I was under the impression that the wmv family of formats was proprietary to Microsoft. I'd be shocked into next Tuesday if MS licensed those codecs to be used for a Linux-based player. How is it possible that FIC is doing this legally?
SED is much lighter than a monolithic CRT because it isn't one big vacuum tube that has to be supported from the sides. There are spacers every few scanlines, much as in a plasma screen, to hold the front glass up. So, the front glass can be much thinner than the glass on a big monolithic CRT. SEDs should weigh about as much as plasma screens. Of course, they're still vapor, so they're infinitely light:)
Previously mentioned power consumption is a little more interesting. When Sony made JumboTron stadium displays, they found that the CRT mechanism is an astonishingly efficient way of turning electricity into colored light with the characteristics they wanted. JumboTrons are huge arrays of tens- to hundreds-of-thousands of little CRTs. What makes normal CRTs such power hogs is that the electrons have to be accelerated agressively to fly the distance from the back of the tube to the phosphors, and 2/3 of the electrons hit the shadow mask, and so that energy is wasted.
Similarly, in LCD screens, in the best case 2/3 of the light of the backlight is wasted by the color filters (all the green and blue light is always blocked by the red pixels, for example) and typically at least 5/6th of the light is wasted (if the pixels are half-on).
For the SED, every electron generated lights up a bit of phosphor, and only those pixels that need to be lit have any power usage at all.
Again, it'll be great if they can actually be produced in quantity. Apparently they are planning to ink-jet print the electron-emitter array, so that shouldn't be too expensive. I'm sure that the first few years of production will be as expensive as plasma screens, but hopefully that price will come down. We'll see.
I've been following the story about the joint venture between Toshiba and Canon on SED TV's for some time. Apparently it has become somewhat more real, as shown in this article.
Apparently things are going well enough with the new factory that Toshiba is stopping plasma-panel production, and staking its future on SED TV's.
SEDs are like CRTs, in that they use electron guns to shoot electrons across a vacuum at a phosphor scren to generate light. The difference is that SEDs have a semiconductor-based electron emitter at each pixel. This allows the screen to be flat, shallow (a few centimeters) and relatively lightweight, while preserving the fast response, brightness, and wide viewing-angle of regular CRTs. Also, somewhat surprisingly, SEDs are significantly less power-hungry than plasma panels or even big LED screens.
Toshiba and Canon have built a factory to start building these TVs, and apparently they are going to be trickling into the market toward the end of 2005. I can't wait!
Back in 1980, Tracy Peterson was working with me at the NYIT Computer Graphics Laboratory. He had built, with funding from Alex Schure, a digital sound facility at the lab that was state-of-the-art for the time.
The heart of the system was a SoundStream digital audio box, that would do the D-to-A and A-to-D conversions in real time, with very low noise. Again, for the day, it was revolutionary.
We were working with a system that recorded data optically onto cards of the 3x5 for factor mentioned and shown in the article. It was extremely cool. The data was recorded in the same circular arcs as show in the photo, so I have to believe it was the same system.
The recorder and players both kept the media still, and scanned a laser across the surface -- unlike the ubiquitous CD and DVD systems today that move the media and keep the head relatively still. There was an ingenious system of five prisms that whirled above the data card, that scanned the laser across the surface in an arc of 72 degrees or so. When the beam reflected from one prism hit the edge of the card, the beam was shifted to the next prism.
I believe that the whole card was slowly shifted into and out of the player as it went from track to track, much like the laser is shifted along a rail in a CD player.
Apparently the biggest problem was registration -- getting the card to align as precisely as required and getting each of the five prisms aligned perfectly.
It was interesting technology, but the CD mechanism is much simpler, and probably should have won out (as it did).
Thanks ishmalius for pointing me to Mr Gilmour's journal. I lived in the Silicon Valley from 1982-1997, and I hadn't been aware that Gilmour's writing has shifted from the pure-tech stuff he was doing back then. I find both the current article and his other recent writing fascinating.
I have to say, though, that as soon as you say "both sides of an issue" you've already severly endangered good journalism. There are extremely few issues for which there are two neat sides, and casting issues in that way is a way of stop-think. Unfortunately, it is such a useful paradigm for politicians that few people can imagine any other mode of reporting, or even thinking.
While I completely sympathize with Ursula Le Guin, and am horrified that the people behind the miniseries didn't see fit to work with her on the project, I do have an interesting story to relate that puts perhaps a different light on this.
My daughter is the world's greatest young artist. Well, in the top ten, anyway. It's been years since she had an art teacher that she had much respect for. So, this year, she gets to high school, and on the first day she waltzes into her new art teacher's office and says "I think all those people who analyze artist's paintings for meaning and symbolism are full of shit! The artists are just painting pictures! People who say otherwise are just making stuff up!"
Her new teacher says "OK, give me your sketchbook". My daughter carries it religiously everywhere, and hands it over. The teacher proceeds to tell my daughter everything about every detail of the drawings. The teacher explains in embarrassing detail my daughter's relationship toward each of the subjects -- things like "Hmmm -- you really like this person, but you just had a big fight with them." The teacher describes exactly what my daughter was thinking as she drew each part of each of the drawings. My daughter is, of course, thunderstruck -- the teacher got everything exactly right -- and my daughter had absolutely no idea that it would have been possible.
With that kind of an introduction, perhaps it is surprising that they have a truly wonderful student-teacher relationship now, but they do. I expect great things.
I believe that porting KDE tools to Windows would be a good thing, for any number of reasons.
1. It lets people who would never abandon Windows cold-turkey sample the benefits of open-source software. Note how well Firefox is doing, it is actaully making significant inroads against a monopoly MS app, and people like it. [off-topic, but I don't know if this has happened before -- that any entrenched Microsoft monopoly app has had its market share eroded.]
2. It will encourage openness within the Windows-using community. To get significant functionality, the KDE programming team will have to reverse-engineer some basic Linux operating system essentials -- to all Windows programmers benefit.
3. It will encourage MS to make better tools. Regardless of how much one hates MS (and I defy anybody to stand before me on that score), it's a good thing for them to have competition.
4. Office is the lifeblood of MS, not Windows. If MS has a monopoly on applications for Windows, they will continue to amass a warchest that will enable them to buy the planet. OEM Windows brings in just a few dollars/sale (a lot of sales, to be sure) but Office brings in a larger share of revenue. If you want to attack them MS it really hurts, go after their Office sales. Again, look at Linux desktop vs. Firefox market penetration. While the Linux desktop has progressed spectacularly well over the last few years, the market penetration is small because it forces people to make a leap they are uncomfortable making. Firefox has exploded into the market, because it's an easy step to make.
I have two big fears, though, with this move.
1. It's going to take significant resources to do this. The KDE programmers are remarkably prolific, but a Windows version of everything will be a major undertaking, and will inevitably slow development of the Linux versions. Jumping through the (small, moving, flaming) hoops that MS will require to get the functionality and performance that the KDE people will demand could well be seen as -- if not a waste of effort -- then an huge effort with little reward.
2. Lock-in. It's not out of the question by any means that MS will attempt to creatively co-op the KDE team by building in cool, but MS-only features, that would then be hard or impossible to port back to Linux -- for patent or other reasons. It would be a disaster (to me) if the Windows version of KDE tools were better than Linux versions.
Anyway -- I believe that the risks are worth the benefits. As I am just a devoted KDE user, but a non-contributor, I of course will defer to the good judgments of the KDE developers, I just hope to stimulate debate a little bit.
I agree with the parent poster that it's quite difficult to see a way that SGI could have survived as a workstation-class computer company, even with the benefits of hindsight. I bought a few million dollars worth of SGIs while working for PDI and Hammerhead, and worked at SGI in '91-'92, and they really were the only machines in their class at the time.
SGI did, in fact, attempt to sell PC graphics cards back in the early 90's, but it was an almost complete failure -- so you can't say that they didn't try to get into that market. I agree with the parent poster, again, that SGI wasn't prepared for the incredibly short product cycle that is the PC market. As an example of that, I was talking to an engineer at Nvidia at Siggraph this year (an ex-SGI guy, as many of their senior engineers are) who was well into the design of their chip for the second generation past the just-released NV40 chip -- SGI didn't have a corporate, research, or financial structure that would support that.
What finally killed them in the small computer market was games. Games pumped billions of dollars into the PC card market, and with those billions the PC card companies could create companies that could (and had to!) innovate like lightning. Once SGI got behind in performance, it was all over -- because that was their only advantage. The fact that their software was quirkly and deeply proprietary (although often state-of-the-art) was a significant burden.
I think it's just the way it goes. In almost every respect, the magic of SGI still survives, it's just that the employee nametags now say "Nvidia" and "ATi" insteady of "SiliconGraphics". The corporate structures that constrained these guys while they worked at SGI no longer apply, to the benefit of all of us.
There is a bright cloud in the upper right, but I believe that the sun is to the upper left. It's impossible to say, except for the shadow-beam:)
Seriously, this looks so exactly like a typical shadow-beam that I can't believe it's anything else. Note well that the beam ends at a cloud. I would be extremely surprised if the cloud wasn't long in the direction between the viewer and the cloud.
Others have noted that the beam doesn't show up in the previous or subsequent photos. These shadow beams are typically of short duration, as the geometry of the sun, cloud, and veiwer needs to be fairly exact for a high-contrast shadow beam to be visible, and the clouds move pretty fast.
I've attempted to use sun-angle calculators, but haven't succeeded yet. Still, if you are looking east from the coast of Australia at 18:52 on Nov 22, as this picture apparently is, you would expect the shadow to be at an angle that matches what is seen in this picture very closely. (note that it is summer in australia now, and I believe they have a kind of daylight-savings time, so 18:52 is the middle of the evening.)
You can see this dark streak almost every day in southern california, or almost anyplace that has contrails visible in the sky. When the contrail goes between you and the sun, you can see a dark band coming down from it. Watch for it!
Basically, what you are seeing is the equivalent of a sunbeam, except that it's a shadow-beam. A sunbeam occurs when there is a small hole in the cloud, and the light going through the hole illuminates the dust particles and water droples in the air along the path of the light. If the light is strong and the background relatively dark, it is easy to see these sunbeams (or God rays.)
Shadows through the sky are somewhat harder to see, because the contrast is not as great. When they are dramatic, as in this picture, you have to have the fortuitous situation of looking through a long, well defined slab of shadowed air, with well-lit air on either side. Airplane contrails are the perfect shadow source for this.
Imagine a 3D volume of a shadow cast by a contrail. It is a long thin slab of shadowed air. If you are within that slab, and looking along it, all the air in that direction is shadowed, for many miles, so the contrast between the shadowed air and the surrounding air is strong.
A good bit of the light around the shadow beam is not light scattered by dust or water droplets, but is just the same Raliegh-scattered light that makes the sky blue. The dark streak through the sky will be noticably darker and especially less blue than the surrounding air.
As you can tell, this is one of my favorite (of many!) atmospheric optical phenomena. Once you start to look for them, they are quite easy to see. Occasionally you can see them from natual cloud formations or even topographical or architechural features when the conditions are just right.
You've gotten a tremendous amount of good advice from long-time contractors, and I'd advise you to take it to heart. Taxes, insurance, and retirement are things that you will have to do on your own, and that's a substantial burden.
On the other hand, you're young and have no kids. This is a time for experimentation. If you are really hot shit, then you will thrive as a contractor. If you were working for some big company, you'd never be appreciated (read: compensated) for that, most likely. As a contractor, you can really see what you can do, in your most ideal work environment.
If it doesn't work out, oh well. You can try to get a more stable job somewhere else. Note well those that have said in previous comments that the idea of a "stable job" is almost as extinct as the dodo and passenger pigeon.
I own a small visual effects company, and pay my contract employees more than I take home myself -- and happily do it! No kidding, it's the way of the future, and can work out well for all concerned.
I know for a fact that EA management is aware of these three articles in Slashdot. They're not stupid, and they don't have their collective heads in the sand.
One must wonder, then, why they let these articles go unchallenged? Would Roblimo post an article from so high-up EA manager describing their point of view? Of course he would, it's a guaranteed 1,000 comment article!
So, one has to conclude that EA is being intentionally silent. The only reasons that I can think of for doing that are:
1. They can think of nothing to say, they have no justification for the practices that seem borderline illegal and certainly cruel.
or
2. They like the articles, and enjoy the reputation for complete balls-out insane production schedules.
HoshiToshi -- thanks for the links. I hadn't seen these, and I can't say that I have anything more informative than what I've included in the previous articles.
The flat-out-rave from guidetohometheater.com is extremely encouraging. The reduced power consumption with respect to plasma screens and even LCD screens is also quite exciting. I'm moving from cautiously optimistic to panting with anticipation.
Canon has a page that describes their proposed system. They describe using a similar ink-jet system for building the semiconductor array.
Admittedly, OLEDs, if they ever are practical, will be far more flexible, lightweight, just flat-out cooler than SEDs. But, the way that they keep moving the release date out a-year-per-year has me wondering if they will *ever* be practical.
I've been following the SED technology for quite some time. It seems to be finally moving from the labs into production. In this article, Toshiba and Canon announce the creation of a company to begin producing the displays, with limited production in 2005, ramping up after that.
SED displays are CRTs, after a fashion. They have electron guns that fire modulated electron beams through a vacuum at phosphor screens. As such, they have the brightness, color purity, and response rate of regular CRTs. What is different is that there is an electron gun for every pixel, instead of just one that is scanned across the screen. This allows the screen to be flat and shallow, and gives the geometric flatness and sharpness usually associated with LCDs.
This was attempted before with a slightly different technology, and went by the acronym/buzzword FED, for Field Effect Display. As this article points out, there was tremendous anticipation of this technology quite some time ago, they were planning to go into production in 1996. FED's had an array of tiny, very sharp needles behind the phosphor screen. Unfortunately, the production and maintenance of this array of needles proved to be next to impossible.
SED's use a much more producable and durable semiconductor array of electron guns. The technology of creating large, dense arrays of semiconductors on substrates has been developed and perfected by the LCD process, so I feel that there is hope this time around that the machines will actually be mass-produced on the aggressive schedule that Canon and Toshiba have laid out.
The first generation of SED's, it is claimed, will unfortunately not have the resolution that would make them good for computer displays or home TV's, as the spacing of the pixels will be somewhat large. They'll be used for business displays of various kinds. But, in the not-too-distant future (three-to-five years) Canon and Toshiba predict that SED's will come to dominate TV and monitor production.
Read the article. It's remarkably good, and makes a good case for temporary "cask" storage for a hundred years or so. There is little that you can say for certain about the future, but the one thing you can say is that it will be very different than the present, and different in unforseeable ways.
If you're really ambitious, read the Yucca Mountain reports from the goverment, available at John Young's indispensible cryptome.org among other places. The documents are amazingly detailed and well researched, and describe the truly monumental efforts proposed to make the best of the sadly misguided site that is Yucca Mountain. Radical alloys, glass matrices to bind the material, titanium drip shields, it just goes on and on and on. (The word "monumental" is actually literal, not just figurative. Part of the proposal describes the need for monuments to warn people away from the site for the next 10,000 years.)
The engineers and scientists working on Yucca Mountain were given the task to keep the amount of radiation leaking out of the site to low levels for 10,000 years. If everything goes exactly right, if there are no unforseen events, and the experimental materials they are using perform exactly as predicted under high radiation and hydrological stress for that time, the site will meet that mission. Astonishingly, the radiation release graphs go off the chart after 10,000 years -- there's still enough radiation there after that time to be terribly dangerous, and all protective measures will hae failed by that point.
Yucca Mountain was chosen and designed based on the assumption that it was dry. It's wet. That's such a huge difference that the original decision was simply wrong.
It's an appealing idea, but suffers from the slight problem of being completely wrong.
Indeed, natural uranium in the ground is really not very hazardous -- U235 is the most radioactive isotope, but is only a very small percentage of natural uranium and has a half-life of many millions of years. It's so benign that it was used as a pigment in early Fiesta Ware dishes and blue-blocker optical components (admittedly, it is not quite benign enough for these purposes...these have been recalled, but it's close.)
But, nuclear fission creates a spectacular kaleidescope of new isotopes. These are hundreds of thousands of times more radioactive than the natural uranium that was in the ground. It's true that they will only be extremely dangerous for a limited time, but that limited time is still in the many thousands of years.
While just reburying nuclear waste has some naive (although as show above, wrong) appeal, releasing them to the atmosphere is completely insane. This has been done already, in Chernobyl, on a relatively small scale. The area around the plant will be uninhabitable for a few thousand years.
What most people outside of California (and most inside California too, as people here have the average attention span of a three-day-old kitten) fail to realize is that Prop 13 was really the result of a series of California Supreme Court decisions in Serrano v. Priesta couple of years prior to it. That Supreme Court decision said that education was a state mandate, and that all schools must be funded equally. This was a dramatic departure from the status quo ante, which let the property taxes of relatively wealthy areas fund very good public schools in those areas, and left the poorer areas with significantly inadequate funding.
These court decisions, and the various legislative responses to them, led people in the relatively wealthy areas (read "people who care enough to vote") to limit their property taxes dramatically. In the end, school funding is equal -- yes -- but equal at a very low level. People remember the tax cut and the 2/3 vote necessary to raise taxes, but they don't remember or even know the political environment that caused it to happen.
The older Prius (pre-2004) was a tiny car -- in fact, it was basically a re-engined Echo.
The newer ones are far larger. It's quite a nice four-passenger car, with a reasonably roomy hatchback and other bits of storage space.
And, the new ones are more powerful and more fuel efficient than the older ones. You just cannot imagine the lengths that Toyota engineers went to to get the last few percent of fuel efficieny and pollution-control. Thermos bottles to retain heat in the coolant, carbon canister to trap startup hydrocarbons, drive-by-wire braking to do only regenerative braking until below 5 MPH, fins and baffles under the car to route air more efficiently...the list is almost endless. It basic Synergy drive, which throws in for free a CVT by basically using electricity the way other cars use transmission fluid, is the best known radical system, but it's only the beginning.
Ob-topic -- this is an insane scheme. I have to agree with the tinfoil-hat crowd that the only reason this makes sense is to get the GPS units into the cars for some other purposes -- like making it more expensive to drive through downtowns in rush hour (as they do in London, Singapore, and other rediculously congested cities.)
There are so very many ways that the State (and the state) benefit from more fuel efficient cars, that reducing the incentive to drive them is remarkably short-sighted. Treating the fuel tax as a carbon-dioxide tax really does make sense -- those Hummers and SUV's really do impose a cost on everybody else. Reducing the gasoline-delivery infrastructure is a good thing, from reducing the number of tankers that need to port in California, reducing the number of tankers on the roads, reducing the number of leaky gas tanks under service stations...these reflect costs on everybody that the gas tax goes some part of the way to paying for. Fewer kids dying of asthma would be a good thing.
If they want more money from the gas tax, they should just raise the tax.
Thad Beier
I feel that the professor let his students down.
on
A College Guide to EA
·
· Score: 4, Interesting
How could the professor not enlighten his students about the work schedule at EA, that from the previous two articles here is rather different than what might be expected? Several times, and in different ways, he states that you have to "work hard" and that EA is a "meritocracy" and that mediocre results will not be tolerated. That's all good, but your average CMU student is substantially brighter than most students (just an observation, I didn't go there) and probably feels that he would be able to excel at EA by working a normal, or maybe somewhat extended workweek.
I can well imagine that the student arriving at EA to the expectation that he will work 12/6 would feel blindsided. He does mention that there are "crunch times" before deadlines, but I would think that a little more elaboration on that topic would be appropriate for his students. The facts that crunch times seem to be scheduled even when projects are on track, that the extra hours are uncompensated by overtime pay, and that the ratio of "crunch time" to "down time" seems to be greater than one (based on admittedly biased, but believable comments here so far.)
It's got to be tough to be in his position -- appropriate jobs are hard to find for even the most qualified new graduates -- but presenting a balanced picture would be a good thing to do, IMHO.
Seriously, quitting is almost certainly painless to EA, as they can get other people to do the job pretty easily. Just send an email saying that you're only going to work 50 hours a week, and stick to it, and see what happens.
Because firing people has consequences. I run a small visual effects production company, and we hire freelance people as we get projects, for the length of the project. The State of California doesn't see it that way, though, and to the state it appears that we hire and fire people at a high rate.
This causes our unemployment insurance rate to be insanely high -- we pay about 10% of our employee's earnings into the state unemployment insurance system. Now, we consider that the cost of doing business -- we could even avoid it if we wanted to by various means but it does seem to us a reasonable price to pay for the privilege of hiring people just when we need them.
But, if EA's unemployment insurance rate skyrocketed, it'd hit them right in the wallet. They might even do something about it.
Just a suggestion. Any EA exec reading this (Hi!) can thank me privately -- as you must know, long term, that these "crunch" policies will destroy the company.
It turns out that a huge part of the film restoration/cleanup work is done using programs from MTI Right now huge film libraries are undergoing scanning and cleanup using MTI's products in preparation for high-def DVD release (once the standards get, uhm, standardized)
These MTI workstations might have highest software/hardware cost ratio of any widely deployed system. The hardware costs are a couple of thousand dollars, and the software is well over $50,000 per system. But, they get the job done like nothing else, and it is my experience that studios demand that particular software for their restoration.
It's not a completely automatic process by any means. The software can do a lot on its own, but it does require an artist to painstakingly review and correct the things that the software misses, or to guide the software to a correct solution.
A friend of mine who is building a large restoration facility would love to have a Linux solution, but unfortunately none exists at this point.
It's possible to build an extremely-low-functionality graphics card with open source software. It's so easy, that it's already been done. The 'nv' driver gives decent 2D performance from very cheap NVidia cards, and it ships with all modern Linux distributions. What's the peoblem?
What's not possible is to build competitive 3D cards. NVidia spent a few billion dollars making the NV40 chips. They're currently finalizing design on the NV60 range of chips, two generations in the future. What's truly amazing is that NVidia has found it in their interest to make extremely competitive Linux drivers for these chips/boards, and they give them away. And they "just work". While I deeply appreciate these Linux drivers, I cannot for the life of me understand the business case for spending a few to ten million dollars to write them.
Thad Beier
Disappointed in Pike's flip answer to patent Q
on
Rob Pike Responds
·
· Score: 5, Interesting
But I suppose it's not too surprising, considering the havoc that he and ATT wreaked upon X for Pike's save-under patent.
Save-under was/is a good idea, and so insanely simple it's hard to believe that a patent was granted -- much less weilded with such force. For youngsters (and as an oldster, perhaps my memory isn't quite perfect on this) some early machines had overlay planes for menus. You could draw the menu over the frame, then clear the overlay plane, without disturbing the contents of the window beneath. To do this on a bitmapped display without overlays, the idea was that you would screen-grab the image under where the menu would be, then paste it back when the menu disappeared.
Pike defended ATT's refusal to allow the X consortium to use save-under without royalty at the time.
Slashdot Mirror
I was under the impression that the wmv family of formats was proprietary to Microsoft. I'd be shocked into next Tuesday if MS licensed those codecs to be used for a Linux-based player. How is it possible that FIC is doing this legally?
Thad
SED is much lighter than a monolithic CRT because it isn't one big vacuum tube that has to be supported from the sides. There are spacers every few scanlines, much as in a plasma screen, to hold the front glass up. So, the front glass can be much thinner than the glass on a big monolithic CRT. SEDs should weigh about as much as plasma screens. Of course, they're still vapor, so they're infinitely light :)
Previously mentioned power consumption is a little more interesting. When Sony made JumboTron stadium displays, they found that the CRT mechanism is an astonishingly efficient way of turning electricity into colored light with the characteristics they wanted. JumboTrons are huge arrays of tens- to hundreds-of-thousands of little CRTs. What makes normal CRTs such power hogs is that the electrons have to be accelerated agressively to fly the distance from the back of the tube to the phosphors, and 2/3 of the electrons hit the shadow mask, and so that energy is wasted.
Similarly, in LCD screens, in the best case 2/3 of the light of the backlight is wasted by the color filters (all the green and blue light is always blocked by the red pixels, for example) and typically at least 5/6th of the light is wasted (if the pixels are half-on).
For the SED, every electron generated lights up a bit of phosphor, and only those pixels that need to be lit have any power usage at all.
Again, it'll be great if they can actually be produced in quantity. Apparently they are planning to ink-jet print the electron-emitter array, so that shouldn't be too expensive. I'm sure that the first few years of production will be as expensive as plasma screens, but hopefully that price will come down. We'll see.
Thad
I've been following the story about the joint venture between Toshiba and Canon on SED TV's for some time. Apparently it has become somewhat more real, as shown in this article.
Apparently things are going well enough with the new factory that Toshiba is stopping plasma-panel production, and staking its future on SED TV's.
SEDs are like CRTs, in that they use electron guns to shoot electrons across a vacuum at a phosphor scren to generate light. The difference is that SEDs have a semiconductor-based electron emitter at each pixel. This allows the screen to be flat, shallow (a few centimeters) and relatively lightweight, while preserving the fast response, brightness, and wide viewing-angle of regular CRTs. Also, somewhat surprisingly, SEDs are significantly less power-hungry than plasma panels or even big LED screens.
Toshiba and Canon have built a factory to start building these TVs, and apparently they are going to be trickling into the market toward the end of 2005. I can't wait!
Thad Beier
Back in 1980, Tracy Peterson was working with me at the NYIT Computer Graphics Laboratory. He had built, with funding from Alex Schure, a digital sound facility at the lab that was state-of-the-art for the time.
The heart of the system was a SoundStream digital audio box, that would do the D-to-A and A-to-D conversions in real time, with very low noise. Again, for the day, it was revolutionary.
We were working with a system that recorded data optically onto cards of the 3x5 for factor mentioned and shown in the article. It was extremely cool. The data was recorded in the same circular arcs as show in the photo, so I have to believe it was the same system.
The recorder and players both kept the media still, and scanned a laser across the surface -- unlike the ubiquitous CD and DVD systems today that move the media and keep the head relatively still. There was an ingenious system of five prisms that whirled above the data card, that scanned the laser across the surface in an arc of 72 degrees or so. When the beam reflected from one prism hit the edge of the card, the beam was shifted to the next prism.
I believe that the whole card was slowly shifted into and out of the player as it went from track to track, much like the laser is shifted along a rail in a CD player.
Apparently the biggest problem was registration -- getting the card to align as precisely as required and getting each of the five prisms aligned perfectly.
It was interesting technology, but the CD mechanism is much simpler, and probably should have won out (as it did).
Thad Beier
Thanks ishmalius for pointing me to Mr Gilmour's journal. I lived in the Silicon Valley from 1982-1997, and I hadn't been aware that Gilmour's writing has shifted from the pure-tech stuff he was doing back then. I find both the current article and his other recent writing fascinating.
I have to say, though, that as soon as you say "both sides of an issue" you've already severly endangered good journalism. There are extremely few issues for which there are two neat sides, and casting issues in that way is a way of stop-think. Unfortunately, it is such a useful paradigm for politicians that few people can imagine any other mode of reporting, or even thinking.
Thad Beier
While I completely sympathize with Ursula Le Guin, and am horrified that the people behind the miniseries didn't see fit to work with her on the project, I do have an interesting story to relate that puts perhaps a different light on this.
My daughter is the world's greatest young artist. Well, in the top ten, anyway. It's been years since she had an art teacher that she had much respect for. So, this year, she gets to high school, and on the first day she waltzes into her new art teacher's office and says "I think all those people who analyze artist's paintings for meaning and symbolism are full of shit! The artists are just painting pictures! People who say otherwise are just making stuff up!"
Her new teacher says "OK, give me your sketchbook". My daughter carries it religiously everywhere, and hands it over. The teacher proceeds to tell my daughter everything about every detail of the drawings. The teacher explains in embarrassing detail my daughter's relationship toward each of the subjects -- things like "Hmmm -- you really like this person, but you just had a big fight with them." The teacher describes exactly what my daughter was thinking as she drew each part of each of the drawings. My daughter is, of course, thunderstruck -- the teacher got everything exactly right -- and my daughter had absolutely no idea that it would have been possible.
With that kind of an introduction, perhaps it is surprising that they have a truly wonderful student-teacher relationship now, but they do. I expect great things.
Thad Beier
I believe that porting KDE tools to Windows would be a good thing, for any number of reasons.
1. It lets people who would never abandon Windows cold-turkey sample the benefits of open-source software. Note how well Firefox is doing, it is actaully making significant inroads against a monopoly MS app, and people like it. [off-topic, but I don't know if this has happened before -- that any entrenched Microsoft monopoly app has had its market share eroded.]
2. It will encourage openness within the Windows-using community. To get significant functionality, the KDE programming team will have to reverse-engineer some basic Linux operating system essentials -- to all Windows programmers benefit.
3. It will encourage MS to make better tools. Regardless of how much one hates MS (and I defy anybody to stand before me on that score), it's a good thing for them to have competition.
4. Office is the lifeblood of MS, not Windows. If MS has a monopoly on applications for Windows, they will continue to amass a warchest that will enable them to buy the planet. OEM Windows brings in just a few dollars/sale (a lot of sales, to be sure) but Office brings in a larger share of revenue. If you want to attack them MS it really hurts, go after their Office sales. Again, look at Linux desktop vs. Firefox market penetration. While the Linux desktop has progressed spectacularly well over the last few years, the market penetration is small because it forces people to make a leap they are uncomfortable making. Firefox has exploded into the market, because it's an easy step to make.
I have two big fears, though, with this move.
1. It's going to take significant resources to do this. The KDE programmers are remarkably prolific, but a Windows version of everything will be a major undertaking, and will inevitably slow development of the Linux versions. Jumping through the (small, moving, flaming) hoops that MS will require to get the functionality and performance that the KDE people will demand could well be seen as -- if not a waste of effort -- then an huge effort with little reward.
2. Lock-in. It's not out of the question by any means that MS will attempt to creatively co-op the KDE team by building in cool, but MS-only features, that would then be hard or impossible to port back to Linux -- for patent or other reasons. It would be a disaster (to me) if the Windows version of KDE tools were better than Linux versions.
Anyway -- I believe that the risks are worth the benefits. As I am just a devoted KDE user, but a non-contributor, I of course will defer to the good judgments of the KDE developers, I just hope to stimulate debate a little bit.
Thad Beier
I agree with the parent poster that it's quite difficult to see a way that SGI could have survived as a workstation-class computer company, even with the benefits of hindsight. I bought a few million dollars worth of SGIs while working for PDI and Hammerhead, and worked at SGI in '91-'92, and they really were the only machines in their class at the time.
SGI did, in fact, attempt to sell PC graphics cards back in the early 90's, but it was an almost complete failure -- so you can't say that they didn't try to get into that market. I agree with the parent poster, again, that SGI wasn't prepared for the incredibly short product cycle that is the PC market. As an example of that, I was talking to an engineer at Nvidia at Siggraph this year (an ex-SGI guy, as many of their senior engineers are) who was well into the design of their chip for the second generation past the just-released NV40 chip -- SGI didn't have a corporate, research, or financial structure that would support that.
What finally killed them in the small computer market was games. Games pumped billions of dollars into the PC card market, and with those billions the PC card companies could create companies that could (and had to!) innovate like lightning. Once SGI got behind in performance, it was all over -- because that was their only advantage. The fact that their software was quirkly and deeply proprietary (although often state-of-the-art) was a significant burden.
I think it's just the way it goes. In almost every respect, the magic of SGI still survives, it's just that the employee nametags now say "Nvidia" and "ATi" insteady of "SiliconGraphics". The corporate structures that constrained these guys while they worked at SGI no longer apply, to the benefit of all of us.
Thad
Dominux,
:)
There is a bright cloud in the upper right, but I believe that the sun is to the upper left. It's impossible to say, except for the shadow-beam
Seriously, this looks so exactly like a typical shadow-beam that I can't believe it's anything else. Note well that the beam ends at a cloud. I would be extremely surprised if the cloud wasn't long in the direction between the viewer and the cloud.
Others have noted that the beam doesn't show up in the previous or subsequent photos. These shadow beams are typically of short duration, as the geometry of the sun, cloud, and veiwer needs to be fairly exact for a high-contrast shadow beam to be visible, and the clouds move pretty fast.
I've attempted to use sun-angle calculators, but haven't succeeded yet. Still, if you are looking east from the coast of Australia at 18:52 on Nov 22, as this picture apparently is, you would expect the shadow to be at an angle that matches what is seen in this picture very closely. (note that it is summer in australia now, and I believe they have a kind of daylight-savings time, so 18:52 is the middle of the evening.)
Thad Beier
You can see this dark streak almost every day in southern california, or almost anyplace that has contrails visible in the sky. When the contrail goes between you and the sun, you can see a dark band coming down from it. Watch for it!
Basically, what you are seeing is the equivalent of a sunbeam, except that it's a shadow-beam. A sunbeam occurs when there is a small hole in the cloud, and the light going through the hole illuminates the dust particles and water droples in the air along the path of the light. If the light is strong and the background relatively dark, it is easy to see these sunbeams (or God rays.)
Shadows through the sky are somewhat harder to see, because the contrast is not as great. When they are dramatic, as in this picture, you have to have the fortuitous situation of looking through a long, well defined slab of shadowed air, with well-lit air on either side. Airplane contrails are the perfect shadow source for this.
Imagine a 3D volume of a shadow cast by a contrail. It is a long thin slab of shadowed air. If you are within that slab, and looking along it, all the air in that direction is shadowed, for many miles, so the contrast between the shadowed air and the surrounding air is strong.
A good bit of the light around the shadow beam is not light scattered by dust or water droplets, but is just the same Raliegh-scattered light that makes the sky blue. The dark streak through the sky will be noticably darker and especially less blue than the surrounding air.
As you can tell, this is one of my favorite (of many!) atmospheric optical phenomena. Once you start to look for them, they are quite easy to see. Occasionally you can see them from natual cloud formations or even topographical or architechural features when the conditions are just right.
Thad Beier
Oh my god!
This is the most magical thing I have seen in quite some time.
You've gotten a tremendous amount of good advice from long-time contractors, and I'd advise you to take it to heart. Taxes, insurance, and retirement are things that you will have to do on your own, and that's a substantial burden.
On the other hand, you're young and have no kids. This is a time for experimentation. If you are really hot shit, then you will thrive as a contractor. If you were working for some big company, you'd never be appreciated (read: compensated) for that, most likely. As a contractor, you can really see what you can do, in your most ideal work environment.
If it doesn't work out, oh well. You can try to get a more stable job somewhere else. Note well those that have said in previous comments that the idea of a "stable job" is almost as extinct as the dodo and passenger pigeon.
I own a small visual effects company, and pay my contract employees more than I take home myself -- and happily do it! No kidding, it's the way of the future, and can work out well for all concerned.
Thad Beier
I know for a fact that EA management is aware of these three articles in Slashdot. They're not stupid, and they don't have their collective heads in the sand.
One must wonder, then, why they let these articles go unchallenged? Would Roblimo post an article from so high-up EA manager describing their point of view? Of course he would, it's a guaranteed 1,000 comment article!
So, one has to conclude that EA is being intentionally silent. The only reasons that I can think of for doing that are:
1. They can think of nothing to say, they have no justification for the practices that seem borderline illegal and certainly cruel.
or
2. They like the articles, and enjoy the reputation for complete balls-out insane production schedules.
Think about it.
Thad Beier
HoshiToshi -- thanks for the links. I hadn't seen these, and I can't say that I have anything more informative than what I've included in the previous articles.
The flat-out-rave from guidetohometheater.com is extremely encouraging. The reduced power consumption with respect to plasma screens and even LCD screens is also quite exciting. I'm moving from cautiously optimistic to panting with anticipation.
Thad Beier
Canon has a page that describes their proposed system. They describe using a similar ink-jet system for building the semiconductor array.
Admittedly, OLEDs, if they ever are practical, will be far more flexible, lightweight, just flat-out cooler than SEDs. But, the way that they keep moving the release date out a-year-per-year has me wondering if they will *ever* be practical.
Thad
SED displays are CRTs, after a fashion. They have electron guns that fire modulated electron beams through a vacuum at phosphor screens. As such, they have the brightness, color purity, and response rate of regular CRTs. What is different is that there is an electron gun for every pixel, instead of just one that is scanned across the screen. This allows the screen to be flat and shallow, and gives the geometric flatness and sharpness usually associated with LCDs.
This was attempted before with a slightly different technology, and went by the acronym/buzzword FED, for Field Effect Display. As this article points out, there was tremendous anticipation of this technology quite some time ago, they were planning to go into production in 1996. FED's had an array of tiny, very sharp needles behind the phosphor screen. Unfortunately, the production and maintenance of this array of needles proved to be next to impossible.
SED's use a much more producable and durable semiconductor array of electron guns. The technology of creating large, dense arrays of semiconductors on substrates has been developed and perfected by the LCD process, so I feel that there is hope this time around that the machines will actually be mass-produced on the aggressive schedule that Canon and Toshiba have laid out.
The first generation of SED's, it is claimed, will unfortunately not have the resolution that would make them good for computer displays or home TV's, as the spacing of the pixels will be somewhat large. They'll be used for business displays of various kinds. But, in the not-too-distant future (three-to-five years) Canon and Toshiba predict that SED's will come to dominate TV and monitor production.
We'll see.
Thad Beier
If you're really ambitious, read the Yucca Mountain reports from the goverment, available at John Young's indispensible cryptome.org among other places. The documents are amazingly detailed and well researched, and describe the truly monumental efforts proposed to make the best of the sadly misguided site that is Yucca Mountain. Radical alloys, glass matrices to bind the material, titanium drip shields, it just goes on and on and on. (The word "monumental" is actually literal, not just figurative. Part of the proposal describes the need for monuments to warn people away from the site for the next 10,000 years.)
The engineers and scientists working on Yucca Mountain were given the task to keep the amount of radiation leaking out of the site to low levels for 10,000 years. If everything goes exactly right, if there are no unforseen events, and the experimental materials they are using perform exactly as predicted under high radiation and hydrological stress for that time, the site will meet that mission. Astonishingly, the radiation release graphs go off the chart after 10,000 years -- there's still enough radiation there after that time to be terribly dangerous, and all protective measures will hae failed by that point.
Yucca Mountain was chosen and designed based on the assumption that it was dry. It's wet. That's such a huge difference that the original decision was simply wrong.
Thad Beier
Saven,
It's an appealing idea, but suffers from the slight problem of being completely wrong.
Indeed, natural uranium in the ground is really not very hazardous -- U235 is the most radioactive isotope, but is only a very small percentage of natural uranium and has a half-life of many millions of years. It's so benign that it was used as a pigment in early Fiesta Ware dishes and blue-blocker optical components (admittedly, it is not quite benign enough for these purposes...these have been recalled, but it's close.)
But, nuclear fission creates a spectacular kaleidescope of new isotopes. These are hundreds of thousands of times more radioactive than the natural uranium that was in the ground. It's true that they will only be extremely dangerous for a limited time, but that limited time is still in the many thousands of years.
While just reburying nuclear waste has some naive (although as show above, wrong) appeal, releasing them to the atmosphere is completely insane. This has been done already, in Chernobyl, on a relatively small scale. The area around the plant will be uninhabitable for a few thousand years.
Some kind of waste treatment plan is necessary.
Thad Beier
These court decisions, and the various legislative responses to them, led people in the relatively wealthy areas (read "people who care enough to vote") to limit their property taxes dramatically. In the end, school funding is equal -- yes -- but equal at a very low level. People remember the tax cut and the 2/3 vote necessary to raise taxes, but they don't remember or even know the political environment that caused it to happen.
Thad Beier
The older Prius (pre-2004) was a tiny car -- in fact, it was basically a re-engined Echo.
The newer ones are far larger. It's quite a nice four-passenger car, with a reasonably roomy hatchback and other bits of storage space.
And, the new ones are more powerful and more fuel efficient than the older ones. You just cannot imagine the lengths that Toyota engineers went to to get the last few percent of fuel efficieny and pollution-control. Thermos bottles to retain heat in the coolant, carbon canister to trap startup hydrocarbons, drive-by-wire braking to do only regenerative braking until below 5 MPH, fins and baffles under the car to route air more efficiently...the list is almost endless. It basic Synergy drive, which throws in for free a CVT by basically using electricity the way other cars use transmission fluid, is the best known radical system, but it's only the beginning.
Ob-topic -- this is an insane scheme. I have to agree with the tinfoil-hat crowd that the only reason this makes sense is to get the GPS units into the cars for some other purposes -- like making it more expensive to drive through downtowns in rush hour (as they do in London, Singapore, and other rediculously congested cities.)
There are so very many ways that the State (and the state) benefit from more fuel efficient cars, that reducing the incentive to drive them is remarkably short-sighted. Treating the fuel tax as a carbon-dioxide tax really does make sense -- those Hummers and SUV's really do impose a cost on everybody else. Reducing the gasoline-delivery infrastructure is a good thing, from reducing the number of tankers that need to port in California, reducing the number of tankers on the roads, reducing the number of leaky gas tanks under service stations...these reflect costs on everybody that the gas tax goes some part of the way to paying for. Fewer kids dying of asthma would be a good thing.
If they want more money from the gas tax, they should just raise the tax.
Thad Beier
How could the professor not enlighten his students about the work schedule at EA, that from the previous two articles here is rather different than what might be expected? Several times, and in different ways, he states that you have to "work hard" and that EA is a "meritocracy" and that mediocre results will not be tolerated. That's all good, but your average CMU student is substantially brighter than most students (just an observation, I didn't go there) and probably feels that he would be able to excel at EA by working a normal, or maybe somewhat extended workweek.
I can well imagine that the student arriving at EA to the expectation that he will work 12/6 would feel blindsided. He does mention that there are "crunch times" before deadlines, but I would think that a little more elaboration on that topic would be appropriate for his students. The facts that crunch times seem to be scheduled even when projects are on track, that the extra hours are uncompensated by overtime pay, and that the ratio of "crunch time" to "down time" seems to be greater than one (based on admittedly biased, but believable comments here so far.)
It's got to be tough to be in his position -- appropriate jobs are hard to find for even the most qualified new graduates -- but presenting a balanced picture would be a good thing to do, IMHO.
Thad Beier
Seriously, quitting is almost certainly painless to EA, as they can get other people to do the job pretty easily. Just send an email saying that you're only going to work 50 hours a week, and stick to it, and see what happens.
Because firing people has consequences. I run a small visual effects production company, and we hire freelance people as we get projects, for the length of the project. The State of California doesn't see it that way, though, and to the state it appears that we hire and fire people at a high rate.
This causes our unemployment insurance rate to be insanely high -- we pay about 10% of our employee's earnings into the state unemployment insurance system. Now, we consider that the cost of doing business -- we could even avoid it if we wanted to by various means but it does seem to us a reasonable price to pay for the privilege of hiring people just when we need them.
But, if EA's unemployment insurance rate skyrocketed, it'd hit them right in the wallet. They might even do something about it.
Just a suggestion. Any EA exec reading this (Hi!) can thank me privately -- as you must know, long term, that these "crunch" policies will destroy the company.
Thad
It turns out that a huge part of the film restoration/cleanup work is done using programs from MTI Right now huge film libraries are undergoing scanning and cleanup using MTI's products in preparation for high-def DVD release (once the standards get, uhm, standardized)
These MTI workstations might have highest software/hardware cost ratio of any widely deployed system. The hardware costs are a couple of thousand dollars, and the software is well over $50,000 per system. But, they get the job done like nothing else, and it is my experience that studios demand that particular software for their restoration.
It's not a completely automatic process by any means. The software can do a lot on its own, but it does require an artist to painstakingly review and correct the things that the software misses, or to guide the software to a correct solution.
A friend of mine who is building a large restoration facility would love to have a Linux solution, but unfortunately none exists at this point.
Thad Beier
It's possible to build an extremely-low-functionality graphics card with open source software. It's so easy, that it's already been done. The 'nv' driver gives decent 2D performance from very cheap NVidia cards, and it ships with all modern Linux distributions. What's the peoblem?
What's not possible is to build competitive 3D cards. NVidia spent a few billion dollars making the NV40 chips. They're currently finalizing design on the NV60 range of chips, two generations in the future. What's truly amazing is that NVidia has found it in their interest to make extremely competitive Linux drivers for these chips/boards, and they give them away. And they "just work". While I deeply appreciate these Linux drivers, I cannot for the life of me understand the business case for spending a few to ten million dollars to write them.
Thad Beier
But I suppose it's not too surprising, considering the havoc that he and ATT wreaked upon X for Pike's save-under patent.
Save-under was/is a good idea, and so insanely simple it's hard to believe that a patent was granted -- much less weilded with such force. For youngsters (and as an oldster, perhaps my memory isn't quite perfect on this) some early machines had overlay planes for menus. You could draw the menu over the frame, then clear the overlay plane, without disturbing the contents of the window beneath. To do this on a bitmapped display without overlays, the idea was that you would screen-grab the image under where the menu would be, then paste it back when the menu disappeared.
Pike defended ATT's refusal to allow the X consortium to use save-under without royalty at the time.
Thad