Strictly speaking, those aren't really traffic jams, just weather delays. But that reminds me of an invention I saw a long time ago, an alarm clock with a remote moisture sensor. You put the sensor outside and if it rained or snowed, the alarm clock went off 30 minutes earlier than you set it. I never understood the reason you'd want something like this until I moved to LA, where it hardly ever rains and traffic gets snarled due to idiots who lose their ability to drive when it rains.
Well, I don't know if it still works, this was a few years ago. Back in those days, I lived downtown so I knew every shortcut and alley. My friends used to call me "The Duke of Downtown" (a lame reference to "Escape from New York"). The trick is getting out of downtown, there's a way through the tunnel (2nd street I think) that gets you into a tricky interchange under the freeways, you end up on Beverly. Take Beverly to the Pico-Rivera area and you have multiple choices from there, I usually stick to Beverly. It works fine the other direction, WLA to Downtown. The freeways into Downtown are hopeless, but Beverly is a straight shot into the core of the city. I remember I was working downtown once, in the middle of the financial district, when a customer in WLA called and yelled at me, I was late with some contracts and he had to have them signed in the next 30 minutes, and there was no way in hell I could get them to him in time. I said, "oh yeah? I'll be there in 15 minutes!" He said there was no way, I said, "oh yeah, just watch me!" and hung up. I hopped in my car and was at his doorstep in 15 minutes, he couldn't believe it.
QTSS, or the free multiplatform Darwin Streaming Studio, works great for streaming MP3s (and other formats too). Runs on Windoze, RedHat, MacOS X, and probably BSD and other Unix variants if you're good at porting. Has a nice web-based playist editor system and administration is easy. Runs well even on low-spec hardware.
Although he seems quite correct about CUPS, almost all other distributions I know of provide better UIs to configure it, and KDE ships an absolutely fantastic set of printing tools that rival those found in Windows and MacOS.
Apparently you aren't aware that MacOS X uses CUPS. It has a wonderful GUI that conceals the technical details beneath the candy colored Aqua shell.
This sort of discussion always amazes me, nobody thinks of the obvious solutions to avoiding traffic jams. Here are some solutions that worked for me. Notice that none of these solutions involve spending money on computer gadgets.
1. Get a new apartment that's closer to work. 2. Take mass transit. The subway never has traffic jams. 3. Live downtown, that way you're commuting the opposite way from rush hours and have the other side of the freeway all to yourself. You're trying to get out of downtown when people are trying to get in, and vice versa. 4. Change the schedule to avoid rush hours, work the night shift, or go to work at 5AM and work till 1PM. 5. Find alternate routes on surface roads that avoid freeway congestion. I knew a secret route between West Hollywood and Downtown LA that took only 15 minutes, freeway drivers can't even get off at the downtown exits in 15 minutes.
There is a special corner of Avichi Hell reserved for people who want to solve auto traffic problems with computers. That reserved spot will be right next to Bill Gates, who would today merely be a moderately rich lawyer at Preston, Gates & Ellis, except that one day he got intrigued by a computer programming problem in traffic measurement and founded his first company, Traf-O-Data.
Hey, this wasn't in v4 was it? It must be new in v5. I haven't done serious particles since v4, I learned to avoid them due to the pain involved.
So there are new workarounds, but it looks like there are still problems for render farms. Like for example, the docs say it's endian-dependent, so I can't whip up a particle cache on my Mac and run it on a PC render farm.
Your point is essentially true for the bulk of the CG biz (and it's the same argument I was making) but it isn't totally true. There are companies like ILM and Pixar that have self-sustaining business models. They even own their own render farms. But those studios are definitely the exception. WETA is no ILM or Pixar, they're the upstart trying to finance their projects with their credit cards (so to speak) as must now be obvious by their need to raise cash by renting their farm.
That's a very complex subject, but it boils down to this: there's no technical reason you can't precalculate all the particle positions over time, but that's not how they did it in Maya.
Let's say for example, you set a particle generator to run 60 frames, emitting smoke from a point, like from a cigarette tip. Smoke particles start emitting on frame 1, and continue on their path, particles persist through frame 60 as they drift upwards in a path influenced by random air currents. If you roll forward to frame 30 and render the last half only, you start all over with no smoke from the first 30 frames, it starts from scratch, they emit right from the tip in a new smoke trail, there's no history of past particle movement. So you'd get a huge discontinuity if you rendered the frames in batches. As far as I can tell, the actual image rendering doesn't influence the positions of the particles. It's just that they're calculated sequentially as each frame is rendered. Yeah, it's a huge pain and there aren't many good workarounds. But that's what you have to work with in order to use the particle generators, which are hugely powerful. Its the worst possible method, except for everything else anyone's ever thought of.
The point isn't that WETA is going to go bankrupt. It's that their render farm has a bad business model, it's going to suck up money. It's unsustainable. They wouldn't be trying to rent out the farm if they had a current business model that was sustainable. I don't know how WETA's structured, but if it's anything like Hollywood, it's all phony accounting and even the most profitable movies have zero profits on paper, it gets siphoned off by investors and never gets plowed back into the production companies. That's just how it works. So the production companies have to stand on their own two feet financially, they have to make a profit on their own.
The article says their primary advantage is that people can use WETA's proprietary software which kind of defeats the whole idea of outsourcing your render to them, you can't set it up yourself and let it run on their hardware without having WETA do it for you. You can't just rent the farm, you have to rent the whole company and outsource the design too.
And BTW, whoever modded my oringinal msg down as flamebait is abusing their privileges, someone mod it back up. It may be unpopular to criticize WETA, being godlike figures to LOTR fanatics and all, but I actually worked in the industry and know how it works, unlike 99.99% of/. users with mod points.
If you really did the risk assessment, you'd give up the stressful job you work at to earn money to buy overpriced computer hardware, it will give you a heart attack. You'd find a job without a daily commute where your risks of traffic death are high, you'd go work on a chicken ranch in Montana, selling eggs. Or maybe you should move to Alaska and become a hermit that avoids all human contact, so you don't pick up communicable diseases like influenza or AIDS.
You don't get it. They're going to throw away more money trying to rent the server farm than if they just send them to a toxic waste dump in China. Unless, of course they can get nutball otaku like you to buy the old blades at extortionate prices as your little trophy, just so you can say "hey dude we can host your pr0n website on a machine they used to make LOTR!"
You obviously have never worked in CG. Many common, simple effects cannot be parallelized. For example, Maya's particle effects are notorious for their inability to be parallelized and run on render farms, if they use randomness (and most particle fx do use randomness in positioning). Those fx must be rendered sequentially on a single CPU. Each frame's particle positions are used to calculate the next frame's particle positions, they're all calculated at runtime.
And the big liability here is the same, it's WETA's farm. It is obvious they didn't really think it through when they set up this farm.
My buddies in Australia and NZ tell me their internet costs are extremely high, they pay little for incoming data but pay up the wazoo for outgoing data. So let's say I zap a few gigs of model and map data from my studio in Hollywood, no problem, but when they want to send back terabytes or more of uncompressed, rendered frames, it's going to cost a whole lot more than that render farm down the street which is easily accessible by cheap US networking, or I could just drive in and pick up a stack of Exabyte tapes. No wonder you heard tales of WETA fedexing iPods around with compressed rushes. I worked with a lot of Hollywood producers and they're always in a hurry, Fedex from NZ isn't going to work too well.
One big lesson I learned in Hollywood is that you'd have to be an idiot to actually BUY a huge CG render rig. Everyone who builds a big rig is absolutely the state of the art when they BEGIN the project. But by the time they finish, the rig is obsolete, the studio down the street with the NEW project is now the state of the art. I've seen CG studios take huge losses on Big Iron like Connection Machines, Crays, etc, they were all acclaimed for their work, but after the big project, they all limped along trying to rent out the CPU for other projects, until they went bankrupt after a few more months. So it makes no sense to build your own render farm when you can lease it and create a virtual studio. Why accumulate hardware, which depreciates so quickly? Boy did WETA fuck up on this one.
Using your backassward logic, it seems more logical to devote your CPU time to researching automotive traffic patterns, so you don't get killed in an auto accident or get hit by a bus.
I remember performance artist Chris Burden set up a huge piece called "scale model of the solar system" way back in the early 1980s. I saw a picture of a couple of the locations, I think the sun was down in Newport Beach, Mercury was installed at a Ferrari dealership in Beverly Hills, etc. Burden even went to the trouble of calcluating the REAL positions of the planets on the opening day of the exhibit, and positioning the planets on the earth in relation to position, not just distance. You would have needed to travel all across southern California to see all the "planet" sites. California is a lot bigger than England, I think these new guys haven't got a chance of beating the size of previous efforts.
"about on par"...?? You're stoned. Yes, RH comes with Apache. It's as easy to set up as.. installing RH. Which I've done but your grandma couldn't. In contrast, here is how you set up an iMac to run Apache: 1. Turn machine on 2. Click "personal web sharing" 3. There is no step 3.
You apparently are incapable of comprehending this simple point: there is a huge difference between a program and a packaged product. A program is a package full of problems to solve to get it to work; a product is a solution to your problems.
You're not paying attention. Apache and Linux are programs. They don't come with a CPU, hard drive storage, RAM, and other essential features that make it a product. An iMac running Apache (it's part of the default install) is a product. Provide the hardware, support, documentation, extensively user-tested ease of use and quality control, easy configuration and setup, etc, and you're on your way to creating a product. Let me make this even simpler for you. TiVo is a product. Do it yourself Linux PVR software is a program.
Excuse me, but Apache is not Linux, you're way off in left field here. Apache runs on Linux, Windows, MacOS X, and probably a few others. But the "ease of use" of Apache is closely related to the OS. On Linux, I wouldn't be surprised if your 2+ yr newbie Linux admins had trouble installing and setting up Apache. On MacOS X it's in the default install, to start it up you just go to System Preferences, click one little box marked "Personal Web Sharing" and it's running. Your grandma could get Apache running on MacOS X. And there is the crux of the problem right there. As I have commented elsewhere in this topic, Linux (and Apache) aren't products, they're programs. A box that I can run my business website is a commercial product. A box that someone can use to publish their grandkids' photos is a consumer product. Stop making programs and start making products.
There you go. People don't want Linux products. They want products. Linux isn't a consumer product. A PDA, a TiVo, a file server are products. Nobody gives a damn that their TiVo runs on linux. Quit trying to sell linux and get busy selling products.
Since this got modded up, I suppose I should point out that.Mac supports WebDAV and you can upload HTML content from any platform, and of course the webmail is universal as well.
Right. I described that a little bit unclearly, sorry. Obviously I meant that the email headers, which (usually) indicate the text encoding, must be used by the webmail system to present the correct code type to the browser. I even made an error, Japanese is ISO-2022, oopsie. That's what I get from going off the top of my head, there are too many damn codes in this world to remember them all with precision.
You have to learn to let go. I remember reading an interview with the head of a courier company. He started all by himself, the sole employee, and built up a big clientele. He prided himself for never ever losing a single package, and he did his utmost to always deliver on time. But as his clientele grew, he had to hire more couriers. And suddenly, the new employees occasionally lost packages and were behind schedule. He finally came to the conclusion that nobody would ever be as conscientious at the job as he was, and he had to take human nature into account, and built procedures to allow for human error. And most of all, he had to learn to let go of his tendency to be a control freak.
You're missing a lot here. The Honda ad has scenes lifted directly from The Way Things Go, like the tires counterweighted so they "roll uphill," or the wiper blade motors that move exactly the same way as the mechanized shoes. It was a direct copy. You are apparently unable to see beyond the literal interpretation of an artwork. Fischli & Weiss's film is about alchemy, notice that the devices have 4 basic themes, fire, air, water, earth. There is a real narrative in the film, if you look for it. There is far more in this film I could explain, but I am probably casting pearls before swine.
After repeated viewings, I can see the feet move. But it's still a fake. Multiple light sources can't account for the (relatively) distinct shadow on the bot and extremely diffuse shadows elsewhere. The CG uses a standard "shadow catcher" plane and composited the shadow plane onto the floor.
You want to believe. Go ahead. This fraud has been debunked so thoroughly, if you want to look like a gullible fool, be my guest.
Slashdot Mirror
Strictly speaking, those aren't really traffic jams, just weather delays. But that reminds me of an invention I saw a long time ago, an alarm clock with a remote moisture sensor. You put the sensor outside and if it rained or snowed, the alarm clock went off 30 minutes earlier than you set it. I never understood the reason you'd want something like this until I moved to LA, where it hardly ever rains and traffic gets snarled due to idiots who lose their ability to drive when it rains.
Well, I don't know if it still works, this was a few years ago. Back in those days, I lived downtown so I knew every shortcut and alley. My friends used to call me "The Duke of Downtown" (a lame reference to "Escape from New York").
The trick is getting out of downtown, there's a way through the tunnel (2nd street I think) that gets you into a tricky interchange under the freeways, you end up on Beverly. Take Beverly to the Pico-Rivera area and you have multiple choices from there, I usually stick to Beverly. It works fine the other direction, WLA to Downtown. The freeways into Downtown are hopeless, but Beverly is a straight shot into the core of the city.
I remember I was working downtown once, in the middle of the financial district, when a customer in WLA called and yelled at me, I was late with some contracts and he had to have them signed in the next 30 minutes, and there was no way in hell I could get them to him in time. I said, "oh yeah? I'll be there in 15 minutes!" He said there was no way, I said, "oh yeah, just watch me!" and hung up. I hopped in my car and was at his doorstep in 15 minutes, he couldn't believe it.
QTSS, or the free multiplatform Darwin Streaming Studio, works great for streaming MP3s (and other formats too). Runs on Windoze, RedHat, MacOS X, and probably BSD and other Unix variants if you're good at porting. Has a nice web-based playist editor system and administration is easy. Runs well even on low-spec hardware.
Apparently you aren't aware that MacOS X uses CUPS. It has a wonderful GUI that conceals the technical details beneath the candy colored Aqua shell.
This sort of discussion always amazes me, nobody thinks of the obvious solutions to avoiding traffic jams. Here are some solutions that worked for me. Notice that none of these solutions involve spending money on computer gadgets.
1. Get a new apartment that's closer to work.
2. Take mass transit. The subway never has traffic jams.
3. Live downtown, that way you're commuting the opposite way from rush hours and have the other side of the freeway all to yourself. You're trying to get out of downtown when people are trying to get in, and vice versa.
4. Change the schedule to avoid rush hours, work the night shift, or go to work at 5AM and work till 1PM.
5. Find alternate routes on surface roads that avoid freeway congestion. I knew a secret route between West Hollywood and Downtown LA that took only 15 minutes, freeway drivers can't even get off at the downtown exits in 15 minutes.
There is a special corner of Avichi Hell reserved for people who want to solve auto traffic problems with computers. That reserved spot will be right next to Bill Gates, who would today merely be a moderately rich lawyer at Preston, Gates & Ellis, except that one day he got intrigued by a computer programming problem in traffic measurement and founded his first company, Traf-O-Data.
Hey, this wasn't in v4 was it? It must be new in v5. I haven't done serious particles since v4, I learned to avoid them due to the pain involved.
So there are new workarounds, but it looks like there are still problems for render farms. Like for example, the docs say it's endian-dependent, so I can't whip up a particle cache on my Mac and run it on a PC render farm.
Your point is essentially true for the bulk of the CG biz (and it's the same argument I was making) but it isn't totally true. There are companies like ILM and Pixar that have self-sustaining business models. They even own their own render farms. But those studios are definitely the exception. WETA is no ILM or Pixar, they're the upstart trying to finance their projects with their credit cards (so to speak) as must now be obvious by their need to raise cash by renting their farm.
That's a very complex subject, but it boils down to this: there's no technical reason you can't precalculate all the particle positions over time, but that's not how they did it in Maya.
Let's say for example, you set a particle generator to run 60 frames, emitting smoke from a point, like from a cigarette tip. Smoke particles start emitting on frame 1, and continue on their path, particles persist through frame 60 as they drift upwards in a path influenced by random air currents. If you roll forward to frame 30 and render the last half only, you start all over with no smoke from the first 30 frames, it starts from scratch, they emit right from the tip in a new smoke trail, there's no history of past particle movement. So you'd get a huge discontinuity if you rendered the frames in batches.
As far as I can tell, the actual image rendering doesn't influence the positions of the particles. It's just that they're calculated sequentially as each frame is rendered. Yeah, it's a huge pain and there aren't many good workarounds. But that's what you have to work with in order to use the particle generators, which are hugely powerful. Its the worst possible method, except for everything else anyone's ever thought of.
The point isn't that WETA is going to go bankrupt. It's that their render farm has a bad business model, it's going to suck up money. It's unsustainable. They wouldn't be trying to rent out the farm if they had a current business model that was sustainable. I don't know how WETA's structured, but if it's anything like Hollywood, it's all phony accounting and even the most profitable movies have zero profits on paper, it gets siphoned off by investors and never gets plowed back into the production companies. That's just how it works. So the production companies have to stand on their own two feet financially, they have to make a profit on their own.
/. users with mod points.
The article says their primary advantage is that people can use WETA's proprietary software which kind of defeats the whole idea of outsourcing your render to them, you can't set it up yourself and let it run on their hardware without having WETA do it for you. You can't just rent the farm, you have to rent the whole company and outsource the design too.
And BTW, whoever modded my oringinal msg down as flamebait is abusing their privileges, someone mod it back up. It may be unpopular to criticize WETA, being godlike figures to LOTR fanatics and all, but I actually worked in the industry and know how it works, unlike 99.99% of
You obviously have no sense of irony. Or humor.
If you really did the risk assessment, you'd give up the stressful job you work at to earn money to buy overpriced computer hardware, it will give you a heart attack. You'd find a job without a daily commute where your risks of traffic death are high, you'd go work on a chicken ranch in Montana, selling eggs. Or maybe you should move to Alaska and become a hermit that avoids all human contact, so you don't pick up communicable diseases like influenza or AIDS.
Sheesh!
You don't get it. They're going to throw away more money trying to rent the server farm than if they just send them to a toxic waste dump in China. Unless, of course they can get nutball otaku like you to buy the old blades at extortionate prices as your little trophy, just so you can say "hey dude we can host your pr0n website on a machine they used to make LOTR!"
You obviously have never worked in CG. Many common, simple effects cannot be parallelized. For example, Maya's particle effects are notorious for their inability to be parallelized and run on render farms, if they use randomness (and most particle fx do use randomness in positioning). Those fx must be rendered sequentially on a single CPU. Each frame's particle positions are used to calculate the next frame's particle positions, they're all calculated at runtime.
And the big liability here is the same, it's WETA's farm. It is obvious they didn't really think it through when they set up this farm.
My buddies in Australia and NZ tell me their internet costs are extremely high, they pay little for incoming data but pay up the wazoo for outgoing data. So let's say I zap a few gigs of model and map data from my studio in Hollywood, no problem, but when they want to send back terabytes or more of uncompressed, rendered frames, it's going to cost a whole lot more than that render farm down the street which is easily accessible by cheap US networking, or I could just drive in and pick up a stack of Exabyte tapes. No wonder you heard tales of WETA fedexing iPods around with compressed rushes. I worked with a lot of Hollywood producers and they're always in a hurry, Fedex from NZ isn't going to work too well.
One big lesson I learned in Hollywood is that you'd have to be an idiot to actually BUY a huge CG render rig. Everyone who builds a big rig is absolutely the state of the art when they BEGIN the project. But by the time they finish, the rig is obsolete, the studio down the street with the NEW project is now the state of the art. I've seen CG studios take huge losses on Big Iron like Connection Machines, Crays, etc, they were all acclaimed for their work, but after the big project, they all limped along trying to rent out the CPU for other projects, until they went bankrupt after a few more months. So it makes no sense to build your own render farm when you can lease it and create a virtual studio. Why accumulate hardware, which depreciates so quickly? Boy did WETA fuck up on this one.
Using your backassward logic, it seems more logical to devote your CPU time to researching automotive traffic patterns, so you don't get killed in an auto accident or get hit by a bus.
I remember performance artist Chris Burden set up a huge piece called "scale model of the solar system" way back in the early 1980s. I saw a picture of a couple of the locations, I think the sun was down in Newport Beach, Mercury was installed at a Ferrari dealership in Beverly Hills, etc. Burden even went to the trouble of calcluating the REAL positions of the planets on the opening day of the exhibit, and positioning the planets on the earth in relation to position, not just distance. You would have needed to travel all across southern California to see all the "planet" sites. California is a lot bigger than England, I think these new guys haven't got a chance of beating the size of previous efforts.
Yes, you have great difficulty understanding a +5 Insightful post, which says more about you than about me.
Better cut back on the drugs, it's messing with your cognitive functions. Sober up, troll.
"about on par"...?? You're stoned. Yes, RH comes with Apache. It's as easy to set up as.. installing RH. Which I've done but your grandma couldn't. In contrast, here is how you set up an iMac to run Apache:
1. Turn machine on
2. Click "personal web sharing"
3. There is no step 3.
You apparently are incapable of comprehending this simple point: there is a huge difference between a program and a packaged product. A program is a package full of problems to solve to get it to work; a product is a solution to your problems.
You're not paying attention. Apache and Linux are programs. They don't come with a CPU, hard drive storage, RAM, and other essential features that make it a product. An iMac running Apache (it's part of the default install) is a product. Provide the hardware, support, documentation, extensively user-tested ease of use and quality control, easy configuration and setup, etc, and you're on your way to creating a product.
Let me make this even simpler for you. TiVo is a product. Do it yourself Linux PVR software is a program.
Excuse me, but Apache is not Linux, you're way off in left field here. Apache runs on Linux, Windows, MacOS X, and probably a few others. But the "ease of use" of Apache is closely related to the OS. On Linux, I wouldn't be surprised if your 2+ yr newbie Linux admins had trouble installing and setting up Apache. On MacOS X it's in the default install, to start it up you just go to System Preferences, click one little box marked "Personal Web Sharing" and it's running. Your grandma could get Apache running on MacOS X.
And there is the crux of the problem right there. As I have commented elsewhere in this topic, Linux (and Apache) aren't products, they're programs. A box that I can run my business website is a commercial product. A box that someone can use to publish their grandkids' photos is a consumer product. Stop making programs and start making products.
There you go. People don't want Linux products. They want products. Linux isn't a consumer product. A PDA, a TiVo, a file server are products. Nobody gives a damn that their TiVo runs on linux. Quit trying to sell linux and get busy selling products.
Since this got modded up, I suppose I should point out that .Mac supports WebDAV and you can upload HTML content from any platform, and of course the webmail is universal as well.
Right. I described that a little bit unclearly, sorry. Obviously I meant that the email headers, which (usually) indicate the text encoding, must be used by the webmail system to present the correct code type to the browser.
I even made an error, Japanese is ISO-2022, oopsie. That's what I get from going off the top of my head, there are too many damn codes in this world to remember them all with precision.
You have to learn to let go. I remember reading an interview with the head of a courier company. He started all by himself, the sole employee, and built up a big clientele. He prided himself for never ever losing a single package, and he did his utmost to always deliver on time. But as his clientele grew, he had to hire more couriers. And suddenly, the new employees occasionally lost packages and were behind schedule. He finally came to the conclusion that nobody would ever be as conscientious at the job as he was, and he had to take human nature into account, and built procedures to allow for human error. And most of all, he had to learn to let go of his tendency to be a control freak.
You're missing a lot here. The Honda ad has scenes lifted directly from The Way Things Go, like the tires counterweighted so they "roll uphill," or the wiper blade motors that move exactly the same way as the mechanized shoes. It was a direct copy.
You are apparently unable to see beyond the literal interpretation of an artwork. Fischli & Weiss's film is about alchemy, notice that the devices have 4 basic themes, fire, air, water, earth. There is a real narrative in the film, if you look for it. There is far more in this film I could explain, but I am probably casting pearls before swine.
After repeated viewings, I can see the feet move. But it's still a fake. Multiple light sources can't account for the (relatively) distinct shadow on the bot and extremely diffuse shadows elsewhere. The CG uses a standard "shadow catcher" plane and composited the shadow plane onto the floor.
You want to believe. Go ahead. This fraud has been debunked so thoroughly, if you want to look like a gullible fool, be my guest.