Slashdot Mirror

Definately for the middle of the road server/workstation, X86 CPU prices are driven down by the amazing volume of them that are sold. Go to the lower or high end and it's different

Maybe, depends on how much you spend and what you are doing, Apple's G5s are just bastards of much MUCH faster computers

The PPC design is simpler and generally going to be cheaper to implement.

Cool. I worked on it too. I designed some of the most kick-a$$ cables especially for it:

040-681-3925, Serial Port Converter Cable, 9-Pin to 25-Pin. $22.50
Converts signals both ways, handles voltages up to 100 volts! Necessary for syncing your Palm.

7040-681-3125, Serial to Serial Port Cable for Rack/Rack. $72.00
Flexible, with two connectors, one at each end. No cheap $50 serial cables for you -- this computer demands the best! Doubles as a tie-down strap when transporting your p690.

Yeah. Except that with .Net Code Access Security, you could say that only binaries signed with a certain key could run. And that the binary could only access a certain file on the drive. Or just about any permission you can think of. "aka applet" cannot do this. This is critical to application deployment on the internet.

You know what I find funny about that? Is that Java 1.1 could also do all that! It's called Signed Applets, and the Java Security Manager controls access to specific resources just as you stated.

Except that "aka webstart" does not have side by side execution, versioning and rollback.

You need to Read Up on Webstart. It does versioning. You could if you wished run different versions side by side (though probaby only a developer would ever make use of that feature). If you think about what the words "Web" and "Start" mean together, you will wonder why the word "rollback" has any meaning. Since when do you "rollback" a read-cache? You re-read the data.

Comparing Avalon and XAML with "aka SwiXML" is funny.

Ha-Ha funny, or "that's strange" funny?
How about XUL, or (dramatic pause...) FLASH!!!!! ha Ha Ha Ha Ho! Now that's funny. Flash does everything you want to do with XAML later on, across a million platforms, TODAY!

I'm sure the virus writers will love XAML though. Should be a tremendous boon for them when they start playing around with buffer exploits from 100% translucent video playing in the background of the cool XAML form app using some obscure and poorly-written codec.

The point is, maybe you could do a lot with java on the client. But doesn't that fact that not many people are using Java on the client suggest something else?

Actually a lot of people are, just in intranets. Shouldn't the fact that Flash has taken over all rich interactive browser UI like you wna to do with XAML scare you just a little? I don't know if you've been browsing recently but there are a lot of sites using Flash to pretty good effect.

I guess the first step of Longhorn is to not ship IE with Flash installed anymore...

The IBM version is pretty decent though.

The development tools under windows blow everything else out of the water.

Yeah because we all know that Fortran is completely dead.

Does this mean they might finally break that 7-qubit barrier that quantum computers up until this point had seemed to have been limited to?

There are many ways in which to perform quantum computing. The "7-qubit barrier" probably refers the factorization of 15 using 7 qubits. This was done using NMR-qubits (nuclear magnetic resonance). NMR-computing uses the nuclear spin of atoms in some molecule to represent the qubits. This is relatively easy to achieve, as the atoms are bound close together by the molecule, but it doesn't scale well, because there is a limit to the complexity of the molecule. NMR-qubits are therefore ideal for a proof-of-consept and the research with them is invaluable in the development of quantum computers, but they probably won't have any practical applications. (Of course, it's always possible that some breakthrough comes along...)

The next big thing will probably be trapped ion qubits, where ions trapped with a magnetic and/or electrical field are used to represent the qubits. This allows a much larger number of qubits, but it, too, has its practical limits. There are also other methods (eg. holding particles on a liquid-helium surface), which are not currently practical, but show great potential for the future.

In any development there often seems to be some barriers which cannot be broken, until someone does. This will probably be the case of quantum computers, too.

Well, some folks at IBM have a theory at least.

Not that I know enough about it to make any conclusion myself. The whole thing just sounds suspiciously like Douglas Adam's infinite probability engine to me. :P

One? linuxthreads is a subset of POSIX threads, and is part of glibc now. There's three different implementations of threads on Linux, the original linuxthreads(1:1), ngpt(M:N), and nptl(1:1). They might be implemented in different libraries, but they're just one API -- POSIX. Of course, maybe you mean Apache's APR, wxWidget, and the other platform abstraction libraries?

I've written a webmail client that does threading, as well as thread arcing. Check it out.

I'd write pooling code to handle the objects instead

Actually, object pooling may hurt!

I've found the linux kernel's built-in RAID capabilities more than adequate for most of my fault tolerance needs. The best part is I can move the drives to pretty much any system - a new motherboard, whatever - without having to worry about kernel support or finding that IDE driver. If a drive fails I can boot its mirror up in any system and be in great shape. I also use the utility mdadm to email me if one of the drives fails. For some linux firewall systems I've built, I use old crappy 6GB drives, but mirror them so there's no risk if one of them goes out. Looking at my basement firewall now and...

root@fw01:~# cat /proc/mdstat Personalities : [linear] [raid0] [raid1] [raid5] read_ahead 1024 sectors md0 : active raid1 hdb2[0] hda2[1] 38796864 blocks [2/2] [UU] unused devices: <none>

everything is cool!

Am I missing something, or did you just suggest writing a JVM in Java?

Check out this recent IBM Developerworks article which looks at how modern JVMs handle allocation and garbage collection.

Some very surprising tidbits there. For instance:

"Performance advice often has a short shelf life; while it was once true that allocation was expensive, it is now no longer the case. In fact, it is downright cheap, and with a few very compute-intensive exceptions, performance considerations are generally no longer a good reason to avoid allocation. Sun estimates allocation costs at approximately ten machine instructions. That's pretty much free -- certainly no reason to complicate the structure of your program or incur additional maintenance risks for the sake of eliminating a few object creations."

Read the article for an excellent nuts-and-bolts analysis of many current performance considerations. From the posts in this thread, it's pretty clear a lot of people haven't looked into what's actually done in a server JVM these days.

Link:

Jikes RVM

Written in Java. It was codenamed Jalapeno.

Another Physics Fanboy speaks out! Hi there, Physics Fanboy!

I read your "reference" (or at least the Google cache of it), and it doesn't even contain the word "computer", so I fail to see how you've supported the claim that QC can help with teleportation. See, your (attempted) sarcastic point was actually literally true; I do know that stuff. Evidentally better than you do, since I can describe why we aren't teleporting stuff around right now. Can you? After all, we teleported a photon years ago; why haven't we done anything significantly larger? (Maybe because it's impossible? Give the idea a fair shot.)

Anyone want to take a crack at providing a reference that actually, well, refers to WarriorPoet42's claim?

Of course there is no reason for you to read this, since you know so much already. I just wanted to point out that everything WarriorPoet42 said was true. I'm not so sure about your statements.

For those of you who don't know: The biggest problem with quantum computing is that you can never extract all the information you compute. So you can process y=f(x) for 2^31 values of x simultaneously, but when you go to read y from the computer, you just get one solution, and what's worse, you don't even know which value of X it corresponds to!

Using Shor's factoring algorythm, however, you can extract one of the factors of a large number without knowing all the other factors. That would be useful for public key encryption. I wouldn't worry about your PGP key just yet though. 7 q-bit computers are incredibly difficult to make. The process used to make the 7-bit QC does not scale to larger numbers easily. 2048 bit computers are way beyond our technical skills.

On a side-note, I wonder if each computer simulates a q-bit (with one responsible for management). It would be the most obvious way to run the simulation, but may or may not be the fastest. There would need to be a lot of cross-communication since all the q-bits are entangled in any interesting quantum computation.

If you could rebuild everything from the ground up (and had tons of money to throw at it), you'd most likely want to build a system based on a very expensive vendor solution.

Assuming that you can't do that, your best bet would be to go with some sort of parallel filesystem, the likes of Lustre, GFS, Ibrix, GPFS or CxFS. The architectures of these vary, but the basic principle they share is performance scalability based on increasing the number of data paths to the disk. So if you have, say, 100 nodes on a high-speed network, you take 10 of them and attach them to your SAN. The parallel filesystem spans the entire SAN and so requests from the nodes can reach any bit on the SAN from any of the ten paths. If you need more performance, you add more paths: controllers, HBAs and storage nodes. I know GPFS scales linearly in performance based on the number of paths to the data, and I believe the others scale well also.

I haven't hit 50 TB on disk (I have on tape, but your post suggests that tape wouldn't give you the performance you need), but I have set up several 4-8 TB GPFS filesystems that could easily grow to 50 TB if I had the spindles.

Good luck finding a solution; symlink-based solutions on a convnentional *NIX filesystem are a nightmare; I sympathize.

If you could rebuild everything from the ground up (and had tons of money to throw at it), you'd most likely want to build a system based on a very expensive vendor solution.

Assuming that you can't do that, your best bet would be to go with some sort of parallel filesystem, the likes of Lustre, GFS, Ibrix, GPFS or CxFS. The architectures of these vary, but the basic principle they share is performance scalability based on increasing the number of data paths to the disk. So if you have, say, 100 nodes on a high-speed network, you take 10 of them and attach them to your SAN. The parallel filesystem spans the entire SAN and so requests from the nodes can reach any bit on the SAN from any of the ten paths. If you need more performance, you add more paths: controllers, HBAs and storage nodes. I know GPFS scales linearly in performance based on the number of paths to the data, and I believe the others scale well also.

I haven't hit 50 TB on disk (I have on tape, but your post suggests that tape wouldn't give you the performance you need), but I have set up several 4-8 TB GPFS filesystems that could easily grow to 50 TB if I had the spindles.

Good luck finding a solution; symlink-based solutions on a convnentional *NIX filesystem are a nightmare; I sympathize.

1) I haven't seen any Linux wristwatches. While I'm sure there may be some out there, I seriously doubt they're anything more than cheap toys.

Here is one and here's another. Both were built around 5 years ago, back when people were still actually surprised by such a feat.

While they aren't in production, they aren't exactly cheap toys either. The IBM Linux wristwatch aimed to have the same capabilities as a Palmpilot. The other is a videophone developed by a wearable-computer inventor.

But as for other small embedded devices, I have seen Free/NetBSD on consumer routers and bridges.

Yes, FreeBSD exists on routers and bridges and rightly so; it's got stellar networking performance and stability. But there are many more embedded Linux systems simply because it is a far more flexible code base with support for a LOT more architectures than FreeBSD. And you tend to see Linux in a wider variety of devices, such as the myriad of set top box devices and the like.

There is no reason why FreeBSD couldn't be put on a big-iron mainframe.

Yes there is: because it would be a lot more work porting Linux. There's no one (or at least not many) currently working on putting FreeBSD on big iron whereas every few months you hear about some company or university adding enterprise and mainframe features to Linux.

But as far as I can tell, only IBM ever bothered to put Linux on a mainframe anyway.

They're possibly the only ones to put it on a mainframe and sell it commercially, but this is IBM we're talking about. They're not exactly tiny in the mainframe market. And they aren't the only ones working on getting Linux to scale to the mainframe level by any means.

So it's a pointless argument.

No, it is a perfectly valid argument.

I like and appreciate FreeBSD as much as anyone else, but I'm not a very good system administrator if I don't evaluate and acknowledge a product's weaknesses as much as its strengths.

Do you mean this white paper? Your URL is broken.

Wrong. Solid state memory has MORE failures per time period. Check out the IBM chipkill memory whitepaper. IBM has decided that multiword ECC simply isn't enough. To get acceptable reliability you need multiword ECC AND RAID across modules.

. . . but try launching Slashdot using this.

Though for news and entertainment in general, I usually take my pick from one of these (World service and Radio 4 I consider to be almost essential for my existence).

From a fellow INTP, been there, done that...

I've had the best success when I approached a class a a 'new game' rather than an opportunity to learn. If you think that college is a citadel of learning, you are in for a big reality check. I have a pal who holds several degrees (BS, MS, going for the MBA) and he put it best, "Professors are not there to teach you. They are there to give you a grade. And I mean 'give you' the grade they feel you deserve."

They are people just like you and I, no more and no less. Imagine yourself as a professor for a moment.

Ok, so it's a game. The object of the game is to get a professor to give you a good grade. Learn who they are, how they operate, what they expect, and do some work.

My favorite example is a Socio-cultural Anthropology class I took (requirement filler). The prof. was about as PC as they get. We had two texts and an autobiography to read. I managed a B without opening any of them (just to see if I could do it I think.) I just answered the insanely easy multiple-guess exams in the most PC way I could. I hit it right on the head, that's what she wanted us to "learn" - the PC crap, not any real anthropology methodology (hint: we had movies to watch ever other week, that was a dead givaway we were not going to 'learn' anything).

Anyway, enough of my ramblings... remember, college is a big new game to you. One that you haven't learned the rules to yet, that you haven't mastered yet, that's rather difficult and many people can't master. It's the grade game ultimately, with a side bet on if you manage to make a few close friends there and learn a bit from it as you go. Go win it if you think you can ;)

p.s. I'm a hotshot developer with a good job I enjoy, and I almost finished my BS (got enticed by the boom, or more precisely, the money that was available back in the 90's). I may still get that degree yet !

OK, is there (should there be?) anything like this for PowerPC assembly? I'm aware of these web pages:

• Stuart Ball: Beginners Guide to PowerPC assembly language
• Hollis Blanchard: Introduction to assembly on the PowerPC

You're right, the OSS community doesn't have the resources for this kind of marketing blitz. But have hope, we're trying to help... ibm.com

http://www-306.ibm.com/chips/techlib/techlib.nsf/t echdocs/7874C7DA8607C0B287256BF3006FBE54/$file/PPC _QRG_2-22-04.pdf

Not happ with something like this?

...and multiple fans: Apple does it because they want to keep the machine as quiet as possible while still as cool as possible (as opposed to being forced to do it, lest the processor become hotter than the surface of the Sun).

(They don't do it because the PowerPC 970 family is "so hot", either; the PowerPC 970, and the 970FX even moreso, run much cooler, and require less power, than even the newest generation PowerPC 74xx (G4) family processors: )

Also, new PowerPC 970FX information from IBM is now available.

Actually that number includes all of their partners in that area.

In 2003, IBM received 3,415 U.S. patents from the USPTO. This is the eleventh consecutive year that IBM has received more U.S. patents than any other company in the world.
linky.

So not quite 6K, but more than I thought (almost 10 a day!) Their 10 year average is closer to 7 a day, and if you go back 26 years I'm sure it's even lower. Of course the rediculous number makes my point even more clear that fighting IBM in a patent battle is sheer stupidity.

Dude, no one competes with IBM on patents, they have averaged more than a patent a day for as long as any currently enforceable patent has been in existance.

I think your numbers are just a *tad* off. Yes, they do a bit more than a patent per day. In fact, according to IBM, they get over 6,000 patents per year. That's over 16 every day of the year, or about 24 per business day.

To anyone familiar with the topic being discussed -- as opposed to a government bureaucrat that couldn't recognize a piece of code if it fell on their head -- it seems obvious that Ken Brown is being disingenuous.

AdTI did not publish Samizdat with the expectation that rabidly pro-Linux developers would embrace it. Its purpose is to provide U.S. leadership with a researched presentation on attribution and intellectual property problems with the hybrid source code model, particularly Linux.

Clearly not, because the pro-Linux developers are aware that the claims, as outlined in this document, are preposterous. "rabidly" is simply a thinly disguised ad hominem attack trying to lend some semblance of credence to claims which deserve none.

The United States is the home of the United States Patent and Trademark Office, an internationally respected agency which contributes to the worldwide effort to protect and govern intellectual property. In addition, the U.S. government is one of the largest patent holders in the world, owning the rights to 20-30,000 patents.

The USPTO cannot easily be said to be "internationally respected", when in fact huge organizations and a tidal wave of dissatisfcation have arisen, because the USPTO grants 95% of all patents filed, has no manpower to assess the validity of patents and has spawned an industry of patent litigation and manipulation. The government's portfolio is relatively unimpressive; IBM, a single US company, has 23,000 active patents and was granted 3415 in 2003 alone. (And, it should be noted, IBM is one of the largest corporate supported of Linux; funny how they are simultaneously the largest patent 'consumer', but also one of Linux's biggest supporters; that in itself seems to imply a disconnection in the documents conclusions)

The disturbing reality is that the hybrid source model depends heavily upon sponging talent from U.S. corporations and/or U.S. proprietary software. Much of this questionable borrowing is a) not in the best interest U.S. corporations b) not in the best interest of IT workers in America c) at a serious expense to the investment community, an entity betting on the success of intellectual property in the marketplace.

This is very cute, and the derogatory language immediately gives away the intent. "Sponging"? Linux inspires grateful contributions of improvements from developers excited to be a part of a phenomenon. That's not "sponging". And I'd like to see Ken Brown justify his argument of theft of proprietary software, which sounds slanderous to me. Note that OSRM now offers intellectual property insurance because their EXTENSIVE review of Linux concluded that it was legally unencumbered. I doubt Ken Brown was so thorough.

Moreover, Brown's paragraphs conclusions about 'best interests' is fallacious. Because with respect to: (a) U.S. Corporations like IBM support Linux because it gives them access to powerful technology that they can build services and other revenue models on top of without forcing them to develop it themselves and Linux inspires more confidence than their proprietary products (such as AIX) ever did; (b) IT workers can benefit enormously from cheaper and more accessible software; expensive proprietary products simply allow soaking customers for IP, which does not benefit IT workers who do less work and see corporations charge more for it; moreover, OSS gives independant and small IT shops the ability to build powerful services on top of openly available tools; I make my living this way (and a good one), and (c) the investment community is irrelevent -- their "bet" on the IP marketplace is not one that government policy makers need to hedge for them; if someone betting on an "IP marketplace", if in fact such a bet has been made, th

All the big players are buying into this. For the Google-challenged:

EMC's take from todays Cnet
CA
HP
IBM(tho' they prefer the acronym HSM or Hierarchical Storage Management)
Veritas
Random googled pundit on the entire field.

Ok, Apple has a near monopoly on the PPC market. How many other PPC based computer

manufacturers are there? What is their market share, compared to Apple?

Not really yet ready though.

Generating Word documents using XSLT

Thinking XML

Opening Open Formats with XSLT

Office 2003 XML Reference Schemas Overview

Why You Should Choose MS Office Over OO.org

Why not complicate a complicated world a little more. Each standard unit of complication renders X standard monetary units in someones pockets.

CC.

We looked into something a lot like what you suggest (and actually have it up and running inside our intranet with 2k or so users). The problem with doing this on the internet is that p2p techniques are MUCH more susceptible to spamming than centralized techniques in general (because, for one, p2p reputation systems are very difficult to get right). Another problem is that most existing p2p search methods work great for finding popular content but not very well for finding that very specific peice of information that maybe only you are looking for at the current moment. Kademlia/Chord are DHT's and do not solve the text search problem on their own. While some p2p networks have adapted DHT's for keyword searching, the results still leave a lot to be desired (IMO).

You mention a good point. Email messages have different purposes. Some are chit-chat, some worthless jokes, some spam, some messages external to the organization, some internal, some from your boss, some from your employees, etc. But it's all the same in the inbox, governed by the same disk quota and archival policies. Perhaps the file/folder dynamic for email is a little lacking (I think of IBM's Remail, though I'm not sure if this really hits the problem). We just recently got "junk" as a new category with our filters. People use email differently of course, but we have few methods of treating messages differently outside of the user's direct intervention.

There are different ways to improve this. Perhaps one is to look at making email messages more like documents (like XML) and give the user more options in creating, saving, and searching their messages. Another is to improve indexing and categorization apart from the user entirely (something like Google). Another is efficient and longterm storage (like seperating attachments from mail).

Right now we have some more crude solutions like saving or deleting mail based on spam-targeted filters, timestamps, etc. But you are right, I think the solution has to start at the inbox and the user. We may need some fundamental improvements to email. Of course, we risk overcomplicating our busy and over-technical lives.

• A cache model in which there is one cache for instructions and another cache for data is called a "Harvard-style" cache. This is the model assumed by the PowerPC Architecture, e.g., in the descriptions of the Cache Management instructions in Section 3.2, "Cache Management Instructions" on page 16. Alternative cache models may be implemented (e.g., a "combined cache" model, in which a single cache is used for both instructions and data, or a model in which there are several levels of caches), but they support the programming model implied by a Harvard-style cache. The processor is not required to maintain copies of storage locations in the instruction cache consistent with modifications to those storage locations (e.g., modifications caused by Store instructions).

That's the point here. The PowerPC does not guarantee that self-modifying code will work. There's a long discussion of how to do self-modifying code in section 1.8.1 of the architecture manual. This is mainly for debugger support.

Given this instruction/data separation, it's not surprising that the PowerPC architecture defines a no-execute bit (it's bit 61 of each page table entry). Unfortunately, not all PowerPC models implement it.

GNU Classpath
GCJ/GIJ
Kaffe VM
Jikes Java compiler
SableVM
Java-GTK
Documents about how to compile and use QTJava and KDEJava

What else do we need? How does all this not play well with Free software? We've got the tools, why not use them?

A friend of mine's mother worked for IBM back then. One day, after ferreting around in the attic, this friend presented to me a small notepad with the word "THINK" on the cover. An old-school ThinkPad!

Can anyone confirm that this is where the name of their laptop lineup comes from?

There's a nifty utility for IBM Thinkpads, which also works with other computers that lets you set a delay for CAPS LOCK, as well as F1 Tab, and ESC.

If you want to keep these keys, but don't want them acticvating by accident, check it out.

I use it for CAPS LOCK and F1.

Does anyone know the story behind Apple's THINK (not think different) poster?

Heck, if forking is a concern, then they should see the writing on the wall and open source Java as soon humanly possible.

There'd be no better way to nip potential feature drift in the bud in the gcc, IBM, etc re-implimentations of Java in the bud than to open up Java. Yes, the compilers are going to be different, but noone in their right mind is going to reimplement a reasonably licensed set of libraries, and that's where almost all the real meat is in Java.

It's also the patent numer (by IBM!) for a 'graded channel field effect transistor'

Conspiracy indeed!

Now LinuxWorld has an IBM add for their xServer! This is an intel-inside box! (There's even an official intelInside logo on the ad!) Clearly LinuxWorld is biased by accepting advertising from the original cause of the whole Wintel phenomenon - IBM! After all, you can't have PC-compatible without PC!

Depending on their support package it might mean RBC's IT department won't catch all the heat for this.

Depending on their support package it might mean RBC's IT department won't catch all the heat for this.

Create native, cross-platform GUI applications

Create native, cross-platform GUI applications, revisited

Not compile once, run everywhere, maybe write once, compile everywhere, but that is Java, GPLd with a GUI.

Create native, cross-platform GUI applications

Create native, cross-platform GUI applications, revisited

Not compile once, run everywhere, maybe write once, compile everywhere, but that is Java, GPLd with a GUI.