It makes me sad when I see companys trying to hype up their research to pump up their share price. Now when my group does any research which has positive results we are scared to release anything because everyone assumes its simply another con. Currently we have an asynchronous processor which releases so little EMI it looks dead in the graphs. We tried showing this to other people but everyone nowdays refuses to beleve anything unconvesional can be good.
What do you mean? It should be -easy- to convince people your processor works the way you claim. Loan them some of your prototypes (after they sign a non-disclosure agreement) and let them test the processors on their own, with their own equipment.
If your measurements are valid, they will be able to replicate them easily. If not, you will learn what you did wrong quickly enough. Whatever you do, don't sit around blaming close-minded attitudes for your failure to convince people. Otherwise, you start to sound like crackpot researchers.
As to Dr. Schon, I suspect that he will be asked to replicate some of his measurements in the presence of trained observers. He should still have working parts, and if he can do so, he will be vindicated. Otherwise, his career is finished.
Chip makers complain because the "CAD Community" isn't coming up with solutions to some of their problems, but University R&D programs are unable to keep up with fabrication standards as the equipment gets more expensive. Isn't this a problem waiting for a few self-interested chip-makers to whip their wallets in the direction of a few universities?
As someone who has participated in CAD and microelectronics research in academia, I can offer a few informed opinions about this situation. Everyone has a opinion of what academia "should" be doing. Very few people take a hard look at -why- academia research in these areas has come to a standstill.
Research money for CAD development in particular and microelectronics in general is almost impossible to come by. For research in IC processing, the big factor is the cost of capital equipment. Universities simply can't keep up - only industry can afford the multi-multi-million dollar capital equipment costs. Even research in analog and mixed-signal circuit design has pretty much collapsed, simply for lack of funding to fabricate ICs through third-party fabs. What little money is available is quickly soaked up by a tiny number of high-profile researchers, mainly in California schools. Unfortunately, these guys can't even come close to meeting the demand for students and research output.
CAD is a somewhat different story. The capital cost of CAD research is actually pretty small. All you need are a few good workstations. The bigger problem is lack of qualified graduate students, and lack of any money to pay them. Everybody agrees that universities should do more CAD research, but the federal government won't write the checks (applied research should be funded by industry) and industry won't write the checks (too many intellectual property hassles with universities, plus no guaranteed return on investment). In fact, industry often acts as its own worst enemy by hiring away students and faculty both and eating the academic "seed corn" of future research.
To top it off, any CAD researcher with even a moderately innovative tool will do MUCH better by leaving academia, finding venture capital, starting his/her own company, getting bought out by Cadence, and retiring wealthy.
Graduate student recruitment has improved somewhat in the past year due to the recession, but as soon as the microelectronics industry starts to recover, you won't be able to hire a qualified graduate student or faculty member for love or money. It will take massive amounts of cash to change this trend, and frankly I don't see it happening here in the U.S. The microelectronics industry will have to rely on internal R & D from now on. Except for a few remaining high-profile programs, the academic sector is pretty much finished.
As an ECE faculty member, I can offer some observations of my own about the topics of faculty hiring and retention.
To begin with, faculty hiring in electrical and computer engineering is -damned- difficult right now, and won't be getting better anytime soon. Ten years ago you could advertise a faculty position and get hundreds of resumes, with at least half a dozen outstanding candidates in the pile. Nowadays you're lucky to get a dozen resumes, and typically most of them are a complete joke. Most ECE departments have had open positions for -years-, with little prospect for filling them.
The tech sector boom of the 90's, coupled with stagnant faculty salaries and declining budgets, has been the direct cause. Many CS and CE departments suffered enormous attrition during the dot-com boom. Unfortunately, the current recession did almost nothing to improve the situation, since (a) the really bright, good people have remained employed rather than return to university positions, and (b) the CS and CE departments are no longer cranking out Ph.D. candidates in significant numbers. Graduate enrollment by U.S. citizens in Ph.D. programs has simply collapsed. You can barely get domestic students to hang around for an M.S. nowdays, much less a doctorate. Industry money is just too good, even now.
So how do universities fill in the gaps? Adjunct positions are one way, but take my word for it - an adjunct professor is the lowest of the low at most schools. Rotten pay, long hours, and the best you can hope for is that you're left alone by the tenure-track faculty.
However (and this is the part many people do not realize), adjunct and part-time teaching help is politically and economically undesirable at most schools. Parents do not want to hear that their kids are being taught by part-time hired help. Furthermore, schools want full-time faculty who can pursue research grants and bring in the big government bucks. Adjunt professors are a necessary evil, but one we'd rather not have to deal with.
So what's the future of CS, EE, and CE faculties? IMHO, I think faculty hiring is going to go the way of graduate student enrollments over the next ten years. Nowadays the majority of graduate students in our department are international students. This is the norm even at the top universities in the U.S. My guess is that eventually U.S. schools will have to recruit most tenure-track faculty overseas, and sponsor their visas just as companies do. Granted, this is politically even more undesirable than hiring adjuncts, but you're not going to find qualified full-time professors who are able to pursue research funds otherwise.
After reading the article, I checked out Rainmaker's site. These guys have a theory, some patents, and some simulations. What they don't seem to have is any working hardware that proves this 10X bandwidth increase can actually be achieved in residential cable systems.
Does this remind anyone of Transmeta, who promised processors with a fraction of the power consumption at higher speeds? Everybody loved them when all they had was a press release. The actual product didn't work as advertised, and now they've faded away.
If it sounds too good to be true, it probably is. 10X uber-bandwidth schemes sound suspiciously like 10X uber-compression schemes. I'll reserve my enthusiasm when I see working hardware.
You know, doesn't this mean that all this other searching for extra-terrestrial intelligence [seti.org] is pretty counter-productive? If there's water right there on Mars, chances are there would be intelligent life there within a few billion years too. (It's the initial part of the thing that takes awhile...once you've got cells, the growth is like, exponential man.)
Although many people believe that life is very common throughout the universe, and will spread rapidly anywhere it gets a foothold, there's no deterministic path towards intelligent life. The prevailing view is that intelligence is an accident, not an end-product, of evolution. Many scientists believe that intelligent life is extremely rare, and that we may be the first intelligent civilization to evolve in the Milky Way galaxy.
After all, if a large meteor hadn't hit the earth about 65 millions years ago, dinosaurs would still be roaming the earth. Humanity is here due to a freak accident, and in fact we may be very unique.
MPEG-4 may be a better algorithm, but is it a suitable algorithm for real-time or short-turnaround compression?
These reporters want to stream this stuff live, or at least get it transmitted ASAP. I would personally choose the Sorenson compressor, but I know how much time it takes to compress even a short clip. Waiting an hour to compress a 3 minute video clip may not be practical.
I think that H.263 may have been chosen as a more suitable compromise between compression speed and video quality for reporters in the field who may sometimes be running for their lives.
1. Coders get dumped when they turn 35.
2. We can't find any skilled coders.
Hello? Does anyone else see the correlation here? Skill is the product of talent and experience. Talent comes from God Almighty in precious little doses, but experience comes with age.
As one of my friends in the aerospace industry puts it, "Any manger will tell you that of course there's a shortage of engineers - especially really brilliant ones who will work 60 hours a week for $20K a year."
The skilled coder you can't find is probably one of the ones you dumped because his salary was just a little too high. Now you'll pay double his salary in recruitment costs and receive nothing productive in return.
You would think even an MBA could understand this.
You give MBA's too much credit. Several years ago I worked at a company that went through this exact situation. Older, experienced engineers were forced out the door to reduce department budgets by MBAs who had no understanding of the technical contributions of those guys. Many of them were ultimately rehired as consultants at 3X their hourly rates. But the MBA's didn't care, because the consulting fees were paid from a different pot of money, not their departmental budgets. They could cut their salary budgets, still get the work done, and look good to their own bosses, even though it cost the company far more money in the long run.
I studied at a well-known eastern U. What the heck, let's name names, Georgia Tech. It's a corporate research empire. Most profs are spending their time writing proposals and trolling for grants from industry and gov't. The profs are paid on a percentage of the grants they pull in. After a coupla years, that is their only income. And the undergrads are always grumbling about not getting good prof time. For the profs that are good at it (i.e., the entrepeneural types), it's a nice cash cow.
I graduated from Georgia Tech and have worked at two other univerities in faculty positions. I want to point out to everyone that research (and money) are the defining factors in academia nowadays. Among the things that I've observed:
(1) At most schools, excellence in teaching is a distant third in your evaluation by your peers, after total research funding and number of publications. You are encouraged not to devote too much of your time and effort to teaching, particularly if you don't have tenure. Mediocrity is sufficient. As we say in academia, "Good teaching will not help you, but bad teaching can hurt you."
(2) Academia is pretty strongly divided into "have's" and "have-not's". Professors with lots of research money can buy out of their teaching duties, can pay themselves summer salary, and can support lots of graduate students who write papers. Professors without money can do none of these things. They are punished by the administration with the most undesirable service and teaching duties, and are often scorned by their peers who bring in research money. (Unfortunately, most "have's" eventually turn into "have-not's" as their research specializations become mainstream technologies and cease to be funded by the government.) That's why you see so many bitter, burned-out older faculty at most schools.
(3) The amount of money that companies funnel into universities pales in comparison to the money that the government spends. Most schools have been sucking at the government teat for so long that their undergraduate and graduate curriculums have been reshaped to train students to become potential graduate research assistants, not private sector employees. The U.S. government has literally transformed engineering and computer science curriculums in the past 20 years without fully realizing it.
(4) What few companies understand that most universities don't worry if students are employable or learn subjects relevant to the workplace. We faculty have to worry about our own government-funded research programs, and if you don't give us money we can't waste time on you. Once the government decides not to fund a particular research area, it disappears from academia as faculty bail out for industry jobs or retire.
(5) The fact that UC Berkeley sold so much of its research background rights for so little money is a perfect example of the incompetence of most university technology transfer offices. On one hand, the OTT wants to own or control all of the IP produced by the faculty, just like a private corporation. On the other hand, too often they wind up either ignoring the IP, ignoring the companies that want to license it, or giving away too much for too little money. Many years ago, faculty were free to create their own companies and profit from their inventions (e.g. Silicon Valley). This was A Good Thing for our economy. Nowadays most schools are so paranoid about another Netscape or Microsoft "escaping", that they've effectively made it impossible for faculty to be entrepreneurs.
(6) IMHO, more corporate money in universities is a good thing, at least in engineering and computer science. It encourages faculty to teach courses and do research that is relevant to industry. Without it, entire teaching disciplines are fading away at most schools, solely for lack of government funding. For example, if you're a EE, try finding a senior elective in transistor-level circuit design at your school. If you're lucky, you'll find an older faculty member (50+) who may still teach it. Companies are dying to hire circuit designers even now, but most schools have abandoned microelectronics and circuit design as unfundable specializations.
Students at schools like Georgia Tech may complain about the lack of attention they receive from faculty, but in fact they are really better off than they realize. A top ten school can at least hire lots of faculty and offer a wider range of courses and facilities. The nightmare situations occur at smaller schools that try to act like bigger research schools, but wind up doing poorly at both research and teaching. You can complain about professors being research and money hounds, but in fact that is exactly what we are hired to do nowadays.
As a seasoned systems administrator in a college department and former student myself, I know that in a college environment, the efforts to which some students will go to cheat show an astonishing amount of creativity---breaking into accounts, exploiting lack of permission control on other users' accounts, searching through the recycle bins, etc. The use of technology in this environment has made cheating easier, and harder to trace.
Yes, but that's only true if the cheater actually goes to the trouble to thoroughly cover his tracks. A cheater, by definition, is unwilling to do hard work. He might bother to erase a log file or rename a few variables, but the fact that his assignment is a carbon copy of someone else's work sticks out like a sore thumb.
An example from my previous place of employment: I gave an circuit simulation design project to a group of graduate students. Five of the assignments were practically identical. The students had renamed the components in each of their respective assignments, but had forgotten to change the time stamps and license code at the bottom of each plot, which showed that assignment had been run and printed on the same computer about 3 minutes apart. Furthermore, the schematics, when superimposed on each other, were identical (same wires, nodes, component orientations, etc.)
The effort it takes to cover your tracks, and cover them well, is simply more than most cheaters are willing to go to, else they would have done their own work in the first place. Believe me when I say that professors know what to look for, and we can usually find blatant evidence when cheating occurs, if we take the time to look. The only reason that most students get away with the cheating that they do is simply lack of time on the part of the professor to perform the pattern matching. Automating the searching/matching process now makes it feasible for a professor to effectively police a large class.
The risk is that some of the students are probably innocent, merely being guilty of having their own papers copied without their knowledge. Indeed, we've seen many cases here where the person whose work was copied ends up in a situation where they have to prove their own innocence.
In the rare cases I've seen where a student's assignment was copied by a stranger without his/her knowledge, the direction of the copying was obvious. The overwhelming majority of cheating is blatant plagiarism from external sources or else collaborative in nature. One student will let one or more friends copy his work, and again this is easy for a professor to spot.
Unfortunately, the technology of online composition and submission of papers (as typically done at most Universities) lacks sufficient security, encryption, and authentication standards.
However, no amount of security or encryption is going to stop collaborative cheating or outright plagiarism.
I just fear that the cost of this action could possibly end the academic careers of too many students guilty of nothing more than failing to see how their work could be copied.
Believe me, the student will get the benefit of the doubt unless the evidence says otherwise. Universities are very sensitive to lawsuits by outraged students and parents. No professor is going to turn in a student for an honor code violation unless he is sure he has undeniable proof. Any professor can tell you about cases where he let a cheater walk because he didn't have evidence that would stand up in a court of law. Faculty members can be sued for false accusations, and we don't make them lightly.
This is a great idea! I would love to see a computing device in the hands of every person on the planet. But first I would like to see full stomaches on all of them. How are they going to market these to people who can't afford to feed their children? [Free bag of rice with every purchase.] I am not saying these are a bad idea. Or that they cannot work. Hell, I would love to see them all over in countries that could afford them. But I believe that we should at least try and be socially conscience of the thousands of people who not only have never heard of a computer -- they are dying because of starvation.
Okay, I realize that I am endangering my karma here. I am taking a stance that may be seen as flamebait. But I really believe that this should be said by someone. But anyway, what are our plans to bring food to people who need it? Those should be more important than computers.
This is a specious argument. Everyone on Slashdot could sell their computers right now and send the money to feed the starving children. Guess what? Six months from now they'll -still- be starving.
If civilization waited until everyone was fed and happy before investing resources in new ideas, we'd still be squatting in caves. It's the investment in those ideas that makes real advancements in the quality of life possible.
Unfortunately, world hunger is a much harder problem to solve than building a Simputer. Making a Simputer is just a matter of engineering - solving world hunger is more an economic, cultural, and political problem than a matter of growing more food. However, building a Simputer might help some of the best and brightest in Third World countries help themselves, and in the long run that will be the only viable solution to mass starvation.
A final point for everyone to consider - are the creators of the Simputer overestimating the market for these machines? Remember the guy who created the hand cranked radio a few years ago? He designed it to bring modern communications to remote Third World villages. The problem was that no one in the Third World wanted to buy them (or could afford to do so)! Nowadays they sell them as camping gear in the U.S.A. Somehow I think the Simputer may have a bigger market in the First World rather than the Third World.
They are talking about achieving a 25 micron feature size. The current generation of processors is being done with an 0.13 micron feature size, meaning that the number of gates you can fit on your plastic chip is about 40000 (200 times 200) times lower.
Still, if they can get one transistor in 25 microns square, and handle all the wiring in other layers:
* The original 68000 (with 68,000 transistors) fits in a 6.5 by 6.5 mm square.
* The original 80386 (with 275,000 transistors) fits in a 13.1 by 13.1 mm square.
* The original 80486 (1.2M transistors) needs 27.4 by 27.4 mm (just over a square inch). Once we get to this stage a lot of the transistors are L1 cache.
Unfortunately, the definition of "feature size" is not the minimum width or length of a transistor, but instead the smallest dimension that can be reliably fabricated on a chip. On digital ICs, the feature size refers to the minimum gate length. The actual transistor, when factoring in total gate and source/drain diffusion areas, is much larger than the square of the feature size.
To estimate how much area it would take to port a microprocessor to this printing process, the best approximation is to scale the area of the original chip by the ratio of (25 um / (x) um)^2, where (x) is the original feature size. The original 80486 die was 0.414 inches by 0.649 inches with a feature size of 0.8 um. The "printed" 80486 would therefore be 12.94 inches by 20.28 inches! Furthermore, it would be much slower than the "old" 80486 because of the lower mobility and much higher capacitance of the transistors. You would be lucky to clock it at 1 MHz.
Although this printing process would be great for low power embedded applications (e.g. "smart" wallpaper, giant displays, ultra-cheap dumb terminals, throwaway sensors, etc.), you'll never get the kind of high-performance computing you'd expect from a laptop or desktop, or even a Palm PDA for that matter. Performance-wise, this process will never compete with traditional silicon and submicron lithography, although it will find some useful and profitable applications.
I didn't mean to imply that this process would be useless. Certainly there would be lots of applications where they would be extremely useful and novel. The innovative display technologies you mentioned are one. There are lots of embedded, low-computation, low-power applications where a manufacturing technique like this could find a big market. (I was thinking about ultra-cheap stick-on sensors powered by sunlight, communicating via picocellular networking.)
The point I wanted to make was that the high-end laptop or desktop computer with submicron CMOS ICs is not going to be replaced anytime soon. Even if you could replace a 1 GHz CPU with 1000 CPUs operating at 1 MHz, your power requirements would not decrease (direct tradeoff of clock speed vs. number of transistors), and you'd need to write an operating system that could take advantage of a slow, massively-parallel processor. Also, as a previous poster pointed out, you wouldn't even get the 1 MHz clock speed, because the carrier mobility of the polymer semiconductors is much lower than silicon.
This manufacturing technique will find a niche, but we won't be stamping out general-utility PCs with it anytime soon.
Nowhere in Cringley's article is there any discussion of the performance penalty that this process would entail. Let's assume we want to duplicate the equivalent of 50 million transistors clocked at 1 GHz. Right now Intel can squeeze that many components on a 200 square millimeter die by using a CMOS process with a 0.18 micron feature size.
Now assume that your printing process needs transistors with 10 micron feature sizes to ensure proper registration and a high enough yield to be manufacturable. That increases your effective "die" to 956 square inches. (Area increases with the square of feature size.) That's equivalent to 10 sheets of single-sided paper.
For a multi-layer printing process, 10 layers of plastic sandwiched together would definitely be possible. HOWEVER - you are not going to be able to clock your circuit at 1 GHz! Because of the much larger size (and capacitance) of your circuit, you'll do well to get a 1 MHz clock speed (1000X slower).
While this process may be very useful for e-books, displays, etc., I don't see how any high-performance computing could be done with a microprocessor constructed with this technique. Your only alternative to slower clock speeds would be massive parallelism to achieve higher computational throughput. Assuming a direct tradeoff of speed versus number of transistors, you would need 10000 layers instead of 10 layers in your process. There goes your low manufacturing cost.
It's not just enough for a computer to be cheap. It's got to be fast, or it's no good to anyone.
As long as our educational institutions feel the need to play along with corporate sponsors, these situations will continue to occur. And with all the big money grants and donations available, more and more schools will feel the urge to get some of that money for themselves. Unfortunately, this can only hurt the students and faculty, ultimately. The damage has been going on for 30+ years already. The entire university landscape has been dramatically altered by the growth of federal research funding, and any large-scale corporate funding is not going to make things that much worse. As a faculty member in a research university, I can assure everyone that no university is going to think twice about getting money from wherever it can.
My employer, like every other public university, never gets enough money from the state legislature. The only ways to deal with growing operating costs is (a) increased tuition or (b) grants and gifts. Rather than pass costs for new equipment on to the students, or beg for it from the legislature and Board of Regents, a school would much rather take a corporate donation.
Faculty members and university administrators fully realize the potential conflicts, but taking money or equipment from a company like Intel is better than the alternative - no modern lab equipment or computers for the students. If you are unhappy about this situation, consider the alternative: Would you be willing to pay 2X tuition, or a 2% higher state sales tax, or a 5% higher state income tax, etc., to provide your state's public univerities with enough money? If not, then it's moot to complain about corporate influence in higher education.
> Who would care about CS101 notes, since they're pretty much the same world wide?
The professor who teaches the course may care very much, if he/she feels that new technology may -replace- him/her as a teacher. There is a deep-rooted fear among a -lot- of faculty members that universities will eventually use Internet-based instruction to reduce or eliminate professors altogether. The idea runs like this:
(a) The university hires an adjunct professor to teach a class and generate Web-based lectures.
(b) The adjunct signs over his/her IP rights to the course material as a condition of employment.
(c) The university dismisses the professor, and offers the course every semester thereafter using a TA to hold office hours and give exams.
I am a faculty member myself, and I have heard these comments from my colleagues repeatedly. Nor do I doubt some college administrators would do it to cut costs, if they thought they could get away with it. Students have always had the ability to photocopy course notes and give them to other students, but a photocopier can't teach a class, so IP concerns were never a big issue in the past. However, the perception nowadays is that Internet-based multimedia lectures -can- replace a live professor, at least for some courses. The entire issue of IP ownership of course notes is simply a pre-emptive move on the part of faculty members to prevent this scenario from taking place.
Not that it will help, in the long run. If Moore's Law stays on track, I fully expect to see AI-based tutors taking the place of live instructors for many undergraduate courses in the next 20 to 30 years. Universities will embrace the innovation as a cost-cutting measure but will themselves be swept away by the revolution in teaching. Professors and universities alike are unwilling and unable to grasp how much the business of education is going to change in the next two decades.
Slashdot Mirror
What do you mean? It should be -easy- to convince people your processor works the way you claim. Loan them some of your prototypes (after they sign a non-disclosure agreement) and let them test the processors on their own, with their own equipment.
If your measurements are valid, they will be able to replicate them easily. If not, you will learn what you did wrong quickly enough. Whatever you do, don't sit around blaming close-minded attitudes for your failure to convince people. Otherwise, you start to sound like crackpot researchers.
As to Dr. Schon, I suspect that he will be asked to replicate some of his measurements in the presence of trained observers. He should still have working parts, and if he can do so, he will be vindicated. Otherwise, his career is finished.
As someone who has participated in CAD and microelectronics research in academia, I can offer a few informed opinions about this situation. Everyone has a opinion of what academia "should" be doing. Very few people take a hard look at -why- academia research in these areas has come to a standstill.
Research money for CAD development in particular and microelectronics in general is almost impossible to come by. For research in IC processing, the big factor is the cost of capital equipment. Universities simply can't keep up - only industry can afford the multi-multi-million dollar capital equipment costs. Even research in analog and mixed-signal circuit design has pretty much collapsed, simply for lack of funding to fabricate ICs through third-party fabs. What little money is available is quickly soaked up by a tiny number of high-profile researchers, mainly in California schools. Unfortunately, these guys can't even come close to meeting the demand for students and research output.
CAD is a somewhat different story. The capital cost of CAD research is actually pretty small. All you need are a few good workstations. The bigger problem is lack of qualified graduate students, and lack of any money to pay them. Everybody agrees that universities should do more CAD research, but the federal government won't write the checks (applied research should be funded by industry) and industry won't write the checks (too many intellectual property hassles with universities, plus no guaranteed return on investment). In fact, industry often acts as its own worst enemy by hiring away students and faculty both and eating the academic "seed corn" of future research.
To top it off, any CAD researcher with even a moderately innovative tool will do MUCH better by leaving academia, finding venture capital, starting his/her own company, getting bought out by Cadence, and retiring wealthy.
Graduate student recruitment has improved somewhat in the past year due to the recession, but as soon as the microelectronics industry starts to recover, you won't be able to hire a qualified graduate student or faculty member for love or money. It will take massive amounts of cash to change this trend, and frankly I don't see it happening here in the U.S. The microelectronics industry will have to rely on internal R & D from now on. Except for a few remaining high-profile programs, the academic sector is pretty much finished.
As an ECE faculty member, I can offer some observations of my own about the topics of faculty hiring and retention.
To begin with, faculty hiring in electrical and computer engineering is -damned- difficult right now, and won't be getting better anytime soon. Ten years ago you could advertise a faculty position and get hundreds of resumes, with at least half a dozen outstanding candidates in the pile. Nowadays you're lucky to get a dozen resumes, and typically most of them are a complete joke. Most ECE departments have had open positions for -years-, with little prospect for filling them.
The tech sector boom of the 90's, coupled with stagnant faculty salaries and declining budgets, has been the direct cause. Many CS and CE departments suffered enormous attrition during the dot-com boom. Unfortunately, the current recession did almost nothing to improve the situation, since (a) the really bright, good people have remained employed rather than return to university positions, and (b) the CS and CE departments are no longer cranking out Ph.D. candidates in significant numbers. Graduate enrollment by U.S. citizens in Ph.D. programs has simply collapsed. You can barely get domestic students to hang around for an M.S. nowdays, much less a doctorate. Industry money is just too good, even now.
So how do universities fill in the gaps? Adjunct positions are one way, but take my word for it - an adjunct professor is the lowest of the low at most schools. Rotten pay, long hours, and the best you can hope for is that you're left alone by the tenure-track faculty.
However (and this is the part many people do not realize), adjunct and part-time teaching help is politically and economically undesirable at most schools. Parents do not want to hear that their kids are being taught by part-time hired help. Furthermore, schools want full-time faculty who can pursue research grants and bring in the big government bucks. Adjunt professors are a necessary evil, but one we'd rather not have to deal with.
So what's the future of CS, EE, and CE faculties? IMHO, I think faculty hiring is going to go the way of graduate student enrollments over the next ten years. Nowadays the majority of graduate students in our department are international students. This is the norm even at the top universities in the U.S. My guess is that eventually U.S. schools will have to recruit most tenure-track faculty overseas, and sponsor their visas just as companies do. Granted, this is politically even more undesirable than hiring adjuncts, but you're not going to find qualified full-time professors who are able to pursue research funds otherwise.
After reading the article, I checked out Rainmaker's site. These guys have a theory, some patents, and some simulations. What they don't seem to have is any working hardware that proves this 10X bandwidth increase can actually be achieved in residential cable systems.
Does this remind anyone of Transmeta, who promised processors with a fraction of the power consumption at higher speeds? Everybody loved them when all they had was a press release. The actual product didn't work as advertised, and now they've faded away.
If it sounds too good to be true, it probably is. 10X uber-bandwidth schemes sound suspiciously like 10X uber-compression schemes. I'll reserve my enthusiasm when I see working hardware.
Although many people believe that life is very common throughout the universe, and will spread rapidly anywhere it gets a foothold, there's no deterministic path towards intelligent life. The prevailing view is that intelligence is an accident, not an end-product, of evolution. Many scientists believe that intelligent life is extremely rare, and that we may be the first intelligent civilization to evolve in the Milky Way galaxy.
After all, if a large meteor hadn't hit the earth about 65 millions years ago, dinosaurs would still be roaming the earth. Humanity is here due to a freak accident, and in fact we may be very unique.
MPEG-4 may be a better algorithm, but is it a suitable algorithm for real-time or short-turnaround compression?
These reporters want to stream this stuff live, or at least get it transmitted ASAP. I would personally choose the Sorenson compressor, but I know how much time it takes to compress even a short clip. Waiting an hour to compress a 3 minute video clip may not be practical.
I think that H.263 may have been chosen as a more suitable compromise between compression speed and video quality for reporters in the field who may sometimes be running for their lives.
As one of my friends in the aerospace industry puts it, "Any manger will tell you that of course there's a shortage of engineers - especially really brilliant ones who will work 60 hours a week for $20K a year."
You give MBA's too much credit. Several years ago I worked at a company that went through this exact situation. Older, experienced engineers were forced out the door to reduce department budgets by MBAs who had no understanding of the technical contributions of those guys. Many of them were ultimately rehired as consultants at 3X their hourly rates. But the MBA's didn't care, because the consulting fees were paid from a different pot of money, not their departmental budgets. They could cut their salary budgets, still get the work done, and look good to their own bosses, even though it cost the company far more money in the long run.
I graduated from Georgia Tech and have worked at two other univerities in faculty positions. I want to point out to everyone that research (and money) are the defining factors in academia nowadays. Among the things that I've observed:
(1) At most schools, excellence in teaching is a distant third in your evaluation by your peers, after total research funding and number of publications. You are encouraged not to devote too much of your time and effort to teaching, particularly if you don't have tenure. Mediocrity is sufficient. As we say in academia, "Good teaching will not help you, but bad teaching can hurt you."
(2) Academia is pretty strongly divided into "have's" and "have-not's". Professors with lots of research money can buy out of their teaching duties, can pay themselves summer salary, and can support lots of graduate students who write papers. Professors without money can do none of these things. They are punished by the administration with the most undesirable service and teaching duties, and are often scorned by their peers who bring in research money. (Unfortunately, most "have's" eventually turn into "have-not's" as their research specializations become mainstream technologies and cease to be funded by the government.) That's why you see so many bitter, burned-out older faculty at most schools.
(3) The amount of money that companies funnel into universities pales in comparison to the money that the government spends. Most schools have been sucking at the government teat for so long that their undergraduate and graduate curriculums have been reshaped to train students to become potential graduate research assistants, not private sector employees. The U.S. government has literally transformed engineering and computer science curriculums in the past 20 years without fully realizing it.
(4) What few companies understand that most universities don't worry if students are employable or learn subjects relevant to the workplace. We faculty have to worry about our own government-funded research programs, and if you don't give us money we can't waste time on you. Once the government decides not to fund a particular research area, it disappears from academia as faculty bail out for industry jobs or retire.
(5) The fact that UC Berkeley sold so much of its research background rights for so little money is a perfect example of the incompetence of most university technology transfer offices. On one hand, the OTT wants to own or control all of the IP produced by the faculty, just like a private corporation. On the other hand, too often they wind up either ignoring the IP, ignoring the companies that want to license it, or giving away too much for too little money. Many years ago, faculty were free to create their own companies and profit from their inventions (e.g. Silicon Valley). This was A Good Thing for our economy. Nowadays most schools are so paranoid about another Netscape or Microsoft "escaping", that they've effectively made it impossible for faculty to be entrepreneurs.
(6) IMHO, more corporate money in universities is a good thing, at least in engineering and computer science. It encourages faculty to teach courses and do research that is relevant to industry. Without it, entire teaching disciplines are fading away at most schools, solely for lack of government funding. For example, if you're a EE, try finding a senior elective in transistor-level circuit design at your school. If you're lucky, you'll find an older faculty member (50+) who may still teach it. Companies are dying to hire circuit designers even now, but most schools have abandoned microelectronics and circuit design as unfundable specializations.
Students at schools like Georgia Tech may complain about the lack of attention they receive from faculty, but in fact they are really better off than they realize. A top ten school can at least hire lots of faculty and offer a wider range of courses and facilities. The nightmare situations occur at smaller schools that try to act like bigger research schools, but wind up doing poorly at both research and teaching. You can complain about professors being research and money hounds, but in fact that is exactly what we are hired to do nowadays.
As a seasoned systems administrator in a college department and former student myself, I know that in a college environment, the efforts to which some students will go to cheat show an astonishing amount of creativity---breaking into accounts, exploiting lack of permission control on other users' accounts, searching through the recycle bins, etc. The use of technology in this environment has made cheating easier, and harder to trace.
Yes, but that's only true if the cheater actually goes to the trouble to thoroughly cover his tracks. A cheater, by definition, is unwilling to do hard work. He might bother to erase a log file or rename a few variables, but the fact that his assignment is a carbon copy of someone else's work sticks out like a sore thumb.
An example from my previous place of employment: I gave an circuit simulation design project to a group of graduate students. Five of the assignments were practically identical. The students had renamed the components in each of their respective assignments, but had forgotten to change the time stamps and license code at the bottom of each plot, which showed that assignment had been run and printed on the same computer about 3 minutes apart. Furthermore, the schematics, when superimposed on each other, were identical (same wires, nodes, component orientations, etc.)
The effort it takes to cover your tracks, and cover them well, is simply more than most cheaters are willing to go to, else they would have done their own work in the first place. Believe me when I say that professors know what to look for, and we can usually find blatant evidence when cheating occurs, if we take the time to look. The only reason that most students get away with the cheating that they do is simply lack of time on the part of the professor to perform the pattern matching. Automating the searching/matching process now makes it feasible for a professor to effectively police a large class.
The risk is that some of the students are probably innocent, merely being guilty of having their own papers copied without their knowledge. Indeed, we've seen many cases here where the person whose work was copied ends up in a situation where they have to prove their own innocence.
In the rare cases I've seen where a student's assignment was copied by a stranger without his/her knowledge, the direction of the copying was obvious. The overwhelming majority of cheating is blatant plagiarism from external sources or else collaborative in nature. One student will let one or more friends copy his work, and again this is easy for a professor to spot.
Unfortunately, the technology of online composition and submission of papers (as typically done at most Universities) lacks sufficient security, encryption, and authentication standards.
However, no amount of security or encryption is going to stop collaborative cheating or outright plagiarism.
I just fear that the cost of this action could possibly end the academic careers of too many students guilty of nothing more than failing to see how their work could be copied.
Believe me, the student will get the benefit of the doubt unless the evidence says otherwise. Universities are very sensitive to lawsuits by outraged students and parents. No professor is going to turn in a student for an honor code violation unless he is sure he has undeniable proof. Any professor can tell you about cases where he let a cheater walk because he didn't have evidence that would stand up in a court of law. Faculty members can be sued for false accusations, and we don't make them lightly.
This is a specious argument. Everyone on Slashdot could sell their computers right now and send the money to feed the starving children. Guess what? Six months from now they'll -still- be starving.
If civilization waited until everyone was fed and happy before investing resources in new ideas, we'd still be squatting in caves. It's the investment in those ideas that makes real advancements in the quality of life possible.
Unfortunately, world hunger is a much harder problem to solve than building a Simputer. Making a Simputer is just a matter of engineering - solving world hunger is more an economic, cultural, and political problem than a matter of growing more food. However, building a Simputer might help some of the best and brightest in Third World countries help themselves, and in the long run that will be the only viable solution to mass starvation.
A final point for everyone to consider - are the creators of the Simputer overestimating the market for these machines? Remember the guy who created the hand cranked radio a few years ago? He designed it to bring modern communications to remote Third World villages. The problem was that no one in the Third World wanted to buy them (or could afford to do so)! Nowadays they sell them as camping gear in the U.S.A. Somehow I think the Simputer may have a bigger market in the First World rather than the Third World.
They are talking about achieving a 25 micron feature size. The current generation of processors is being done with an 0.13 micron feature size, meaning that the number of gates you can fit on your plastic chip is about 40000 (200 times 200) times lower.
Still, if they can get one transistor in 25 microns square, and handle all the wiring in other layers:
* The original 68000 (with 68,000 transistors) fits in a 6.5 by 6.5 mm square.
* The original 80386 (with 275,000 transistors) fits in a 13.1 by 13.1 mm square.
* The original 80486 (1.2M transistors) needs 27.4 by 27.4 mm (just over a square inch). Once we get to this stage a lot of the transistors are L1 cache.
Unfortunately, the definition of "feature size" is not the minimum width or length of a transistor, but instead the smallest dimension that can be reliably fabricated on a chip. On digital ICs, the feature size refers to the minimum gate length. The actual transistor, when factoring in total gate and source/drain diffusion areas, is much larger than the square of the feature size.
To estimate how much area it would take to port a microprocessor to this printing process, the best approximation is to scale the area of the original chip by the ratio of (25 um / (x) um)^2, where (x) is the original feature size. The original 80486 die was 0.414 inches by 0.649 inches with a feature size of 0.8 um. The "printed" 80486 would therefore be 12.94 inches by 20.28 inches! Furthermore, it would be much slower than the "old" 80486 because of the lower mobility and much higher capacitance of the transistors. You would be lucky to clock it at 1 MHz.
Although this printing process would be great for low power embedded applications (e.g. "smart" wallpaper, giant displays, ultra-cheap dumb terminals, throwaway sensors, etc.), you'll never get the kind of high-performance computing you'd expect from a laptop or desktop, or even a Palm PDA for that matter. Performance-wise, this process will never compete with traditional silicon and submicron lithography, although it will find some useful and profitable applications.
I didn't mean to imply that this process would be useless. Certainly there would be lots of applications where they would be extremely useful and novel. The innovative display technologies you mentioned are one. There are lots of embedded, low-computation, low-power applications where a manufacturing technique like this could find a big market. (I was thinking about ultra-cheap stick-on sensors powered by sunlight, communicating via picocellular networking.)
The point I wanted to make was that the high-end laptop or desktop computer with submicron CMOS ICs is not going to be replaced anytime soon. Even if you could replace a 1 GHz CPU with 1000 CPUs operating at 1 MHz, your power requirements would not decrease (direct tradeoff of clock speed vs. number of transistors), and you'd need to write an operating system that could take advantage of a slow, massively-parallel processor. Also, as a previous poster pointed out, you wouldn't even get the 1 MHz clock speed, because the carrier mobility of the polymer semiconductors is much lower than silicon.
This manufacturing technique will find a niche, but we won't be stamping out general-utility PCs with it anytime soon.
Nowhere in Cringley's article is there any discussion of the performance penalty that this process would entail. Let's assume we want to duplicate the equivalent of 50 million transistors clocked at 1 GHz. Right now Intel can squeeze that many components on a 200 square millimeter die by using a CMOS process with a 0.18 micron feature size.
Now assume that your printing process needs transistors with 10 micron feature sizes to ensure proper registration and a high enough yield to be manufacturable. That increases your effective "die" to 956 square inches. (Area increases with the square of feature size.) That's equivalent to 10 sheets of single-sided paper.
For a multi-layer printing process, 10 layers of plastic sandwiched together would definitely be possible. HOWEVER - you are not going to be able to clock your circuit at 1 GHz! Because of the much larger size (and capacitance) of your circuit, you'll do well to get a 1 MHz clock speed (1000X slower).
While this process may be very useful for e-books, displays, etc., I don't see how any high-performance computing could be done with a microprocessor constructed with this technique. Your only alternative to slower clock speeds would be massive parallelism to achieve higher computational throughput. Assuming a direct tradeoff of speed versus number of transistors, you would need 10000 layers instead of 10 layers in your process. There goes your low manufacturing cost.
It's not just enough for a computer to be cheap. It's got to be fast, or it's no good to anyone.
As long as our educational institutions feel the need to play along with corporate sponsors, these situations will continue to occur. And with all the big money grants and donations available, more and more schools will feel the urge to get some of that money for themselves. Unfortunately, this can only hurt the students and faculty, ultimately.
The damage has been going on for 30+ years already. The entire university landscape has been dramatically altered by the growth of federal research funding, and any large-scale corporate funding is not going to make things that much worse. As a faculty member in a research university, I can assure everyone that no university is going to think twice about getting money from wherever it can.
My employer, like every other public university, never gets enough money from the state legislature. The only ways to deal with growing operating costs is (a) increased tuition or (b) grants and gifts. Rather than pass costs for new equipment on to the students, or beg for it from the legislature and Board of Regents, a school would much rather take a corporate donation.
Faculty members and university administrators fully realize the potential conflicts, but taking money or equipment from a company like Intel is better than the alternative - no modern lab equipment or computers for the students. If you are unhappy about this situation, consider the alternative: Would you be willing to pay 2X tuition, or a 2% higher state sales tax, or a 5% higher state income tax, etc., to provide your state's public univerities with enough money? If not, then it's moot to complain about corporate influence in higher education.
> Who would care about CS101 notes, since they're pretty much the same world wide?
The professor who teaches the course may care very much, if he/she feels that new technology may -replace- him/her as a teacher. There is a deep-rooted fear among a -lot- of faculty members that universities will eventually use Internet-based instruction to reduce or eliminate professors altogether. The idea runs like this:
(a) The university hires an adjunct professor to teach a class and generate Web-based lectures.
(b) The adjunct signs over his/her IP rights to the course material as a condition of employment.
(c) The university dismisses the professor, and offers the course every semester thereafter using a TA to hold office hours and give exams.
I am a faculty member myself, and I have heard these comments from my colleagues repeatedly. Nor do I doubt some college administrators would do it to cut costs, if they thought they could get away with it. Students have always had the ability to photocopy course notes and give them to other students, but a photocopier can't teach a class, so IP concerns were never a big issue in the past. However, the perception nowadays is that Internet-based multimedia lectures -can- replace a live professor, at least for some courses. The entire issue of IP ownership of course notes is simply a pre-emptive move on the part of faculty members to prevent this scenario from taking place.
Not that it will help, in the long run. If Moore's Law stays on track, I fully expect to see AI-based tutors taking the place of live instructors for many undergraduate courses in the next 20 to 30 years. Universities will embrace the innovation as a cost-cutting measure but will themselves be swept away by the revolution in teaching. Professors and universities alike are unwilling and unable to grasp how much the business of education is going to change in the next two decades.