the rate that a country prints money is only one factor in the inflation rate (not to mention that most money today isn't really printed). In fact, a country that has a constant money supply would end up suffering from deflation if they did not increase their money supply over time. This is because everyone is becoming richer (and there are likely more people), but the same amount of money is in circulation. This means the money that people have buys more than it did before. In fact that is one of the biggest problems that economies based on the gold standard (or other similar standards) had to face. Deflation is far more devastating on an economy than inflation, as it discourages investment and savings.
Yeah, because it's not like other OS's haven't gained a significant marketshare in the past few years. I'm pretty sure OSX's marketshare has improved a couple fold in the past 5 years, so it looks like your argument is just full of it.
We all know the only real winners will be Apple, and Windows XP. Linux will likely carry along with it's.8% market share or whatever it has been at for the past 10 years or so. If these allegations are true, and hold in the final version (remember this is BETA software), it could be problematic. However, there could also be good security reasons for these changes (allowing applications to register what dlls they use and not running if they're changed is a good security practice that can prevent third party applications from breaking their software through the insertion of trojans and/or adware). The inability to fix some of the issues is also probably due to the beta nature of Windows 7.
As for the sound issue, do we really know that this is the OS doing it, and not the driver manufacturers not having this feature implemented in their driver yet? Lots of things could be at fault, and to call DRACONIAN DRM on it is a bit hasty.
Actually, this implies that the difference in power of JUST the CPU is far more than 50%. This is because most of the system's power draw is NOT from the CPU. Lets assume for instance that at idle the Atom processor consumes 5 watts of power (the rest of the system consumes the rest). This means that at idle the VIA cpu consumes almost 8 watts of power (both these estimates are perfectly reasonable based on the class of processors that they are). This estimates means that when idling the VIA cpu is consuming 50% more power than the atom.
Now, when the cpus are at load, the Atom processor is consuming ~8.5 watts of power, and the VIA is consuming ~25 or 26 watts of power. This looks to me that the via processor is consuming 4 times the power of the Atom, not merely 50% more.
Of course, this estimate is assuming that the Atom processor's idle power is only 5 watts. In reality, the idle power it consumes is likely even lower, as it was designed to minimize power dissipation. Now, claiming that the VIA's system power is approximately 50% more than the Atom is not accurate, but that doesn't mean that the CPU is not consuming that much more power. Anyone doing a fair comparison between the processors would likely be focusing on the difference in power of the CPUs themselves. Otherwise in a full system, the difference between a CPU that requires 50 Watts of power, and one that requires 100Watts of power wouldn't be that significant.
Actually, this is basically true anytime. If you complain to your cable company and threaten to cancel they will almost always move you to a cheaper rate (assuming you're not on the cheapest introductory rate). They make money off a customer paying $70/month (for cable and internet) and those who pay $100/month. Obviously, they prefer the ones who pay them more. However, they realize that they're better off with a customer paying $70/month than having no customer at all. You just have to be willing to say you want to cancel your account, and either pull the trigger when needed, or hang up before you do.
The real question that no one is asking is what time period is this 23000% sales jump over? IBM could claim a 230,000% sales increase day over day on occasions if they land a big client on one day or whatever. The same is true for most any company. While this sales jump is large, the real question is what is the real impact on DVD sales over months or years timespan. The day they're released on youtube, you'd expect a big jump like this. You have a large fan base who haven't seen the show in years, someone links them to this, or it's advertised on youtubes page, and they go "oh yeah, i've been wanting that" and buy it. However how does this effect sales after the initial excitement dies down. Hell, I'm sure on a normal day, a slashdot story about monty python would increase their DVD sales a couple fold over the previous days sales. It all depends on the initial volume of the previous day etc. Of course, as they're at the top of sales right now, this is a fairly big deal. But if they hold on a sizable increase in sales 2 months from now, then people will seriously consider this model.
of course... in any country owning land is only a technicality. I can't think of any country that doesn't have some form of eminent domain laws on the book. Technically they're supposed to compensate you with full market prices, but I know that doesn't always happen, and doesn't account for the personal significance often associated with a location. If a business is allowed to operate on a set piece of land that they paid for indefinitely, it's the same to them as owning it. While the government can seize it or force it to cease production (known to happen in china, and elsewhere), this obviously isn't the standard case. If it happened to every business that operated in China, no one would be spending the money to open new factories there. It's how the market works.
And by free market, there are many different aspects involved. These include overall tax rates, tariffs (if they exist), regulations, government incentives, etc.
China isn't remotely a communist country by the real meaning of the word. The only reason their economy has been booming in recent years is because they have opened up their markets (unlike a true communist country such as North Korea). Basically, you can have an authoritarian regime without communism (like china), although I've yet to see an instance of a communist regime that isn't authoritarian. At this point in time, I wouldn't be surprised if China's markets are actually freer than most of Europe's markets.
And if the majority of people think the internet should be censored, so be it. If the majority of people think blacks should be slaves, so be it. The whole rule of the majority thing is a very very dangerous way to talk about the government. A majority rule system is just as deadly as being ruled by a tyrant, it's just that now the tyrant happens to be 50.1% of the population who doesn't like you or something about you.
"Democracy is based on the assumption that a million men are wiser than one man. How's that again? I missed something. Autocracy is based on the assumption that one man is wiser than a million men. Let's play that over again too. Who decides?" --Lazarus Long
So under this new amendment to the constitution, a 15 year old hacker who breaks into someone else's computer and illegally accesses their personal information (or reads someone's instant messages sent on an unsecured wireless network, etc) can be tried for treason for committing a search without a warrant . . . You do know what the penalty for treason normally is, right?
Also, such a law would basically make it impossible for congress to operate at all, as laws are regularly shot down by the constitution. One goes just a little too far, or a judge redefines what one particular clause in the constitution means. Personally, I could see some good of it, any federal politician caught trying to bail out the auto industry, or redistribute income etc etc could be punished with treason for violating the constitution.
Basically such an amendment would (like most government solutions) cause more problems than it solves.
Even in the era of muscle cars, or the more recent SUV era, the gas tax modified behavior. It made gas more expensive, and made the cost of driving better equal the actual cost of driving. I am not talking about creating prohibitive gas taxes whose goal is to reduce consumption, and generate more money than it costs to build and maintain the road systems. I am talking about keeping gas taxes at a level to pay for all construction and maintenance needed for the roadways.
There is a big difference between the two, and just because it doesn't change peoples behavior drastically does not mean it does not influence people's behavior in the long run. If anything, before gas prices plummeted, the gas taxes actually influenced people behavior LESS than they did in the past. If gas taxes are say 25 cents/gallon, people will notice it and see it as a bigger deal when gas costs $1 normally ($1.25 after tax). The difference in behavior between when gas is $3.75 a gallon and $4.00/gallon is likely less.
Actually, with gas taxes, part of the reason for them is precisely to modify peoples behavior. In general, I agree that taxes shouldn't be used to modify peoples behavior. However, this is an instance where the tax is levied on something that allows you to use that states resources. The states are building billions of dollars worth of highways, and they don't want to encourage people to use these resources inefficiently. They want to modify people's behavior so that they aren't as likely to crowd the roads (encouraging people to consolidate trips, carpool, etc). The roads need to be funded somehow, and a usage tax is one of the most effective ways to handle it. In fact, I wish all the money paying for roads only came from gas taxes. That would end up with a more even and natural balance where supply meets demand.
Taxing gas to pay for building roads is done precisely to help balance out the supply and demand for using roads. Roads entail a large fixed cost to build, and a smaller marginal cost for wear and tear on the road. If the government didn't tax the usage of the roads we could end up with the tragedy of the commons. Taxing them fixes this so that the supply and demand meet more evenly where the costs of the road best offsets the benefits given by the road. And yes, this requires modifying peoples behavior. Otherwise, people would overuse the roads, and we'd end up needing extra roads that don't provide enough additional utility to the people to justify the cost (as they wouldn't have needed them if they had to pay to reflect the roads cost).
Without even doing any calculations of the amount of fuel needed to ship the CFL, it's obvious that this environmental cost is more than offset by the savings in power of using a CFL vs an incandescent. The simple way to figure this out is to look at the price, and the cost savings of the bulb over a couple months time. A new CFL only costs a couple dollars to buy. That means (unless companies are selling them at a loss) that it costs less than this amount to ship the bulb to you. Assuming for a minute that the cost of energy in fuel and coal powering a power plant are the same (with the same efficiencies etc, in reality the power grid powered by coal is generally cheaper per watt, but that's okay), if you can make up the cost of the bulb in a matter of months by lowering your electric bill, you have more than made up for the cost of shipping.
I always laugh when people start talking about carbon footprints and all that. Currently the simplest and only real way to drastically reduce your carbon footprint is to spend less money. Things that cost more in general have a greater carbon footprint (there are exceptions to this rule, but it often holds true). So the footprint of manufacturing a new stove is roughly proportional to how much the stove costs (obviously if you're buying a gold plated stove with platinum racks this won't quite hold true). So the quickest way to halve the US's carbon footprint is to cut the US's spending power in half. Of course, with the steps the government is taking to reverse the current economic crisis, they could easily accomplish this.
First, you sound bitter from a bad experience with one particular IEEE conference. Every IEEE conference is different, and the IEEE is just a society that sponsors some of the conferences in the field, and doesn't necessarily perform quality control on them. That is handled by the conference organizers themselves, and better conferences tend to have better quality control. The people in the field should know which are the better conferences.
Second, the fact that you reviewed a couple really awful papers has little bearing on whether your paper will be accepted or not. Now, if you could claim that these awful papers got accepted, while yours was rejected, you might have something to say. But if those papers were also rejected, your statements about them have absolutely no merit.
Also, sometimes you get bad luck with reviewers. You might get reviewers who think the entire premise behind your research is crap, or ones who don't like your group for whatever reason. Sometimes you get one who just happened to be having a bad day, and wanted to find fault with your paper. The hope is that with enough reviewers, it is unlikely that you will get only crappy reviewers. If that does happen, I'm sorry, try again. Also, don't put too much stock in the timestamps given. Those could have been modified for any number of reasons. It's highly unlikely that a reviewer would submit final reviews 2 minutes after getting the paper. Chances are good they'd save the paper, and wait a few weeks before looking at them right before the deadline. The only time I'd spend 2 minutes reviewing a paper is if it was either a) Completely unreadable (like the computer generated gibberish), or b) it was obviously plagiarizing other papers. In those cases, it isn't worth more than 2 minutes of my time to give it a valid review.
Hopefully next time you have better luck with reviewers, in my experience I tend to get some informed comments back on my papers that can help guide changes I want to make whether it gets accepted or not.
not everyone who uses a linux workstation is a sysadmin. If someone else has set up the box, and all they do is run a handful of applications on the machine, they don't know anywhere near enough to really run linux at home. If you're in a shop where EVERYONE knows how to admin their own machine, and runs windows, sure. This isn't the case in the real world.
Sure, you could try expanding the instruction set, and reach the point where a single instruction can do the work of 5-10 instructions. . . . Unfortunately doing this will likely reduce your operating clock frequency by a factor of 5-10, while increasing power and area by a similar ratio. Net gain, next to nothing. Plus of course, now instructions end up taking up 5-10 times as much storage space.
The problem is that all reasonable methods for improving the performance of single threaded code in convential ISAs requires changing the programming paradigm (backwards compatibility is still more important than power/performance). The first method people used was pipelining the processor core. This worked, and let chipmakers ramp up the clock frequency. However, it has limited gains. For one, there is a limit to how deep you can pipeline your work, because at the bare minimum your clock cycle has to be the speed of a latch, and one degree of logic (otherwise you're not doing anything in a single pipe stage and/or can't store the results). We reached the point where we couldn't feasibly increase the pipelining depth, and increasing it further pushed up the clock speed a little bit, but didn't actually improve performance (think Pentium 4). What else can be done ? Well, one idea that's been used for a while is executing multiple instructions at a time. We're doing this, modern processors can execute 3-4 instructions in a single clock cycle, of course they average around 1...
We could try going further with this idea, but limit studies have shown that their is no feasibly way to gain considerable performance out of current codes by looking at more instructions simultaneously. There is just a lack of parallelism available in single instruction streams. Secondly, memory and branch prediction greatly reduce the performance that a wide superscalar can achieve. Even a branch predictor that only misses one out of a thousand instructions is missing too frequently to obtain high performance on an 8-16 wide core.
The next logical steps are extensions to the ISA. This is being done, but still requires better compilers. SSE2 etc can greatly speedup vectorizable codes. Of course, the compilers fail here normally, and the only reason this gives a speedup is because it involves people specifically coding for a parallel architecture.
There are other methods being investigated to achiever higher performance, however all of them involve a new programming paradigm to some extent. Personally, I don't think functional programming is the answer. It might be part of an answer, and it might help with some stuff, but it's just an unnatural way to express statements. More explicitly parallel languages like verilog might become more common, however these are much harder to use than traditional programming languages.
What needs to change is the overall way we think about programming. We're taught to take problems and express them serially. Doing anything else is unnatural in most languages. Serial programming allows determinism in the code, and as we go forward, we need to reduce our need for determinism, or first split up an application into parts that can be performed deterministically, and parts that don't need to be. This will help with improving parallelism. Additionally, we need far better debug tools. Running multiple instances of GDB trying to trace threads just doesn't hack it. This will also require far greater unit testing in code development.
The good news however is that many applications don't need to be faster. They are fast enough as they stand, and only small sections need to be accelerated. The vast majority of code is for special case conditions that are rare anyhow, so we need ways to accelerate the cores of applications without worrying about the rest. As we go forward I think this will become an increasingly popular option. It won't be easy, and at a high level, applications need to be restructured to support greater parallelism. Whether we'll see parallelism a
Research is a fuzzy area. Particularly at academic institutions. The general rule is that any research you do while in the university is joint work with your advisor. It's just the way things are. If you don't like this sort of apprenticeship system, don't try to be a researcher. Getting a doctorate requires going to grad school. It's just the way things are.
It doesn't matter if your grant is for doing research on X and you discover Y in the process. The research you are doing is for your advisor. You ARE NOT doing it solely for yourself, and your funding status shows this. However, most professors at academic institutions have a lot of discretion concerning this stuff. Talk to them. Don't bother talking to lawyers or any of that stuff yet, talk to your advisor. Ask them how they plan on allowing you to contribute back to the community, or what you are allowed to take with you.
Most academic work is joint in nature, and oftentimes the code can later be released under some other license (BSD, GPL, limited rights, etc). Normally the code that you wrote isn't very useful outside of an academic setting, and would need to be overhauled to be useful elsewhere. If the idea is extremely good, and you know it is, you might want to talk with them about doing something outside of the university on it, or more likely obtaining a patent on it. Professors are generally allowed to consult a set number of hours per week, however there are always conflicts of interest that they need to be careful about. If the idea is not directly related to any grant they currently have though, they can likely work on it with you . . . That is if they're willing to. Also be warned that if you're not on good terms with your advisor to be careful, as this could backfire.
If you just want the code to be available to you later after you graduate for your students to use (if you plan on going into academia), you definitely want to talk with them about that. In general however you need to differentiate your work as a professor from your work as a graduate student, so often times the tools you'd want your students to use would be different anyhow (depending on how general they are). Many professors are quite willing to help you out, particularly if you're going into academia. Remember, it looks good for them if their former students later go on to become well-respected professors.
The likely case is that you're grossly over valuing your code. For most academic projects, the ideas generated, and the data obtained are far more valuable than the tools used to generate them.
as long as one person uses linux at work, and uses windows at home, this IS a fact. I don't need much proof to back it up. But in the lab I work in, I know enough people who regularly do work on Linux workstations, and have Windows on their personal machines (some do use Mac or Linux machines at home however). For the most part, I wouldn't want to use Linux on a laptop (the most common personal machine), while it's getting better, it tends to be a hassle to get wireless, multiple monitors, power management and such to "just work". I know it can be done on some machines with different hacks and settings, but most people don't want to deal with that. Personally, I don't want to deal with that, and thats why I use a Mac. However, the difference is that I'm a UNIX geek, and many of the people I know are not. They can use UNIX, and view it as a tool. Whatever gets the job done the best is what they tend to use.
actually, your statement is a great exaggeration of what the facts say. You are implying that most people don't want to use windows, but are forced to use it. This is NOT something you can claim from the statistics. It seems more likely that maybe 15% of home users use Macs, and 5-10% of business users use Macs. Therefore you have 5-10% of people who normally use Macs being "forced" to use Windows. There's a big difference between that and saying "people don't want windoze but are forced to use it at work.", where you make a generalization covering all people.
The same thing could be said about Linux actually. There are quite a few people who use Linux workstations at work, but have windows PCs at home (often because their home PC is a family PC). By your logic, I could say "people don't want to use Linux, but are forced to use it at work".
And I'll ask you a question, how far into a system does a virus have to penetrate to achieve the virus-writers objective? Does a virus need to have root access to be part of a Botnet? Does malware need root access to start issuing excessive popups to the user? Does malware need root permission to run when a user logs onto a machine? Didn't think so. The only reason many of these malware programs get root access is because they can, and doing so makes them harder to detect, and harder to remove at a later point. Considering the majority of viruses are out there to make money (popups, botnets, etc), you really don't need root access to do this. No reasonable amount of security at the OS level will remove the threat of locally run malware on the system. It just can't be done unless the system restricts what programs can and cannot be run.
You're right that artists can make labels money with a little marketing. However, you missed the biggest form of marketing that artists have. . . Playing live shows. For small unknown bands, that is how they get known. If you're opening for bands with similar styles (which you should be doing), and play good, the audience will appreciate it. If only 15 people at a show buy your CD afterward, you're doing pretty good with marketing (assuming the show has ~200 people there). These 15 people will likely tell others, and word of mouth spreads about the band. I've seen it many times where good bands can grow remarkably fast from this. Of course, this won't make them huge, but it means they can sell on the order of 20,000 records, which isn't too bad.
There are some record labels that take this into account when looking for new bands to sign on. I know New West records (one of my favorite record labels, where I'll randomly pick up albums by artists I don't know just because they're on the label) is really big on live music. One of the guys who runs the label is a huge music fan, going to tons of concerts a year. If he see's a band he likes that is doing there part, he'll often try to pick them up onto the label. The demos a band has are important to the label, but equally, if not more important is the bands live show. Additionally, the band (assuming they're not well established already) has to be willing to hold their end of the agreement, and tour regularly. If you're not willing to spend 100s of days a year on tour promoting your music, you're not worth the investment to them.
Your last statement is only partially true. The number of releases record labels make in a year has reduced, and they have started catering more to the very large stores. However, this was the case quite a bit before Napster came onto the scene, and the mp3 revolutionized the music industry. The large stores had started swallowing up the small. It is false to say that the small independent record stores are gone. There are still quite a few of them that flourish. The people who love music and are willing to pay for it often choose to go to the small independent stores, and always will. The store is far more about the personal interaction, and the recommendations that can be made there. What has hurt the small stores the most isn't the Walmart's and Best Buys who have small selections, low prices, and high volume, but the Amazon's of the world who have practically unlimited selection, and the benefits of scale that come with being so large that they can get lower prices. Many music fans have started shopping online for CDs they used to buy at the small independent stores.
The large record companies shot for the gold in the late 90s by focusing on the big hit of the day kind of thing. Under such a market, they create demand for music, and sell CDs (albeit to a limited number of artists). However, when you concentrate on the masses and the hit of the minute, you lose out on loyalty. The loyal fan base that goes to the small independent CD stores didn't want to switch to buying CDs at Walmart, and they didn't care about the flavor of the month. They bought lots of CDs by bands that aren't particularly profitable to the labels (but tended to bring in a steady stream of income). The group that they won over with low prices at Walmart, and mass consumed discs has little loyalty, and why should they. They could care less where they get their music from, and napster is as good a place as any, but the price was right. Besides, who cares if the back street boys didn't make a few extra bucks, the bands they were pirating from had more money than they needed anyhow.
At the same time this happened, many more of the smaller bands that struggled before said screw it to the major labels, and found that if they play to their niche they can end up okay. They'll never strike it big, but they can keep doing what they're doing. It used to be that no established artist would be on an independent label unless they decided to create their own. Today we have many many examples of well known artists with loyal fan bases going onto smaller labels that better support their needs. These places are still going strong, and still will. What the labels are crying foul on is the fact that they can no longer create millions of potential fans who will go out and pay $10-$15 (assuming walmart prices) for the mass produced crap that they're selling. That said, I imagine their revenue stream for the millions of ringtones they sell to people is earning them a nice chunk of money . . . Until people find an easy way to do that themselves that is.
No one likes being searched. Well except for some really sick people, but that's a different story. The question is, is this search based off profiling better than a purely random search? Sure, a 1% success rate seems low, but I bet it's a hell of a lot higher than the success rate of a purely random search. I can almost guarantee that far more than 99% of the people entering into an airport aren't doing so with the intention to pull off something illegal. Most just want to get on their plane, and get to their destination. Searches in general annoy them, and they could care less whether they're being searched for acting suspicious or because they were "randomly" selected.
This brings us to why this 1% success rate could be a good thing for people. Say the government has two choices in how to select people for searches, one being a purely random search, and another being this "targeted search" mixed with a random search. In the purely random search they might choose to search one out of over 100 people. In the targeted approach they might target 1 out of 1000 people, and then search 1 out of every 50 people randomly. In the end we are now annoying 1/1000 people for this targeted search, however half as many people are being randomly searched. To the majority of people walking through, this ends up being a very good thing.
If however the number of people "randomly" searched stays the same, then this new search means that almost 10% more people are being searched when going through airport security. This would be considered an annoyance. But it's likely not the case that security had a choice of using a targeted method versus leaving things the same. They were likely told they need to increase the chances of catching someone. If they can do this with only searching 10% more people, it seems to me that it's a lot better than if they doubles or tripled the number of people they randomly searched.
When evaluating a policy like this, you need to think pragmatically. Is it reasonable that the idea of searching people getting on a plane is going to go away? If not, then what is the best way to pull it off? It's like telemarketers who use targeted advertising. Sure it sucks for the people targeted for it, but if the other option was them targeting half the population, it's a net win for society (assuming in this case that no new telemarketers enter the game because of this). TSA is too entrenched to go away any time soon. It's humiliating to go through security, sure. But if they can find ways to reduce the number of people selected for searching, and still increase the number of people caught for doing something illegal, is that really such an awful thing?
I agree, MOO1 is one of the best games ever. You don't have to spend time micromanaging your empire, and there was a lot of strategy involved. While adding one or two other things to the planetary sliders wouldn't hurt, MOO2 really went overboard. I tried playing it not too long ago, and just didn't like it at all. I had no clue what I was doing, and it started to feel way too much like the civ series. MOO1 was simple, and the simplistic nature worked great for it. That way you could focus on things like you're research and fleet building, plus of course which of your enemies to attack. Some things could be changed (like making it so that building planetary shields was not required to build missile bases), but they're really minor points.
Also, after playing MOO2 briefly, I really hated their new combat mechanics. MOO1 oversimplified combat, sure, but they made it fun and intuitive. Of course, some updates to the combat are necessary. There were just too many cheap ways to fool the computers AI and destroy huge fleets at a time. Some of these could be addressed with a better AI that realizes that it's not worth sacrificing half your fleet because you have one scout sitting at a planet with 75 missile bases.
MOO1 is one of those classic games that I regularly end up firing up on dosbox. It's just too much fun not to play, and still challenging on impossible level (although, I wish there were a slightly better midgame.... it seems to go from being massively behind to all of a sudden being unstoppable.
and if OSX or ubuntu had the 90+% market share windows had you'd have the same problem with spyware on them. Much of the spyware is installed by people who don't know what they're doing. This is NOT microsoft's fault. They've actually tried addressing the problem with Window's Defender. As much of the spyware doesn't need root privileges, once it's installed, it's there. If it requires a root kit, than it's the same on most any OS. There will be security holes, and they will need to be patched. This is a fact of life. Plus, there will be users who will enter their root password for an application they've never heard of just because the box pops up. Blaming MS for the spyware problem on windows just doesn't make sense (and for the record, I'm writing this on my Mac laptop, and run many linux machines on my network).
Slashdot Mirror
the rate that a country prints money is only one factor in the inflation rate (not to mention that most money today isn't really printed). In fact, a country that has a constant money supply would end up suffering from deflation if they did not increase their money supply over time. This is because everyone is becoming richer (and there are likely more people), but the same amount of money is in circulation. This means the money that people have buys more than it did before. In fact that is one of the biggest problems that economies based on the gold standard (or other similar standards) had to face. Deflation is far more devastating on an economy than inflation, as it discourages investment and savings.
Phil
Yeah, because it's not like other OS's haven't gained a significant marketshare in the past few years. I'm pretty sure OSX's marketshare has improved a couple fold in the past 5 years, so it looks like your argument is just full of it.
Phil
We all know the only real winners will be Apple, and Windows XP. Linux will likely carry along with it's .8% market share or whatever it has been at for the past 10 years or so. If these allegations are true, and hold in the final version (remember this is BETA software), it could be problematic. However, there could also be good security reasons for these changes (allowing applications to register what dlls they use and not running if they're changed is a good security practice that can prevent third party applications from breaking their software through the insertion of trojans and/or adware). The inability to fix some of the issues is also probably due to the beta nature of Windows 7.
As for the sound issue, do we really know that this is the OS doing it, and not the driver manufacturers not having this feature implemented in their driver yet? Lots of things could be at fault, and to call DRACONIAN DRM on it is a bit hasty.
Phil
Actually, this implies that the difference in power of JUST the CPU is far more than 50%. This is because most of the system's power draw is NOT from the CPU. Lets assume for instance that at idle the Atom processor consumes 5 watts of power (the rest of the system consumes the rest). This means that at idle the VIA cpu consumes almost 8 watts of power (both these estimates are perfectly reasonable based on the class of processors that they are). This estimates means that when idling the VIA cpu is consuming 50% more power than the atom.
Now, when the cpus are at load, the Atom processor is consuming ~8.5 watts of power, and the VIA is consuming ~25 or 26 watts of power. This looks to me that the via processor is consuming 4 times the power of the Atom, not merely 50% more.
Of course, this estimate is assuming that the Atom processor's idle power is only 5 watts. In reality, the idle power it consumes is likely even lower, as it was designed to minimize power dissipation. Now, claiming that the VIA's system power is approximately 50% more than the Atom is not accurate, but that doesn't mean that the CPU is not consuming that much more power. Anyone doing a fair comparison between the processors would likely be focusing on the difference in power of the CPUs themselves. Otherwise in a full system, the difference between a CPU that requires 50 Watts of power, and one that requires 100Watts of power wouldn't be that significant.
Phil
Actually, this is basically true anytime. If you complain to your cable company and threaten to cancel they will almost always move you to a cheaper rate (assuming you're not on the cheapest introductory rate). They make money off a customer paying $70/month (for cable and internet) and those who pay $100/month. Obviously, they prefer the ones who pay them more. However, they realize that they're better off with a customer paying $70/month than having no customer at all. You just have to be willing to say you want to cancel your account, and either pull the trigger when needed, or hang up before you do.
Phil
The real question that no one is asking is what time period is this 23000% sales jump over? IBM could claim a 230,000% sales increase day over day on occasions if they land a big client on one day or whatever. The same is true for most any company. While this sales jump is large, the real question is what is the real impact on DVD sales over months or years timespan. The day they're released on youtube, you'd expect a big jump like this. You have a large fan base who haven't seen the show in years, someone links them to this, or it's advertised on youtubes page, and they go "oh yeah, i've been wanting that" and buy it. However how does this effect sales after the initial excitement dies down. Hell, I'm sure on a normal day, a slashdot story about monty python would increase their DVD sales a couple fold over the previous days sales. It all depends on the initial volume of the previous day etc. Of course, as they're at the top of sales right now, this is a fairly big deal. But if they hold on a sizable increase in sales 2 months from now, then people will seriously consider this model.
Phil
of course... in any country owning land is only a technicality. I can't think of any country that doesn't have some form of eminent domain laws on the book. Technically they're supposed to compensate you with full market prices, but I know that doesn't always happen, and doesn't account for the personal significance often associated with a location. If a business is allowed to operate on a set piece of land that they paid for indefinitely, it's the same to them as owning it. While the government can seize it or force it to cease production (known to happen in china, and elsewhere), this obviously isn't the standard case. If it happened to every business that operated in China, no one would be spending the money to open new factories there. It's how the market works.
And by free market, there are many different aspects involved. These include overall tax rates, tariffs (if they exist), regulations, government incentives, etc.
Phil
China isn't remotely a communist country by the real meaning of the word. The only reason their economy has been booming in recent years is because they have opened up their markets (unlike a true communist country such as North Korea). Basically, you can have an authoritarian regime without communism (like china), although I've yet to see an instance of a communist regime that isn't authoritarian. At this point in time, I wouldn't be surprised if China's markets are actually freer than most of Europe's markets.
Phil
And if the majority of people think the internet should be censored, so be it. If the majority of people think blacks should be slaves, so be it. The whole rule of the majority thing is a very very dangerous way to talk about the government. A majority rule system is just as deadly as being ruled by a tyrant, it's just that now the tyrant happens to be 50.1% of the population who doesn't like you or something about you.
"Democracy is based on the assumption that a million men are wiser than one man. How's that again? I missed something. Autocracy is based on the assumption that one man is wiser than a million men. Let's play that over again too. Who decides?" --Lazarus Long
Phil
So under this new amendment to the constitution, a 15 year old hacker who breaks into someone else's computer and illegally accesses their personal information (or reads someone's instant messages sent on an unsecured wireless network, etc) can be tried for treason for committing a search without a warrant . . . You do know what the penalty for treason normally is, right?
Also, such a law would basically make it impossible for congress to operate at all, as laws are regularly shot down by the constitution. One goes just a little too far, or a judge redefines what one particular clause in the constitution means. Personally, I could see some good of it, any federal politician caught trying to bail out the auto industry, or redistribute income etc etc could be punished with treason for violating the constitution.
Basically such an amendment would (like most government solutions) cause more problems than it solves.
Think before you open your mouth next time.
Phil
Even in the era of muscle cars, or the more recent SUV era, the gas tax modified behavior. It made gas more expensive, and made the cost of driving better equal the actual cost of driving. I am not talking about creating prohibitive gas taxes whose goal is to reduce consumption, and generate more money than it costs to build and maintain the road systems. I am talking about keeping gas taxes at a level to pay for all construction and maintenance needed for the roadways.
There is a big difference between the two, and just because it doesn't change peoples behavior drastically does not mean it does not influence people's behavior in the long run. If anything, before gas prices plummeted, the gas taxes actually influenced people behavior LESS than they did in the past. If gas taxes are say 25 cents/gallon, people will notice it and see it as a bigger deal when gas costs $1 normally ($1.25 after tax). The difference in behavior between when gas is $3.75 a gallon and $4.00/gallon is likely less.
Phil
Actually, with gas taxes, part of the reason for them is precisely to modify peoples behavior. In general, I agree that taxes shouldn't be used to modify peoples behavior. However, this is an instance where the tax is levied on something that allows you to use that states resources. The states are building billions of dollars worth of highways, and they don't want to encourage people to use these resources inefficiently. They want to modify people's behavior so that they aren't as likely to crowd the roads (encouraging people to consolidate trips, carpool, etc). The roads need to be funded somehow, and a usage tax is one of the most effective ways to handle it. In fact, I wish all the money paying for roads only came from gas taxes. That would end up with a more even and natural balance where supply meets demand.
Taxing gas to pay for building roads is done precisely to help balance out the supply and demand for using roads. Roads entail a large fixed cost to build, and a smaller marginal cost for wear and tear on the road. If the government didn't tax the usage of the roads we could end up with the tragedy of the commons. Taxing them fixes this so that the supply and demand meet more evenly where the costs of the road best offsets the benefits given by the road. And yes, this requires modifying peoples behavior. Otherwise, people would overuse the roads, and we'd end up needing extra roads that don't provide enough additional utility to the people to justify the cost (as they wouldn't have needed them if they had to pay to reflect the roads cost).
Phil
Without even doing any calculations of the amount of fuel needed to ship the CFL, it's obvious that this environmental cost is more than offset by the savings in power of using a CFL vs an incandescent. The simple way to figure this out is to look at the price, and the cost savings of the bulb over a couple months time. A new CFL only costs a couple dollars to buy. That means (unless companies are selling them at a loss) that it costs less than this amount to ship the bulb to you. Assuming for a minute that the cost of energy in fuel and coal powering a power plant are the same (with the same efficiencies etc, in reality the power grid powered by coal is generally cheaper per watt, but that's okay), if you can make up the cost of the bulb in a matter of months by lowering your electric bill, you have more than made up for the cost of shipping.
I always laugh when people start talking about carbon footprints and all that. Currently the simplest and only real way to drastically reduce your carbon footprint is to spend less money. Things that cost more in general have a greater carbon footprint (there are exceptions to this rule, but it often holds true). So the footprint of manufacturing a new stove is roughly proportional to how much the stove costs (obviously if you're buying a gold plated stove with platinum racks this won't quite hold true). So the quickest way to halve the US's carbon footprint is to cut the US's spending power in half. Of course, with the steps the government is taking to reverse the current economic crisis, they could easily accomplish this.
Phil
First, you sound bitter from a bad experience with one particular IEEE conference. Every IEEE conference is different, and the IEEE is just a society that sponsors some of the conferences in the field, and doesn't necessarily perform quality control on them. That is handled by the conference organizers themselves, and better conferences tend to have better quality control. The people in the field should know which are the better conferences.
Second, the fact that you reviewed a couple really awful papers has little bearing on whether your paper will be accepted or not. Now, if you could claim that these awful papers got accepted, while yours was rejected, you might have something to say. But if those papers were also rejected, your statements about them have absolutely no merit.
Also, sometimes you get bad luck with reviewers. You might get reviewers who think the entire premise behind your research is crap, or ones who don't like your group for whatever reason. Sometimes you get one who just happened to be having a bad day, and wanted to find fault with your paper. The hope is that with enough reviewers, it is unlikely that you will get only crappy reviewers. If that does happen, I'm sorry, try again. Also, don't put too much stock in the timestamps given. Those could have been modified for any number of reasons. It's highly unlikely that a reviewer would submit final reviews 2 minutes after getting the paper. Chances are good they'd save the paper, and wait a few weeks before looking at them right before the deadline. The only time I'd spend 2 minutes reviewing a paper is if it was either a) Completely unreadable (like the computer generated gibberish), or b) it was obviously plagiarizing other papers. In those cases, it isn't worth more than 2 minutes of my time to give it a valid review.
Hopefully next time you have better luck with reviewers, in my experience I tend to get some informed comments back on my papers that can help guide changes I want to make whether it gets accepted or not.
Phil
not everyone who uses a linux workstation is a sysadmin. If someone else has set up the box, and all they do is run a handful of applications on the machine, they don't know anywhere near enough to really run linux at home. If you're in a shop where EVERYONE knows how to admin their own machine, and runs windows, sure. This isn't the case in the real world.
Phil
Sure, you could try expanding the instruction set, and reach the point where a single instruction can do the work of 5-10 instructions. . . . Unfortunately doing this will likely reduce your operating clock frequency by a factor of 5-10, while increasing power and area by a similar ratio. Net gain, next to nothing. Plus of course, now instructions end up taking up 5-10 times as much storage space.
The problem is that all reasonable methods for improving the performance of single threaded code in convential ISAs requires changing the programming paradigm (backwards compatibility is still more important than power/performance). The first method people used was pipelining the processor core. This worked, and let chipmakers ramp up the clock frequency. However, it has limited gains. For one, there is a limit to how deep you can pipeline your work, because at the bare minimum your clock cycle has to be the speed of a latch, and one degree of logic (otherwise you're not doing anything in a single pipe stage and/or can't store the results). We reached the point where we couldn't feasibly increase the pipelining depth, and increasing it further pushed up the clock speed a little bit, but didn't actually improve performance (think Pentium 4). What else can be done ? Well, one idea that's been used for a while is executing multiple instructions at a time. We're doing this, modern processors can execute 3-4 instructions in a single clock cycle, of course they average around 1...
We could try going further with this idea, but limit studies have shown that their is no feasibly way to gain considerable performance out of current codes by looking at more instructions simultaneously. There is just a lack of parallelism available in single instruction streams. Secondly, memory and branch prediction greatly reduce the performance that a wide superscalar can achieve. Even a branch predictor that only misses one out of a thousand instructions is missing too frequently to obtain high performance on an 8-16 wide core.
The next logical steps are extensions to the ISA. This is being done, but still requires better compilers. SSE2 etc can greatly speedup vectorizable codes. Of course, the compilers fail here normally, and the only reason this gives a speedup is because it involves people specifically coding for a parallel architecture.
There are other methods being investigated to achiever higher performance, however all of them involve a new programming paradigm to some extent. Personally, I don't think functional programming is the answer. It might be part of an answer, and it might help with some stuff, but it's just an unnatural way to express statements. More explicitly parallel languages like verilog might become more common, however these are much harder to use than traditional programming languages.
What needs to change is the overall way we think about programming. We're taught to take problems and express them serially. Doing anything else is unnatural in most languages. Serial programming allows determinism in the code, and as we go forward, we need to reduce our need for determinism, or first split up an application into parts that can be performed deterministically, and parts that don't need to be. This will help with improving parallelism. Additionally, we need far better debug tools. Running multiple instances of GDB trying to trace threads just doesn't hack it. This will also require far greater unit testing in code development.
The good news however is that many applications don't need to be faster. They are fast enough as they stand, and only small sections need to be accelerated. The vast majority of code is for special case conditions that are rare anyhow, so we need ways to accelerate the cores of applications without worrying about the rest. As we go forward I think this will become an increasingly popular option. It won't be easy, and at a high level, applications need to be restructured to support greater parallelism. Whether we'll see parallelism a
Research is a fuzzy area. Particularly at academic institutions. The general rule is that any research you do while in the university is joint work with your advisor. It's just the way things are. If you don't like this sort of apprenticeship system, don't try to be a researcher. Getting a doctorate requires going to grad school. It's just the way things are.
It doesn't matter if your grant is for doing research on X and you discover Y in the process. The research you are doing is for your advisor. You ARE NOT doing it solely for yourself, and your funding status shows this. However, most professors at academic institutions have a lot of discretion concerning this stuff. Talk to them. Don't bother talking to lawyers or any of that stuff yet, talk to your advisor. Ask them how they plan on allowing you to contribute back to the community, or what you are allowed to take with you.
Most academic work is joint in nature, and oftentimes the code can later be released under some other license (BSD, GPL, limited rights, etc). Normally the code that you wrote isn't very useful outside of an academic setting, and would need to be overhauled to be useful elsewhere. If the idea is extremely good, and you know it is, you might want to talk with them about doing something outside of the university on it, or more likely obtaining a patent on it. Professors are generally allowed to consult a set number of hours per week, however there are always conflicts of interest that they need to be careful about. If the idea is not directly related to any grant they currently have though, they can likely work on it with you . . . That is if they're willing to. Also be warned that if you're not on good terms with your advisor to be careful, as this could backfire.
If you just want the code to be available to you later after you graduate for your students to use (if you plan on going into academia), you definitely want to talk with them about that. In general however you need to differentiate your work as a professor from your work as a graduate student, so often times the tools you'd want your students to use would be different anyhow (depending on how general they are). Many professors are quite willing to help you out, particularly if you're going into academia. Remember, it looks good for them if their former students later go on to become well-respected professors.
The likely case is that you're grossly over valuing your code. For most academic projects, the ideas generated, and the data obtained are far more valuable than the tools used to generate them.
Phil
as long as one person uses linux at work, and uses windows at home, this IS a fact. I don't need much proof to back it up. But in the lab I work in, I know enough people who regularly do work on Linux workstations, and have Windows on their personal machines (some do use Mac or Linux machines at home however). For the most part, I wouldn't want to use Linux on a laptop (the most common personal machine), while it's getting better, it tends to be a hassle to get wireless, multiple monitors, power management and such to "just work". I know it can be done on some machines with different hacks and settings, but most people don't want to deal with that. Personally, I don't want to deal with that, and thats why I use a Mac. However, the difference is that I'm a UNIX geek, and many of the people I know are not. They can use UNIX, and view it as a tool. Whatever gets the job done the best is what they tend to use.
Phil
actually, your statement is a great exaggeration of what the facts say. You are implying that most people don't want to use windows, but are forced to use it. This is NOT something you can claim from the statistics. It seems more likely that maybe 15% of home users use Macs, and 5-10% of business users use Macs. Therefore you have 5-10% of people who normally use Macs being "forced" to use Windows. There's a big difference between that and saying "people don't want windoze but are forced to use it at work.", where you make a generalization covering all people.
The same thing could be said about Linux actually. There are quite a few people who use Linux workstations at work, but have windows PCs at home (often because their home PC is a family PC). By your logic, I could say "people don't want to use Linux, but are forced to use it at work".
Phil
And I'll ask you a question, how far into a system does a virus have to penetrate to achieve the virus-writers objective? Does a virus need to have root access to be part of a Botnet? Does malware need root access to start issuing excessive popups to the user? Does malware need root permission to run when a user logs onto a machine? Didn't think so. The only reason many of these malware programs get root access is because they can, and doing so makes them harder to detect, and harder to remove at a later point. Considering the majority of viruses are out there to make money (popups, botnets, etc), you really don't need root access to do this. No reasonable amount of security at the OS level will remove the threat of locally run malware on the system. It just can't be done unless the system restricts what programs can and cannot be run.
Phil
You're right that artists can make labels money with a little marketing. However, you missed the biggest form of marketing that artists have. . . Playing live shows. For small unknown bands, that is how they get known. If you're opening for bands with similar styles (which you should be doing), and play good, the audience will appreciate it. If only 15 people at a show buy your CD afterward, you're doing pretty good with marketing (assuming the show has ~200 people there). These 15 people will likely tell others, and word of mouth spreads about the band. I've seen it many times where good bands can grow remarkably fast from this. Of course, this won't make them huge, but it means they can sell on the order of 20,000 records, which isn't too bad.
There are some record labels that take this into account when looking for new bands to sign on. I know New West records (one of my favorite record labels, where I'll randomly pick up albums by artists I don't know just because they're on the label) is really big on live music. One of the guys who runs the label is a huge music fan, going to tons of concerts a year. If he see's a band he likes that is doing there part, he'll often try to pick them up onto the label. The demos a band has are important to the label, but equally, if not more important is the bands live show. Additionally, the band (assuming they're not well established already) has to be willing to hold their end of the agreement, and tour regularly. If you're not willing to spend 100s of days a year on tour promoting your music, you're not worth the investment to them.
Phil
Your last statement is only partially true. The number of releases record labels make in a year has reduced, and they have started catering more to the very large stores. However, this was the case quite a bit before Napster came onto the scene, and the mp3 revolutionized the music industry. The large stores had started swallowing up the small. It is false to say that the small independent record stores are gone. There are still quite a few of them that flourish. The people who love music and are willing to pay for it often choose to go to the small independent stores, and always will. The store is far more about the personal interaction, and the recommendations that can be made there. What has hurt the small stores the most isn't the Walmart's and Best Buys who have small selections, low prices, and high volume, but the Amazon's of the world who have practically unlimited selection, and the benefits of scale that come with being so large that they can get lower prices. Many music fans have started shopping online for CDs they used to buy at the small independent stores.
The large record companies shot for the gold in the late 90s by focusing on the big hit of the day kind of thing. Under such a market, they create demand for music, and sell CDs (albeit to a limited number of artists). However, when you concentrate on the masses and the hit of the minute, you lose out on loyalty. The loyal fan base that goes to the small independent CD stores didn't want to switch to buying CDs at Walmart, and they didn't care about the flavor of the month. They bought lots of CDs by bands that aren't particularly profitable to the labels (but tended to bring in a steady stream of income). The group that they won over with low prices at Walmart, and mass consumed discs has little loyalty, and why should they. They could care less where they get their music from, and napster is as good a place as any, but the price was right. Besides, who cares if the back street boys didn't make a few extra bucks, the bands they were pirating from had more money than they needed anyhow.
At the same time this happened, many more of the smaller bands that struggled before said screw it to the major labels, and found that if they play to their niche they can end up okay. They'll never strike it big, but they can keep doing what they're doing. It used to be that no established artist would be on an independent label unless they decided to create their own. Today we have many many examples of well known artists with loyal fan bases going onto smaller labels that better support their needs. These places are still going strong, and still will. What the labels are crying foul on is the fact that they can no longer create millions of potential fans who will go out and pay $10-$15 (assuming walmart prices) for the mass produced crap that they're selling. That said, I imagine their revenue stream for the millions of ringtones they sell to people is earning them a nice chunk of money . . . Until people find an easy way to do that themselves that is.
Phil
No one likes being searched. Well except for some really sick people, but that's a different story. The question is, is this search based off profiling better than a purely random search? Sure, a 1% success rate seems low, but I bet it's a hell of a lot higher than the success rate of a purely random search. I can almost guarantee that far more than 99% of the people entering into an airport aren't doing so with the intention to pull off something illegal. Most just want to get on their plane, and get to their destination. Searches in general annoy them, and they could care less whether they're being searched for acting suspicious or because they were "randomly" selected.
This brings us to why this 1% success rate could be a good thing for people. Say the government has two choices in how to select people for searches, one being a purely random search, and another being this "targeted search" mixed with a random search. In the purely random search they might choose to search one out of over 100 people. In the targeted approach they might target 1 out of 1000 people, and then search 1 out of every 50 people randomly. In the end we are now annoying 1/1000 people for this targeted search, however half as many people are being randomly searched. To the majority of people walking through, this ends up being a very good thing.
If however the number of people "randomly" searched stays the same, then this new search means that almost 10% more people are being searched when going through airport security. This would be considered an annoyance. But it's likely not the case that security had a choice of using a targeted method versus leaving things the same. They were likely told they need to increase the chances of catching someone. If they can do this with only searching 10% more people, it seems to me that it's a lot better than if they doubles or tripled the number of people they randomly searched.
When evaluating a policy like this, you need to think pragmatically. Is it reasonable that the idea of searching people getting on a plane is going to go away? If not, then what is the best way to pull it off? It's like telemarketers who use targeted advertising. Sure it sucks for the people targeted for it, but if the other option was them targeting half the population, it's a net win for society (assuming in this case that no new telemarketers enter the game because of this). TSA is too entrenched to go away any time soon. It's humiliating to go through security, sure. But if they can find ways to reduce the number of people selected for searching, and still increase the number of people caught for doing something illegal, is that really such an awful thing?
Phil
I agree, MOO1 is one of the best games ever. You don't have to spend time micromanaging your empire, and there was a lot of strategy involved. While adding one or two other things to the planetary sliders wouldn't hurt, MOO2 really went overboard. I tried playing it not too long ago, and just didn't like it at all. I had no clue what I was doing, and it started to feel way too much like the civ series. MOO1 was simple, and the simplistic nature worked great for it. That way you could focus on things like you're research and fleet building, plus of course which of your enemies to attack. Some things could be changed (like making it so that building planetary shields was not required to build missile bases), but they're really minor points.
Also, after playing MOO2 briefly, I really hated their new combat mechanics. MOO1 oversimplified combat, sure, but they made it fun and intuitive. Of course, some updates to the combat are necessary. There were just too many cheap ways to fool the computers AI and destroy huge fleets at a time. Some of these could be addressed with a better AI that realizes that it's not worth sacrificing half your fleet because you have one scout sitting at a planet with 75 missile bases.
MOO1 is one of those classic games that I regularly end up firing up on dosbox. It's just too much fun not to play, and still challenging on impossible level (although, I wish there were a slightly better midgame.... it seems to go from being massively behind to all of a sudden being unstoppable.
Phil
and if OSX or ubuntu had the 90+% market share windows had you'd have the same problem with spyware on them. Much of the spyware is installed by people who don't know what they're doing. This is NOT microsoft's fault. They've actually tried addressing the problem with Window's Defender. As much of the spyware doesn't need root privileges, once it's installed, it's there. If it requires a root kit, than it's the same on most any OS. There will be security holes, and they will need to be patched. This is a fact of life. Plus, there will be users who will enter their root password for an application they've never heard of just because the box pops up. Blaming MS for the spyware problem on windows just doesn't make sense (and for the record, I'm writing this on my Mac laptop, and run many linux machines on my network).
Phil