs on a handful of linux architectures. Games could just as well be written as Java Applets, which would increase the security and portability of the games. For movies, Flash is just plain wrong and other wrappers should be used, Ogg Theora being the obvious choice after MPEG or QuickTime.

It's a sad day when most know so little about architectures that PPC and Sparc are re-flagged as mere 'alternatives' rather than being recognized for the areas they excel in. The Wintel ideal is a ratio of worse that 2:1 of hardware to services. Each box burning tens if not hundreds of watts. Sparc is an open architecture and handles many threads per core, so for most things you should be able to replace a rack of Wintel boxes with a single Sparc. We'll see how long Oracle allows you to access that Sun paper.

Fujitsu also provides info sparc architecture because it also sells server hardware.

Linky linky
http://wikis.sun.com/display/SunSolve/How+Entitlement+Works

With Z already in place and with [not so] recent inline deduplication feature, I think ZFS should do it.

Check Goldberg's article, 'What Every Computer Scientist Should Know About Floating Point' for a more in depth discussion.

"The floating-point types are float and double, which are conceptually associated with the 32-bit single-precision and 64-bit double-precision format IEEE 754 values and operations as specified in IEEE Standard for Binary Floating-Point Arithmetic , ANSI/IEEE Std. 754-1985 (IEEE, New York)."

http://java.sun.com/docs/books/jvms/second_edition/html/Overview.doc.html

Xserves are hardware, Sun Messaging Server is a piece of software written in C.

I'm not sure I understand what you mean. For one thing, the iPhone sandboxes all apps from one another for security and stability. How do you do that within a JVM? And are you talking about making a blacklist on the JVM itself to try to prevent certain code from running, rather than using the whitelist model all other apps in the store use?

Just like PCs do, run each Java app in its own JVM instance.

As for other security issues, I would assume you'd be dealing with signed JARs.

Not Linux per se, but the same idea is implemented nicely on ZFS through its L2ARC: http://blogs.sun.com/brendan/entry/test [sun.com]

And, of course, you can run Linux on it via iSCSI or NFS. I even have a machine tha runs Xen, a CentOS 5 Dom0, a Nexenta DomU which gets physical disks for ZFS, and shares them to Fedora DomU's, all on one piece of physical hardware.

Really the only weak link is CentOS's iSCSI tools, which go out to lunch on occasion.

The submitter wants something like ZFS's L2ARC, which uses the flash as an intermediate cache between the RAM cache and the disk. This works very well for a lot of workloads. Since Linux users appear to be allowed to say 'switch to Linux' as an answer to questions about Windows, it only seems fair that 'switch to Solaris of FreeBSD' would be a valid solution to this problem.

Not Linux per se, but the same idea is implemented nicely on ZFS through its L2ARC: http://blogs.sun.com/brendan/entry/test

Swapcache on DragonFly BSD 2.6.x was implemented for this very reason IIRC.

http://leaf.dragonflybsd.org/cgi/web-man?command=swapcache&section=ANY

Not Linux per se, but the same idea is implemented nicely on ZFS through its L2ARC: http://blogs.sun.com/brendan/entry/test

In addition, some OSes, such as Linux, support copy-on-write semantics for memory pages in child processes created with fork (note, Solaris is an example of an OS that *doesn't* do this).

What year do you live in? Solaris _9_ had COW and multiple page size support, over half a decade ago. Linux large page size support is a joke, Solaris x86 even does it better on Linux's home turf.

http://www.sun.com/blueprints/0304/817-5917.pdf

Most modern OSes have a native fibre channel stack, except notably, Linux which doesn't have userland utilities for managing SCSI devices or even fibre channel drivers for that matter.

See what I did?

This article claims that on Solaris, "fork() has been improved over the years to use the COW (copy-on-write) semantics". It's sort of an in-passing comment, though, and I can't find a definitive statement in docs anywhere (the Solaris fork() manpage doesn't give any details).

This article claims that on Solaris, "fork() has been improved over the years to use the COW (copy-on-write) semantics". It's sort of an in-passing comment, though, and I can't find a definitive statement in docs anywhere (the Solaris fork() manpage doesn't give any details).

The Java SE page has downloads that don't have the obnoxious toolbar/trial crap in them
http://java.sun.com/javase/downloads/index.jsp

no-one makes their hardware open-source except a few niche vendors

Like my employer.

When you say "niche vendor", I think of the guys who helped design my bicycle.

>Cue fat lady!!

You mean Queue?

Nope - IBM would have outsourced nearly all of Sun to India in the first 90 days.

But what would they have done with these guys' jobs, I wonder...?

I can't believe no-one has mentioned this yet. Oracle's Exadata2 solution uses Sun x4175 and x4275 servers, and runs on NO, not Solaris, but Oracle Enterprise Linux. (Which I believe is just a RedHat variant.)

Its my impression that Oracle bought Sun for the hardware, in order to deliver a one-stop-shop solution for Oracle clusters. The one-throat-to-choke model, if you will.

http://www.oracle.com/technology/products/bi/db/exadata/pdf/exadata-storage-technical-overview.pdf

slides 16, 17, 22, and 57. And that helpful link was provided by Scott Davenport's Sun blog at:
http://blogs.sun.com/sdaven/entry/oracle_exadata_2

IE8 opens the unsigned application right away without prompting. ;-((

http://java.sun.com/javase/technologies/desktop/javawebstart/demos.html

Those apps ARE signed. If you look in the java control panel you'll see that there is a certificate for Sun installed. Remove that certificate and those apps behave just like all the other unsigned apps, and you'll get prompted first.

Of course this is unrelated to the current flaw.

This is worse that I thought, further research reveals that : ;-)

In their default configurations:

1) Firefox prompt you with a dialog similar to "open file abc.exe". ;-))

2) IE8 opens the unsigned application right away without prompting. ;-((

http://java.sun.com/javase/technologies/desktop/javawebstart/demos.html

Also Web Start use some sandboxing, but I have trusted it since I have never looked it up ;-))

At least with the official Sun JRE, it never affected 64-bit Linux, because they don't support Java Web Start on the 64-bit distribution. (The 64-bit Linux OpenJDK does support JWS, though.)

This was fixed in a really old update: http://java.sun.com/javase/6/webnotes/6u17.html

Actually it affected Linux browsers too.

However, it was fixed a few updates ago: http://java.sun.com/javase/6/webnotes/6u17.html

We may soon need similar lessons here in the UK when we want to access those filtered sites suspected of potentially hosting copyrighted material. Damn, that sounds sad.

Hate to break it to you but most web sites you could ever even think of accessing will be hosting copyrighted material. That's right not just potentially hosting copyrighted material but actually hanging up copyrighted material for anyone to download.

To avoid getting copyrighted material, you'd have to find a country that did not sign the Berne Convention treaty, but even then the material might be under copyright. Alternately, even the countries in the Berne Convention treaty might have material online that has been made Public Domain either because the copyright expire or the rights holder (not the creator) put it into the public domain. Even then you'll have to download (and read) pages of copyrighted information to get at the PD stuff.

Alternately you can just download as much copyrighted material as you want. Try starting from these sites:

• SourceForge
• CreativeCommons
• Linux Kernel Archives
• arXiv
• Ubuntu
• Fedora
• NetBSD
• Oracle
• Sun
• Haiku
• Internet Archive
• and so on

And remember, there's more where that came from.

"Very few systems (especially in house ones) require true cross platform development so that's generally irrelevant anyway."

That's too blanket a statement to be valid in the general case, it's certainly true for some companies. As I stated however, our company is Microsoft based, however even here we want to expand some of our apps onto mobile devices and we have a combination of them such that Java is the only real sensible option. Of all the companies I've worked in I've yet to work in one that only ever has a single platform throughout the entire company, they've all had the odd Linux server between their Microsoft servers, a combination of mobile devices and so forth. It's certainly not an uncommon situation to want apps to be portable. There's also the issues of larger companies which have different operating subsidiaries who have to share some apps and data, but who also are given autonomy on IT decisions from subsidiary to subsidiary- Java absolutely excels here, it acts as a common language that just works between subsidiaries pretty much whatever their platform choices.

If you're not developing in house applications and are developing to sell Java makes sense too, because there's no point writing say, a piece of helpdesk software in C# .NET, or C/C++ with multiple binaries to sell when you can just write once with Java and inherently have a product that works across Windows, Linux and Mac OS X greatly expanding your potential clientbase.

"Or maybe its a developer who doesn't have a knee jerk reaction that the tool that leads to the quickest prototype is the best."

Whose talking about prototypes? I'm referring to real working apps.

"I've developed back-end trading apps in the past that required the fastest possible throughput of data (we're talking down to milliseconds being shaved off here) to beat the competition and for that Java simply was not an option."

Really? Apparentlyy the NYSE doesn't agree with you:

http://www.nyse.com/tradingsolutions/transacttools/1204674243385.html

"Not every "app" is some floppy piece of GUI code that sits there doing bugger all 99% of its life - some apps are back end systems that are maxed out all the working day and for that you can't beat C and C++."

Simply put, you're wrong. Java performs just as well as C/C++ in many cases, better in some, slightly worse than others. This is largely because the JIT compiler is better suited to optimising per platform, rather than per architecture like classic compilers. Plenty of case studies here for Java use in HPC for example:

http://www.sun.com/customers/index.xml?soln=31a8487e-0f60-11da-99bc-080020a9ed93&page=1&sort=date&asc=false

The fact that you talk about Java being faster simply for prototyping, the fact you are not aware of the fact that Java performs just as well in many cases as C/C++, and the fact that you do not think Java is used for high load back end processing demonstrates one thing- you do not know enough about Java to be able to correctly evaluate whether it is the right tool for the job or not in the face of C/C++ and are a good example of the type of developer I was referring to as not being a great developer for this reason. It may well be that C/C++ was in fact the right tool for your particular solution after all (i.e. if you had some custom hardware to take advantage of), but as you clearly don't know much about Java, you cannot possibly say for sure whether that was the case or not, despite the fact you are attempting assert otherwise.

The likes of eBay runs on Java and much of Google's back end work is done with Java also. There's a good reason it's the most prominent language in business still today and has been for a while. It's because it does offer advantages, it is versatile, and yes, it ca

there are many allocator implementations that behave way better than the single-thread-optimized default malloc. Hooray for another one. Here is one old article about that: http://developers.sun.com/solaris/articles/multiproc/multiproc.html IIRC even glibc uses a slab allocator these days, and is much better performance wise on multi processor machines.

While there are certainly parts of OSX that are open, they are largely the bits that aren't interesting. Darwin is basically just BSD with an extra dose of weird. Webkit is interesting; but is a derivative of KHTML, which is LGPL, so that is a legal obligation on their part. As for the "don't have to jailbreak" bit, just try using Dtrace without doing some surgery...

On the UI/Bundled applications layer, it is basically a closed shop, even in insultingly small ways. Y'know what "Burn Support: Yes (Apple Shipping Drive) " in device properties implies? That 3rd party drives, lacking Apple's special blessing, are restricted from working with integrated burn support, even though ATAPI was sorted out years ago. Or why the (relatively few) 3rd party wi-fi adapters all ship with their own crappy little applets, just like back in the Windows 98 days? Because, unlike the Windows equivalent, where any remotely recent wireless device with the right driver support can be managed through the OS-provided wi-fi configuration interface, Apple restricts their "airport" configuration interface to their cards only. How nice of them.

Apple has every right to do what they do in regards to their products; but the notion that they are "open" in any useful way, aside from the ability to run unsigned binaries in OSX, is basically fantasy.

I think he was talking about namespace aliasing, so you could e.g. use both org.foo.Table and org.bar.Table in the same source file without having to write out the fully-qualified name of either of them.

While it's true that Java's library is highly hierarchical, you don't type out whole namespace paths all the time. You just import namespaces.

Java 5 also features static imports:
import static java.lang.Math.PI;
import static java.lang.Math.*;

double r = cos(PI * theta);

Free Java IDEs like Eclipse and Netbeans manage imports automatically. You don't have to type them out.

This would introduce a bug in your application, since Calendar.getInstance() always returns a new instance, containing the current time at the moment it is created. Storing it in a static variable and reusing it would return the same time forever.

That's not a bug - that's a feature!

But in all seriousness - I wrote an app for myself to keep track of reminders, which works similar to anacron. It's about 5 years old now, and still working. I used setTimeInMillis periodically rather than constantly creating new Calendar objects every time a method is called. (which could be thousands of times within a second very occasionally, or no times per second)

Moreover, Calendar is not thread-safe and is mutable, so storing it in a shared static variable is a really bad idea.

In the case of my reminder app, it's single-threaded, so who cares?

Read the bloody docs before you use a class. ;) That's advice everyone should follow.

Once you've read Date, Calendar, GregorianCalendar, and TimeZone, you're probably ready to use those classes.

http://java.sun.com/javase/6/docs/api/java/util/Calendar.html

I was serious. Tail recursive call is one of the major feature lacking in the JVM. See http://blogs.sun.com/jrose/entry/tail_calls_in_the_vm http://openjdk.java.net/projects/mlvm/ And there is no way to implemnent tail-recursive call in its full generality (including tail-call optimisation to statically unknown code, e.g. thru virtual methods...) without a proper support from the JVM. This has nothing to do with addressing modes (which you don't need in a VM offering garbage collected values). And the Ocaml virtual machine bytecode does indeed show that other kind of bytecodes can be useful. JVM is not the perfect VM; it could and should be improved.

What is Solaris? Solaris is SunOS plus a GUI.

No it is not, my confused, penguin-worshipping young friend. Solaris == SunOS.

That leaves ZFS, which Sun took special measures to keep out of Linux.

Not true, once again. Sun was investigating a Linux port.

Back in 1996 I was replacing SparcStation workstations, low end ones like SS1+, SS2, SS5, and IPX, with intel-motherboard PCs with onboard Mach64 video

My condolences.

OMG, folks... Some idiot troll made braindead posting and all the slashdot started buzz. FUCKING LEARN TO READ IN ENGLISH.

http://www.sun.com/software/solaris/popup.jsp?info=17

The registration process to receive an Entitlement Document is part of the Solaris download process, with the Entitlement Document being returned to you via e-mail. For this reason, YOU MUST PROVIDE A WORKING E-MAIL ADDRESS AS PART OF YOUR SUN DOWNLOAD CENTER ACCOUNT. If you fail to do so, you will not receive an Entitlement Document and will only have the right to evaluate Solaris for 90 days

Oracle only asks for valid email address. Once valid email passed and Entitlement accepted, 90 day restriction does not apply.

with NexentaStor CE, which is based on OpenSolaris b134. It's free.. and has an excellent Storage WebUI. /plug

For a detailed explanation of OpenSolaris dedup see this blog entry.

~Anil

Tiered Storage - "on the fly" support for that is something I would love to see:

Great idea! I'm using that today on FreeBSD with great success. Furthermore, the idea works well enough in practice with real machines that even a cheap USB flash drive gives a nice boost.

ZFS offers this already, they call it the L2ARC, you can read about it here: http://blogs.sun.com/brendan/entry/test

Oh, wow! I am even more impressed by ZFS now. Thanks for sharing, This is probably worth a separate Slashdot story.

I just want to congratulate Oracle on doing everything it can to kill off Solaris passively so they don't have to admit what they're doing. I need a Solaris support contract in order to keep a few systems running specialized software in a compiance-audited environment up to date. This is software that is run in many environments where the inability to keep them patched is a showstopper. However, I can't seem to purchase a support contract. The only page that even lists the ability to purchase it is broken (see dpfloyd's comment), and I have not receved a call back from Oracle/Sun sales in nearly a week (and that was after getting bounced through 6 different people to a support person who at least knew to forward my info to a Sun-related salesperson, or so they said). Additionally, if you click the "How to Purchase a Contract" it provides no actual info on how to do that, and the link it has to "Learn More" takes you into an infinite loop of "click here, now click here, now click here - oh, wait, I'm back where I started" when you try to find out about Sun Solaris support.

I hope I'm wrong about what's happening, but I can't say that any of this gives me the warm fuzzies. I'd say that if I had control over the platform I'd migrate those systems off of Solaris to another OS, but I'm guessing that's exactly what Oracle wants...

Can SOMEONE at Oracle/Sun please tell me how to purchase a support contract to download OS patches? If not, can someone from Oracle/Sun officially tell me to bugger off so I can tell my boss that we're never going to be able to update those servers again and we can start planning on how we're going to get around that issues?

Thanks.