Rob, as a former resident (now living in sandgroper territory), I'd like to suggest you get the time to visit a few places around Sydney. I'm sure your aussie hosts will show you the way, but here's my list for foreigners visiting Oz:
Blue Mountains/Katoomba area. Fantastic scenery and some really nice walks with real history.
King George IV pub, Picton. Make the finest microbrew stuff and have been doing it for decades, before it became trendy. Try the Dogboulter. I promise that you'll need to use one of the bedrooms there after two or three of these.
Any boat or ferry trip on the Harbour.
Royal National Park - Some place like Bundeena on the beach or the quieter Maainbar. If you can ride a motorbike, or opentop car, do the entire trip and stop off at Stanwell tops. Do this on the weekend because you'll get a heap of hang gliders, R/C types and occasionally a _big_ glider.
There are many other places (like any of the beaches!) but these will give you a real taste of some of the variety offered by the NSW region and get you out of the city. Feel free to email me for more places to see:)
Funny thing is that I didn't hear about an Oz LE until this post, and it's not like I don't hang out in places where I wouldn't be told.
Taco, if you want, there's a spare bed here for a couple of days holiday if you can make it over to Sandgroper territory. Just email the address above !
It's the same as hard links (you know, ln without the -s option. Simple, been around as long as soft links. Try creating hard links to files and then modifying or deleting them.
Already done. Most of the URN work is already hammered out, and a few of the older RFC's need to be updated a bit.
From what I'm reading here, the form of those URLS this guy is generating is actually illegal syntax. That is, with the '?' character, that is intended as a query and any proper web server would attempt to run a CGI type script with it.
I spot a trance fan in CowboyNeal. Completely useless information follows...
One of the earliest and most successful underground tracks was called Two Full Moons and a Trout released on Platipus records under the artist title of Art of Trance (I think that's right, or it might be Terra Firma). Wicked track. Anyway, just happy to see another raver in the/. crew.
I know quite a few members of the Java team at Sun and can assure you that it is not their doing. One guy I personally know very well has a 10+ year history of OSS work before joining them. The developers understand what is going on and so do their immediate managers.
The problem is the higher ups in Marketing/Legal. They don't care. Despite the bitching and screaming by Sun's own development teams, it is not their fault. I'm lucky enough to be privy to some of the internals there and it is certainly not pretty the shit fights about the SCSL and this sort of thing. You can guarantee that Sun's developers would have been jumping down the Legal/PR's throats before even the general community did.
I'm not sure if I agree directly, but I think the whole Apollo program was a great hack. Just think of the memorable things 30 years later - Space Food Sticks, Tang and Velcro....
I had the fun of working with an ex-Apollo veteran for 3 years. He was working in the Simulator side. None of these lovely Onyx boxen for generating graphics - all mechanical star fields and control maintenance. Computing was barely even used for the control and monitoring.
He worked on the simulator side of the Apollo 13 recovery. The story goes that he was clocking off shift on that day. The guy before him left the building through security, but he got turned around and told to go back to work. 48 hours later and he takes the first bit of sleep. Now I've done quite a few 24+ hr coding runs, but this still blows me away every time I think about it. Not only did these guys have to know the entire computing system, they also had to know most of the maths/physics they were simulating _and_ also had to be a half-decent mechanic too. There's not many of todays hackers that could claim that level of capabilities.
The most interesting things you never hear about. I spent a lot of time travelling with him to do various things. The really great hacks of the entire Apollo program will never make general knowledge. I'm pleased that I've had a chance to hear about many of them first hand from someone who really was there.
There was a latish episode of TNG where Data thinks he has found his "mother" (the wife of his inventor?). Near the start of it, the two of them are playing some violin piece IIRC. Then they play it again, half way through and this tweaks Data's memory into some analysis. Although she sounds like a human with all the emotion in the playing, he realises that she's played it exactly the same way twice - something that only an andriod is capable of.
Reading through this article, it seems to be exactly the same thing. There doesn't appear to be any randomness associated with it based on how the "player" is feeling. Pass the piece of music through the machine and you get exactly the same reproduction time after time. (like a previous poster talked about the pissed conductor etc)
I think the music generated would be OK for TV themes and that sort of stuff, but to get real performances you need all the extras. The sounds of the coins rattling in the Double Bass player's pocket, the clicking of the keys on the woodwind instruments (my bassoon sounds like a train when played fast - clicketty, clicketty clack, click, clack...), turning of pages etc that make a performance. There used to be a time 10-20 years ago where the record companies would try to eliminate all of this extraneous sounds from a recording. If a muso dropped a mouthpiece on the floor, they'd stop the recording and start again. These days, the trends are to realistic sounding recordings - just have a listen to the latest re-releases of the Star Wars sound tracks for example.
So the basic message is, real musicians will never be replaced by this because there are so many other factors that go into making a recording than just being able to do a drum-roll correctly. Besides, look at the community orchestra/concert band scene. Do these people play music just so they can be on CD, or for the pure fun and relaxation of it?
We have a huge 80Gb mp3 file storage at work. It's a box with a big RAID array running NFS and Samba. Unofficial, of course.:)
Have to agree with that:). Our entire test network is devoted to MP3 storage. However we pale in insignificance to your collection. At this stage we've got around 20GB before we ran out of storage space. Oh, what's that, the phase II test server has an empty 12GB. Hmmm...... We're all ripping our albums there. Luckily most of the guys in the project like dance/techno so we're ripping about 8 albums a day at the moment for the collection (depending on how fast I wear out the visa card each fortnight!)
While I agree I must issue a warning to all users^H^H^H^H^Hreaders that debugging while listening to Dark Side is a seriously dangerous trip. Those bells and buzzers going off in one ear from almost dead silence, while intensely concentrating can do wicked things to your sense of reality.
I did the Floyd thing while doing the study thang for the HSC then the first few years of uni. I think I know every single word of every album of theirs!
These days, the collection is mainly dance stuff. Primarily UK based (Global Undergrounds, React, Platypus, Perfecto etc). I seem to have this thing for UK music:)
The reason why is no video game out there will support the huge amounts of triangles unless every card can handle them. The game would have to be practically rewritten from scratch for the higher triangle count.
Obviously you haven't heard of implicit surfaces. You know, those things like B-Splines, NURBS et al. Describe a surface by a series of control points and then tesselate according to your performance requirements. Start off with a low figure, or do benchmarking when first installing the game then use this info to up the number of triangles until you hit the frame rate/quality level ratio you want. Very simple, very old technique. I s'pose you haven't heard of the Teapot either.
Alternate to Motorola units that do run Linux
on
911 Calls Linux
·
· Score: 1
As an interesting counter point to this: I'm fairly sure that the radio system they use are an APCO 25 unit (digital radio system. Sort of a cross between mobile phones and CB radio). Coincidently, another part of my company (ADI Ltd, Aussie defence contractor) happens to make APCO 25 base units running on... Linux.
ADI is the only other manufacturer of APCO units in the world. AFAIK, the base stations/call control system have always been running linux. I first saw them around 3 years ago so I can positively date our use of it at least that far back. When I walked in an saw this, I was _extremely_ impressed at the time as I was pretty much assumed I was the only one in the company that knew about it them (To be fair, we'd just purchased that part of the company a couple of months previously as part of a bankruptcy of a company called Stanilite - makers of all those little green "Exit" emergency lighting signs!).
Having written one myself, it is not possible to do a complete implementation of a Telnet client in Java. The main reason - standard Java sockets don't support OOB messages. A number of features of the Telnet protocol rely on OOBM to provide control constructs that Java does not allow you to hook into. You could provide your own socket implementation, but that means using native code and then you don't get the portability....
Does the jre support double-clicking on a.jar file on windows?
As other's have commented, yes you can - and it works cross platform.
In your manifest file add a line:
Main-Class: somepackage.someclass.class
Now double clicking on the file or running
java -jar myjarfile.jar
will automatically launch your application. Also, if you need to rely on other JAR files you can set the class path with:
Class-Path: jar1.jar jar2.jar...
as well. The only thing that is not possible with JAR files currently is to have individual icons available for desktop usage. Enough people are clamouring for this that Sun have at least made statements that they will look into some way of providing it.
Java and Jini are immature technologies that should not be used in a life and death environment under any circumstances.
Hmm... funny that. I've written a lot of defence software. Everything from mine warfare systems to counter-terrorist stuff to desktop simulators. One lot was written on SCO unix. Most unstable, unsupported POS you could imagine. My current work is Java based and is being used today in real war situations (yes, deployed on the big two way range).
Of the two, the Java is the most stable and best piece of software according to the users. Now, to give you an idea of exactly how much better Java is in a defence environment that native code - the original app was 200K lines of code written in 8 weeks (and an estimated 800 litres of Coke/Jolt to accompany it). That was a piece of code for a tender for a paid prototype software. They were so impressed that 6 weeks after signing the contract the system went operational with almost minimal changes. Now this is not just your average army unit either - the local Special Forces (Oz SAS - US equivalent of the Rangers/SEALs). The success of this is such that we roll out new versions of the software every 6 weeks which they happily install and run with. Try doing that with native code software (Yes, we used to do really big monolithic Ada/C software here as well so I do know how long it takes).
Sun, by contrast, has never given up control of the language. Is sun going to decide to raise prices or terminate a license on the DOD in the middle of an operation?
That has got to be the biggest POS that I have read in a long time. Maybe you should actually look at Java at some time. Sun controls the language specification. They aren't the only supplier of runtime environments or development environments. There is no dependency on Sun. How can you pull a license to a specification when one was never granted in the first place? What do you think the CLASSPATH, Blackdown and other similar projects are all about. You might as well argue about what would happen if Borland pulled the license on Delphi in the middle of an operation, or MS with VB (there is more military code written in VB than ADA).
Next time, grab a clue stick before talking about something you obviously have NFI about.
To start with, it is quite obvious that you haven't been hanging around the graphics arena all that long. Goemetry accelaration has been something that is a feature of all high-end (read non-consumer level) cards and architectures for at least the last 7 years. If you want to be really picky about this, the reason the scene graph was invented around 30 years ago on the old vector displays was for dealing with hardware geometry accelaration - and this was 2D, not 3D.
According to a number of different, collated sources and things I saw at Siggraph, the reason the PSII is so fast is because the CPU is devoted almost entirely to the needs of graphics. This includes things like vector and matrix mul-add instructions (just like a DSP chip) and putting the entire thing on a single chip. Internally, the bus width is either 2560 or 4960 bits (can't remember exactly now). The real specs are around 66 million textured and lit polys per second. That is, texturing and lighting make almost no speed difference due to the huge bus width.
Considering that the top of the line IR2 does around 10-12M polys a second the Nvidia chip is still going to fall far behind. Why? Well the IR2 does almost the entire GTXSR pipeline on separate hardware. Consumer level cards don't do that. Although the TNT2 et al claim to do 5-6M polys/sec that is only if nothing else needs to be done. Add in some lighting calcs, alpha blended textures and partially obscured polygons that a real world app uses and watch those figures plummet (due usually to bus bandwidth or CPU limitations). The current cards are raster only. Geometry accelaration will help, but it won't go all the way to beating an SGI pipeline. Besides, if I need more graphics grunt for my SGI, I just add another card or two and I instantly double/triple/etc my polygon performance.
At Siggraph, there was a really interesting panel about large scale data visualisation. This was run by the people from Los Alamos, NCSA etc. They were talking about the PS performance wrt other architectures. They had some really interesting things to say about it - ie, expect to see (or not see, depending on your security clearance) Blue Mountain and the other ASCI computers using PSIIs and other similar boxes doing some of the real time interactive rendering of the data sets. That I find really amazing.
I attended the Fahrenheit BOF and there were some pretty heavy questions being asked of MS there wrt to Linux support. They copped quite a caning about it and Unix support in general. The general attitude was if it isn't really trully supported on Unix, nobody was really interested.
Now to show you how interested MS is in supporting unix, they have hired a 3rd party company to do the porting. Said (unnamed) company does not have any graphics knowledge - by MS's own admissions at the BOF. Kinda says something about it doesn't it.
I managed to get in a question at the end about how heavily MS was going to persue the licensing wrt to the OSS community developing an implementation (yup, I was the blue-haired one standing near the door). Their response - "umm... we hadn't thought of that. Might be worthwhile.". There were also a bunch of other statements about how well specified Fahrenheit is compared to the standard Win32 APIs. I just hope they are actually telling the truth this time and it's not another Chrome type effort. Given the reaction on the guy's face when I asked the question I tend to believe that they honestly hadn't thought about the spec issues and OSS implementations.
Other noteworthy things were that they have extensions for particular application areas. No-one in the MS Fahrenheit team has any decent Viz-Sim experience. They don't even have an extension underway to deal with this effort. Not particularly good if you want it to be used in large scale environments.
Lots of excuses on the MS part about not wanting to speak for SGI and their commitments to the IRIX side of it (the BOF was held a couple of hours after SGI announcement of the big culling going on).
Now as one who has done a pretty significant amount of programming for Plan9 a few years back (92-95) and X[lib|t|motif] (95-97) I can honestly say that X is a *lot* better than what Plan9 gives you.
In those days it was even a case of write your own toolkit. No such things as drop down menubars, just a mouse popup. To create a menu, first create a standard popup, fix it in place and go from there, You even had to do a lot of your own background buffering for repainting of windows after the menu was hidden. You didn't even have generalised scrolling areas - only text. Want a graphics scrolled area - write your own.
It was fun (I was at Uni at the time) but there is no way I would want to use Plan9 for a commercial application. Just to write a file manager it too 3 of us 6 months of heavy hacking.
The one thing that I loved about the Plan9 stuff was 9term. Love that cut & paste of the entire window contents, not just the last command line. Also, sam was a pretty decent editor too.
Ender's Game started VRML
on
Ender's Shadow
·
· Score: 2
Ender's Game was another of the defining series of novels that shaped a lot of what you see around on the web today - particularly the virtual reality field. The roots of VRML came not from Neuromancer, but from Ender's Game. Talk to Pesce, Parisi, BehlenDorf or any of the others that worked for Autodesk on their Cyberspace project (one of the precursors to VRML) about it. The reason they created the original cyberspace banana was not because they wanted to be Case, but because they wanted to replicate the Game that Ender played (remember his tablet where he couldn't get through the part with that figure and he could only do it by punching the eyes out of it and the little room with the Egg in it).
Personally, I'm already queued to get Shadow. Loved the original's so much that I just had to get this. Got a local mob that flys out everything the day it is released and only cost A$1 more for the privilege. Hmmm... can't wait!
Just to add one more rambling point - I'd love to see another view of the universe that looks from Peter's and Valerie's perspectives. There's a big hole there about how Peter became Hegemon and Ender the Speaker for the Dead that I'm dying to find out what happened.
Well, from what I've seen around the traps, and the people that I know, one company is almost definitely Sun and the SCSL.
A lot of the engineers inside Sun are pissed at the SCSL in a big way - particularly the commercial clause. At JavaOne, a number of different BOFs that I attended, when Sun lawyers were present, they copped a hammering about the SCSL (admittedly I was one of the people doing said hammering). From memory the SCSL and JINI BOFs were the two most problematical for their legal types.
Here's what I know is going on inside at the moment. The Java Media folks are pushing through a version of SCSL that does not include the commercial clause. That is, you can pretty much do anything you like with it - no compatibility checks, no license fees etc. Take the source and run. I imagine this would bring it extremely close to satisfying the open source guidelines. Naturally the Sun legal people are having a hard time swallowing this so it is taking some time to get through (It's been going at least for the last 3 months that I know of). Note that this is not changing the core SCSL, just a variant of it minus the crappy bits (like the Jini License is a variant of the SCSL).
The first of the Java Media APIs to undergo this is JSDT (Java Shared Data Toolkit). Once this goes through, pretty much all of the rest of the Java Media stuff will be following in quick succession. (Java3D, Media Framework, Sound and Advanced Imaging).
BTW - This is not leaking any private information. The Sun people at the JSDT BOF were the ones talking about this (Rich Burridge and Mike Bundschuh (sp?)) out in the open. Also, directly talking with the various Java Media marketing types and engineers running the stands there were talking about it too.
In the example you use of Perl, it's not O'Reilly's fault that the officially reference manual is proprietary, it's the authors'. They chose to publish it through O'Reilly, knowing that would mean it was proprietary. They were the ones not being "good citizens of the Open Source community" (in your opinion, not mine).
As an author and a coder I'd have to partially disagree with this.
By far the majority of computer books that are written come from an idea from the publisher. That is, the publisher will have an idea for a brand of book (eg the Dummies series) and then develop titles for it. This is usually done in conjunction with someone knowledgable in that area (usually the writer of the first book). The various titles are developed, a rough outline of the topic coverage and a list of potential authors. At this point, the development editor(s) (sometimes called acqusition editors) will start recruiting writers either by direct contact or through an agency. Now, if you approach the publisher first with a completely outlined, highly detailed (and preferably at least half written) book, then you have some good cause to ask for a different copyright scheme. A lot of the time you will get it too - it just depends on who the publisher is (O'Reilly is much more favorable to this than say Addison-Wesley).
Once an interested author has been found, the title is then developed to a much finer detail with the development editor etc until it is passed of to the series editor, tech editors etc for the actual writing.
Now the writer has a fair say in what happens with the book, but in general it doesn't go to the extents of saying "how would you like to license this book?". The publisher owns the copyrights on the basis that they solicited the work from you, not you proposing the work to them. Depending on your status as an author (prior recognition in your community etc) you can negotiate a reasonable degree of freedom. For example, I always get a release for all of the code (non-exclusive rights for redistribution yada yada yada...). This allows me to publish and further develop the code from the book as Open Source (usually LGPL'd because it is library type code). One of the benefits of this is that it helps to further enhance the author's reputation and also sells a lot more copies of the books.
(Side note: I've recently been contacted by a publisher that wants to do a Linux book under the Open Content licensing scheme. That was one of the first things they stated, so it does happen).
To prevent a few further comments - This code is all Java stuff or VRML. The documentation is always freely available online. What the book gives the reader is many of the inside tips of why things were done in certain ways. At least in my approach to writing I use the book to teach people how to approach designing stuff using the language APIs given to develop various types of libraries. I then proceed to develop a particular library, which is what ends up being the OSS product I put on my website. The book is not the design documentation or official API guide, but a way of approaching design philosophically. From most of the authors that I chat with, this is typically the case. The big exception to the rule is the "Bible" or "Nutshell" books that are just API references with a bit of extra stuff thrown in.
It's interesting that I seem some really striking similarities between what is happening now and what Stephenson wrote about in Snow Crash. There, in a virtual world (that happened to be 64K square) inhabitants could buy their virtual property. There we good areas, bad areas, and sparesly populated areas (like Rev Bobs place). People live in the virtual world while it maintains there look on the real world (That dude in the Van with the wicked toys).
Although the VR version of cyberspace has not yet hit critical mass to start producing 3D versions (despite my, others valiant efforts with things like VRML and AlphaWorlds), the 2D version is very much starting to take on these Matrix-like effects. Some parts of the matrix are very popular (.com,.net,.org) while others are not (.us) while others are barely populated (.ro,.pt). People buy and sell these virtual addresses for some very huge amounts of money. Other people are the architects providing home building services, and down the bottom are a bunch of hackers like me trying to get the plumbing working.
What is happening now is the move from the 2D virtual property to the 3D. Although 3D is populated by game engines (and you can class, Isometric, 2.5D and 3D in this) eventually the same thing will spread to the general populace. As Katz points out, the arrival of the Middle-class into once what was hacker territory is a significant thing. We legitimize both the worth of the property and the fact that it exists. Stock options - how much virtual property are they? They are no more than a bunch of bytes on someone's harddrive these days, so what is the difference between that and a couple of pieces of armour and a good character.
The early adoptors always end up with a bunch of the most valued parts. Look around at the domain names, slashdot logins (I got in real early to get mine for example and the newer crowd haven't got such a range anymore. Maybe the login here will be worth something someday?). An interesting thing about these early adoptors is that normally they take the best stuff, not because it is, but because it reflects their personality or other trait that they like to express themselves with. Then, as time progresses, they become some of the most valuable simply because others want it. The gaming examples are just following the same trend already established in the flatlands. How long before some of these game environments start "meeting" at the seems?
The flip side to this is that, because we are digital, we can create as much property as we like. Just because a few games are hot today does not mean that they will be the only property available in a year down the track. Effectively limitless space can have some interesting effects on what and how items are valued. At what point do people come in to make a hedged risk on a new piece of virtual real-estate.
The funny thing that I find about all of this is that it will create jobs for the middle men. OK, so eBay is probably one of only a couple of sites on the 'net doing this now. How long do you think it will be before we start to see specialist virtual game property trading companies/sites being formed (darn, I should go out and patent that business idea!). That is, what we have in the real world will be mirrored in the virtual which will be mirrored in the real world... Just like the Matrix.
We're writing as quickly as we can. There are actually a whole heap of documents there, but the guy that did the main page hasn't linked them in yet. You have to read the CVS update messages to know where to find things:) (I remember that at least there is a sendmail page there)
Rik has just moved the list over to majordomo (message came through all of 5 mikes ago). You can join the list by sending mail to
Note that we are taking a completely different approach to tunelinux. We're breaking it down into functional areas rather than specific applications. For example, to tune a, oh.. let's say samba server, you need to tune SAMBA, the kernel and a number of other things to get the best from your box. If you take the tunelinux approach, you'll only end up with a 50% solution because you'll just tune SAMBA, and not the kernel. The Mindcraft report is a classic for that. Yes, tune samba, but hey, the kernel still only uses 960MB of RAM. The "tuning samba" doc approach wouldn't cover that sort of information all in the one spot. Effectively you end up with the same problem as we have now - lots of documentation but so fragmented that you can't do anything useful with it.
Rik van Riel has already started a linux tuning site over at nl.linux.org. Already there are some 50+ folks on the alias with documents starting to fly around the CVS system. Nobody has mentioned this site to us, so what's the deal?
Slashdot Mirror
- Blue Mountains/Katoomba area. Fantastic scenery and some really nice walks with real history.
- King George IV pub, Picton. Make the finest microbrew stuff and have been doing it for decades, before it became trendy. Try the Dogboulter. I promise that you'll need to use one of the bedrooms there after two or three of these.
- Any boat or ferry trip on the Harbour.
- Royal National Park - Some place like Bundeena on the beach or the quieter Maainbar. If you can ride a motorbike, or opentop car, do the entire trip and stop off at Stanwell tops. Do this on the weekend because you'll get a heap of hang gliders, R/C types and occasionally a _big_ glider.
There are many other places (like any of the beaches!) but these will give you a real taste of some of the variety offered by the NSW region and get you out of the city. Feel free to email me for more places to seeFunny thing is that I didn't hear about an Oz LE until this post, and it's not like I don't hang out in places where I wouldn't be told.
Taco, if you want, there's a spare bed here for a couple of days holiday if you can make it over to Sandgroper territory. Just email the address above !
Maybe he should ask about Dropbears - a relative of the koala bear.....
It's the same as hard links (you know, ln without the -s option. Simple, been around as long as soft links. Try creating hard links to files and then modifying or deleting them.
From what I'm reading here, the form of those URLS this guy is generating is actually illegal syntax. That is, with the '?' character, that is intended as a query and any proper web server would attempt to run a CGI type script with it.
If you want to know more about URNs, and my implementation of them in Java (replaces most of java.net) go to http://www.vlc.com.au/~justin/java/urn/
One of the earliest and most successful underground tracks was called Two Full Moons and a Trout released on Platipus records under the artist title of Art of Trance (I think that's right, or it might be Terra Firma). Wicked track. Anyway, just happy to see another raver in the /. crew.
The problem is the higher ups in Marketing/Legal. They don't care. Despite the bitching and screaming by Sun's own development teams, it is not their fault. I'm lucky enough to be privy to some of the internals there and it is certainly not pretty the shit fights about the SCSL and this sort of thing. You can guarantee that Sun's developers would have been jumping down the Legal/PR's throats before even the general community did.
I had the fun of working with an ex-Apollo veteran for 3 years. He was working in the Simulator side. None of these lovely Onyx boxen for generating graphics - all mechanical star fields and control maintenance. Computing was barely even used for the control and monitoring.
He worked on the simulator side of the Apollo 13 recovery. The story goes that he was clocking off shift on that day. The guy before him left the building through security, but he got turned around and told to go back to work. 48 hours later and he takes the first bit of sleep. Now I've done quite a few 24+ hr coding runs, but this still blows me away every time I think about it. Not only did these guys have to know the entire computing system, they also had to know most of the maths/physics they were simulating _and_ also had to be a half-decent mechanic too. There's not many of todays hackers that could claim that level of capabilities.
The most interesting things you never hear about. I spent a lot of time travelling with him to do various things. The really great hacks of the entire Apollo program will never make general knowledge. I'm pleased that I've had a chance to hear about many of them first hand from someone who really was there.
Reading through this article, it seems to be exactly the same thing. There doesn't appear to be any randomness associated with it based on how the "player" is feeling. Pass the piece of music through the machine and you get exactly the same reproduction time after time. (like a previous poster talked about the pissed conductor etc)
I think the music generated would be OK for TV themes and that sort of stuff, but to get real performances you need all the extras. The sounds of the coins rattling in the Double Bass player's pocket, the clicking of the keys on the woodwind instruments (my bassoon sounds like a train when played fast - clicketty, clicketty clack, click, clack...), turning of pages etc that make a performance. There used to be a time 10-20 years ago where the record companies would try to eliminate all of this extraneous sounds from a recording. If a muso dropped a mouthpiece on the floor, they'd stop the recording and start again. These days, the trends are to realistic sounding recordings - just have a listen to the latest re-releases of the Star Wars sound tracks for example.
So the basic message is, real musicians will never be replaced by this because there are so many other factors that go into making a recording than just being able to do a drum-roll correctly. Besides, look at the community orchestra/concert band scene. Do these people play music just so they can be on CD, or for the pure fun and relaxation of it?
Have to agree with that :). Our entire test network is devoted to MP3 storage. However we pale in insignificance to your collection. At this stage we've got around 20GB before we ran out of storage space. Oh, what's that, the phase II test server has an empty 12GB. Hmmm...... We're all ripping our albums there. Luckily most of the guys in the project like dance/techno so we're ripping about 8 albums a day at the moment for the collection (depending on how fast I wear out the visa card each fortnight!)
I did the Floyd thing while doing the study thang for the HSC then the first few years of uni. I think I know every single word of every album of theirs!
These days, the collection is mainly dance stuff. Primarily UK based (Global Undergrounds, React, Platypus, Perfecto etc). I seem to have this thing for UK music :)
And by the way, which one's Pink?
Obviously you haven't heard of implicit surfaces. You know, those things like B-Splines, NURBS et al. Describe a surface by a series of control points and then tesselate according to your performance requirements. Start off with a low figure, or do benchmarking when first installing the game then use this info to up the number of triangles until you hit the frame rate/quality level ratio you want. Very simple, very old technique. I s'pose you haven't heard of the Teapot either.
ADI is the only other manufacturer of APCO units in the world. AFAIK, the base stations/call control system have always been running linux. I first saw them around 3 years ago so I can positively date our use of it at least that far back. When I walked in an saw this, I was _extremely_ impressed at the time as I was pretty much assumed I was the only one in the company that knew about it them (To be fair, we'd just purchased that part of the company a couple of months previously as part of a bankruptcy of a company called Stanilite - makers of all those little green "Exit" emergency lighting signs!).
Having written one myself, it is not possible to do a complete implementation of a Telnet client in Java. The main reason - standard Java sockets don't support OOB messages. A number of features of the Telnet protocol rely on OOBM to provide control constructs that Java does not allow you to hook into. You could provide your own socket implementation, but that means using native code and then you don't get the portability....
As other's have commented, yes you can - and it works cross platform.
In your manifest file add a line:
Main-Class: somepackage.someclass.class
Now double clicking on the file or running
java -jar myjarfile.jar
will automatically launch your application. Also, if you need to rely on other JAR files you can set the class path with:
Class-Path: jar1.jar jar2.jar ...
as well. The only thing that is not possible with JAR files currently is to have individual icons available for desktop usage. Enough people are clamouring for this that Sun have at least made statements that they will look into some way of providing it.
Hmm... funny that. I've written a lot of defence software. Everything from mine warfare systems to counter-terrorist stuff to desktop simulators. One lot was written on SCO unix. Most unstable, unsupported POS you could imagine. My current work is Java based and is being used today in real war situations (yes, deployed on the big two way range).
Of the two, the Java is the most stable and best piece of software according to the users. Now, to give you an idea of exactly how much better Java is in a defence environment that native code - the original app was 200K lines of code written in 8 weeks (and an estimated 800 litres of Coke/Jolt to accompany it). That was a piece of code for a tender for a paid prototype software. They were so impressed that 6 weeks after signing the contract the system went operational with almost minimal changes. Now this is not just your average army unit either - the local Special Forces (Oz SAS - US equivalent of the Rangers/SEALs). The success of this is such that we roll out new versions of the software every 6 weeks which they happily install and run with. Try doing that with native code software (Yes, we used to do really big monolithic Ada/C software here as well so I do know how long it takes).
Sun, by contrast, has never given up control of the language. Is sun going to decide to raise prices or terminate a license on the DOD in the middle of an operation?
That has got to be the biggest POS that I have read in a long time. Maybe you should actually look at Java at some time. Sun controls the language specification. They aren't the only supplier of runtime environments or development environments. There is no dependency on Sun. How can you pull a license to a specification when one was never granted in the first place? What do you think the CLASSPATH, Blackdown and other similar projects are all about. You might as well argue about what would happen if Borland pulled the license on Delphi in the middle of an operation, or MS with VB (there is more military code written in VB than ADA).
Next time, grab a clue stick before talking about something you obviously have NFI about.
According to a number of different, collated sources and things I saw at Siggraph, the reason the PSII is so fast is because the CPU is devoted almost entirely to the needs of graphics. This includes things like vector and matrix mul-add instructions (just like a DSP chip) and putting the entire thing on a single chip. Internally, the bus width is either 2560 or 4960 bits (can't remember exactly now). The real specs are around 66 million textured and lit polys per second. That is, texturing and lighting make almost no speed difference due to the huge bus width.
Considering that the top of the line IR2 does around 10-12M polys a second the Nvidia chip is still going to fall far behind. Why? Well the IR2 does almost the entire GTXSR pipeline on separate hardware. Consumer level cards don't do that. Although the TNT2 et al claim to do 5-6M polys/sec that is only if nothing else needs to be done. Add in some lighting calcs, alpha blended textures and partially obscured polygons that a real world app uses and watch those figures plummet (due usually to bus bandwidth or CPU limitations). The current cards are raster only. Geometry accelaration will help, but it won't go all the way to beating an SGI pipeline. Besides, if I need more graphics grunt for my SGI, I just add another card or two and I instantly double/triple/etc my polygon performance.
At Siggraph, there was a really interesting panel about large scale data visualisation. This was run by the people from Los Alamos, NCSA etc. They were talking about the PS performance wrt other architectures. They had some really interesting things to say about it - ie, expect to see (or not see, depending on your security clearance) Blue Mountain and the other ASCI computers using PSIIs and other similar boxes doing some of the real time interactive rendering of the data sets. That I find really amazing.
Now to show you how interested MS is in supporting unix, they have hired a 3rd party company to do the porting. Said (unnamed) company does not have any graphics knowledge - by MS's own admissions at the BOF. Kinda says something about it doesn't it.
I managed to get in a question at the end about how heavily MS was going to persue the licensing wrt to the OSS community developing an implementation (yup, I was the blue-haired one standing near the door). Their response - "umm... we hadn't thought of that. Might be worthwhile.". There were also a bunch of other statements about how well specified Fahrenheit is compared to the standard Win32 APIs. I just hope they are actually telling the truth this time and it's not another Chrome type effort. Given the reaction on the guy's face when I asked the question I tend to believe that they honestly hadn't thought about the spec issues and OSS implementations.
Other noteworthy things were that they have extensions for particular application areas. No-one in the MS Fahrenheit team has any decent Viz-Sim experience. They don't even have an extension underway to deal with this effort. Not particularly good if you want it to be used in large scale environments.
Lots of excuses on the MS part about not wanting to speak for SGI and their commitments to the IRIX side of it (the BOF was held a couple of hours after SGI announcement of the big culling going on).
In those days it was even a case of write your own toolkit. No such things as drop down menubars, just a mouse popup. To create a menu, first create a standard popup, fix it in place and go from there, You even had to do a lot of your own background buffering for repainting of windows after the menu was hidden. You didn't even have generalised scrolling areas - only text. Want a graphics scrolled area - write your own.
It was fun (I was at Uni at the time) but there is no way I would want to use Plan9 for a commercial application. Just to write a file manager it too 3 of us 6 months of heavy hacking.
The one thing that I loved about the Plan9 stuff was 9term. Love that cut & paste of the entire window contents, not just the last command line. Also, sam was a pretty decent editor too.
Personally, I'm already queued to get Shadow. Loved the original's so much that I just had to get this. Got a local mob that flys out everything the day it is released and only cost A$1 more for the privilege. Hmmm... can't wait!
Just to add one more rambling point - I'd love to see another view of the universe that looks from Peter's and Valerie's perspectives. There's a big hole there about how Peter became Hegemon and Ender the Speaker for the Dead that I'm dying to find out what happened.
A lot of the engineers inside Sun are pissed at the SCSL in a big way - particularly the commercial clause. At JavaOne, a number of different BOFs that I attended, when Sun lawyers were present, they copped a hammering about the SCSL (admittedly I was one of the people doing said hammering). From memory the SCSL and JINI BOFs were the two most problematical for their legal types.
Here's what I know is going on inside at the moment. The Java Media folks are pushing through a version of SCSL that does not include the commercial clause. That is, you can pretty much do anything you like with it - no compatibility checks, no license fees etc. Take the source and run. I imagine this would bring it extremely close to satisfying the open source guidelines. Naturally the Sun legal people are having a hard time swallowing this so it is taking some time to get through (It's been going at least for the last 3 months that I know of). Note that this is not changing the core SCSL, just a variant of it minus the crappy bits (like the Jini License is a variant of the SCSL).
The first of the Java Media APIs to undergo this is JSDT (Java Shared Data Toolkit). Once this goes through, pretty much all of the rest of the Java Media stuff will be following in quick succession. (Java3D, Media Framework, Sound and Advanced Imaging).
BTW - This is not leaking any private information. The Sun people at the JSDT BOF were the ones talking about this (Rich Burridge and Mike Bundschuh (sp?)) out in the open. Also, directly talking with the various Java Media marketing types and engineers running the stands there were talking about it too.
As an author and a coder I'd have to partially disagree with this.
By far the majority of computer books that are written come from an idea from the publisher. That is, the publisher will have an idea for a brand of book (eg the Dummies series) and then develop titles for it. This is usually done in conjunction with someone knowledgable in that area (usually the writer of the first book). The various titles are developed, a rough outline of the topic coverage and a list of potential authors. At this point, the development editor(s) (sometimes called acqusition editors) will start recruiting writers either by direct contact or through an agency. Now, if you approach the publisher first with a completely outlined, highly detailed (and preferably at least half written) book, then you have some good cause to ask for a different copyright scheme. A lot of the time you will get it too - it just depends on who the publisher is (O'Reilly is much more favorable to this than say Addison-Wesley).
Once an interested author has been found, the title is then developed to a much finer detail with the development editor etc until it is passed of to the series editor, tech editors etc for the actual writing.
Now the writer has a fair say in what happens with the book, but in general it doesn't go to the extents of saying "how would you like to license this book?". The publisher owns the copyrights on the basis that they solicited the work from you, not you proposing the work to them. Depending on your status as an author (prior recognition in your community etc) you can negotiate a reasonable degree of freedom. For example, I always get a release for all of the code (non-exclusive rights for redistribution yada yada yada...). This allows me to publish and further develop the code from the book as Open Source (usually LGPL'd because it is library type code). One of the benefits of this is that it helps to further enhance the author's reputation and also sells a lot more copies of the books.
(Side note: I've recently been contacted by a publisher that wants to do a Linux book under the Open Content licensing scheme. That was one of the first things they stated, so it does happen).
To prevent a few further comments - This code is all Java stuff or VRML. The documentation is always freely available online. What the book gives the reader is many of the inside tips of why things were done in certain ways. At least in my approach to writing I use the book to teach people how to approach designing stuff using the language APIs given to develop various types of libraries. I then proceed to develop a particular library, which is what ends up being the OSS product I put on my website. The book is not the design documentation or official API guide, but a way of approaching design philosophically. From most of the authors that I chat with, this is typically the case. The big exception to the rule is the "Bible" or "Nutshell" books that are just API references with a bit of extra stuff thrown in.
Although the VR version of cyberspace has not yet hit critical mass to start producing 3D versions (despite my, others valiant efforts with things like VRML and AlphaWorlds), the 2D version is very much starting to take on these Matrix-like effects. Some parts of the matrix are very popular (.com, .net, .org) while others are not (.us) while others are barely populated (.ro, .pt). People buy and sell these virtual addresses for some very huge amounts of money. Other people are the architects providing home building services, and down the bottom are a bunch of hackers like me trying to get the plumbing working.
What is happening now is the move from the 2D virtual property to the 3D. Although 3D is populated by game engines (and you can class, Isometric, 2.5D and 3D in this) eventually the same thing will spread to the general populace. As Katz points out, the arrival of the Middle-class into once what was hacker territory is a significant thing. We legitimize both the worth of the property and the fact that it exists. Stock options - how much virtual property are they? They are no more than a bunch of bytes on someone's harddrive these days, so what is the difference between that and a couple of pieces of armour and a good character.
The early adoptors always end up with a bunch of the most valued parts. Look around at the domain names, slashdot logins (I got in real early to get mine for example and the newer crowd haven't got such a range anymore. Maybe the login here will be worth something someday?). An interesting thing about these early adoptors is that normally they take the best stuff, not because it is, but because it reflects their personality or other trait that they like to express themselves with. Then, as time progresses, they become some of the most valuable simply because others want it. The gaming examples are just following the same trend already established in the flatlands. How long before some of these game environments start "meeting" at the seems?
The flip side to this is that, because we are digital, we can create as much property as we like. Just because a few games are hot today does not mean that they will be the only property available in a year down the track. Effectively limitless space can have some interesting effects on what and how items are valued. At what point do people come in to make a hedged risk on a new piece of virtual real-estate.
The funny thing that I find about all of this is that it will create jobs for the middle men. OK, so eBay is probably one of only a couple of sites on the 'net doing this now. How long do you think it will be before we start to see specialist virtual game property trading companies/sites being formed (darn, I should go out and patent that business idea!). That is, what we have in the real world will be mirrored in the virtual which will be mirrored in the real world... Just like the Matrix.
Rik has just moved the list over to majordomo (message came through all of 5 mikes ago). You can join the list by sending mail to
with the body
and then come join the fun.
Note that we are taking a completely different approach to tunelinux. We're breaking it down into functional areas rather than specific applications. For example, to tune a, oh.. let's say samba server, you need to tune SAMBA, the kernel and a number of other things to get the best from your box. If you take the tunelinux approach, you'll only end up with a 50% solution because you'll just tune SAMBA, and not the kernel. The Mindcraft report is a classic for that. Yes, tune samba, but hey, the kernel still only uses 960MB of RAM. The "tuning samba" doc approach wouldn't cover that sort of information all in the one spot. Effectively you end up with the same problem as we have now - lots of documentation but so fragmented that you can't do anything useful with it.
Rik van Riel has already started a linux tuning site over at nl.linux.org. Already there are some 50+ folks on the alias with documents starting to fly around the CVS system. Nobody has mentioned this site to us, so what's the deal?