Efficient price discovery is the purpose of the market. Liquidity is essential to that end. Traders provide that liquidity and, overall, the market benefits because of it.
Perhaps the problem is that what we really want is efficient value discovery. Price and value tend to diverge if you have a whole bunch of players betting on what the value will be in the future, because the underlying mechanisms creating value (manufacturing, development, etc.) can't change that quickly, but the market perception ("price") can change very quickly.
I would not want to rent my books, because I want to keep them for reference in the career that they are supposedly providing us!
I felt the same way in college but realized that I couldn't afford to keep the latest editions for reference. So I sold all textbooks where the same edition was being used by the next class, and either purchased an older edition for $5 or waited a few years until a new edition came out and then purchased it for $5. In instances where I had purchased a used book and then sold it, I was out a total of $5.
I saved a little more than a grand over 4 years and still have my reference copies.:-)
I've hired 40+ engineers over the last 4 years, and here's my take on a Masters degree.
Best option: combined 5-year Bachelor's + Master's program. You get more technical depth and a Master's on your resume for very little additional money. Your starting pay will be higher, and you can expect to break even in 3 years.
Next best option: 2 year Master's program at a top 5 or top 10 school in your field of interest. If it is not a top ranked program, or you're not changing your field of study (e.g. EE to CS or CS to Robotics) it's not worth the money.
Otherwise, get a job and work on your Masters part-time. Either negotiate an accelerated career track while you're working on it, negotiate a pay increase after you get it, or switch jobs for more pay / more relevance to your new field of study afterward.
The bug is an out-of-order sequencing issue. The application sequence is CreateFile, WriteData, RenameFile. What is actually happening on disk is CreateFile, RenameFile, WriteData. If the crash happens between RenameFile and WriteData, you lose the data written to disk and have a zero length file. This is a filesystem / kernel issue.
The length of time between disk writes exacerbates the problem. sync() forces a write and reduces the window when the filesystem is susceptible, but the bug is still there.
This is a common bug when designing caches, because the sequence of writes of interdependent data must be in-order to maintain integrity.
Sounds great in theory. In practice, it's like getting Ma Bell back, except with less innovation. Nationalizing and merging is the easy part, retaining the talent to keep it effective is next to impossible. -- Hiten
While I do not support drastic political positions on this particular issue such as the "windfall profits tax", I do want to point out that your analysis and conclusions are incorrect.
Wal-Mart made a net profit of $12.8B on revenues of $374.5B for the YEAR ending 1/31/08, or roughly 3.4%.
Exxon made a net profit of $11.7B on revenues of $138.1B or 8.4% for the QUARTER ending 3/31/08, with a percentage profit more than twice that of Wal-Mart, which is one of the most efficiently run businesses today.
Exxon's sells roughly 21.2B gallons a quarter (5.6M barrels/day * 91 days/quarter * 42 gallons/barrel). So their profits are roughly 55 cents a gallon, and that is the real "markup" (markup is a real business term and you and I are both using it loosely). I bet a lot of us would love to have 50 cents back per gallon of gas that you bought in Q1 of this year.
And yes, this is price gouging because these are not GROSS margins but NET margins - after every cost has been recouped, including all product cost, depreciation of assets, R&D, marketing, executives salaries and bonuses and perks, and shareholders dividends. Moreover, because of GAAP rules and Wall Street expectations, profit money can only be spent on assets - buying other companies, property, etc. So this money is not going to yield any substantial R&D investments in alternative energy either.
At the same time, General Motors is making huge losses, partly because of their lack of innovation but also because of the high price of gas, resulting in plant closures. Plant closures in a particular business in and of themselves are not bad as long as they happen over time. When they happen quickly, the people who have been working in the auto industry cannot simply start working in the oil industry because their skills don't translate. That is one of the many *real* tragedies of the windfall profits of Exxon.
One could say the same about filesystems - but we figured out how to abstract the filesystem API in UNIX a long time ago. This led to a lot of innovation in filesystems - ext2, ext3, ReiserFS, AFS, ZFS, etc. I think we might see similar innovation in schedulers if the scheduler was pluggable. At the very least, I suspect that Con Klivas would still be a kernel developer had we supported pluggable schedulers, and that alone might justify making the scheduler pluggable.
I expect that there would be a performance impact if the scheduler were pluggable - modular and optimized do not generally go together. However, the worst case performance of any scheduler dominates the user experience, so IMHO, it is worth accepting a small performance penalty to enables competition and innovation toward reducing the worst case performance.
Your idea is pretty interesting, but it needs to be developed further.
The two examples that you cited are moral improvements that occurred as a result of scientists to work to improve something. What also needs to be considered is why the scientists were working on those tasks. Assuming that the research was not being done in academia, in both instances the motivation would be economics - the first one is convenience -> productivity -> faster research -> lower costs, and the second one is low pollution -> tax incentives for car companies or revenues for alternative energy companies -> profits.
While improvements over current paradigms can in certain instances address moral issues, we must also remember that using rats in the first place, or designing gas guzzlers would not be possible without scientists / engineers. Perhaps there were moral isues that we solved in the past that created what we consider moral problems today - e.g. you could argue that using rats is more moral than using prisoners on death row, or that using cars was more moral than using horses, but I would argue that it's economics and convenience (which is also economics). Cars are cheaper than horses and require less maintenance, and rats are easier to obtain than prisoners (also economics in a twisted way).
I think we end up in the same place that all technology related moral arguments end up - technology itself is amoral, but can certainly be directed at solving or creating moral problems. I do agree that scientists / engineers are the ones building the technology that solves or creates these problems, but generally the motivation for this is economic unless it is done in a non-profit environment such as academia.
Now if we could align ethics and economics, everyone could be a moral crusader. Although at that point, either ethics or ecnomics would redudant, and therefore one of them would have to be obsolete (no economics in Star Trek).
-- Hiten
The REAL problem with hydrogen, which everyone loves to ignore, is that there IS ABSOLUTELY NO WAY to produce hydrogen efficiently, from a renewable resource, without leaving toxic byproducts; current methods either involve hideously inefficient electrolysis, toxic catalysts, or non-renewable resources.
Agreed. To drive cars on kinetic energy, three things must happen:
1. Energy production
2. Energy transportation
3. Energy conversion
As you pointed out, if we use Hydrogen instead of oil as a fuel, we only solve the the energy transportation problem. The big win, of course, is that it changes the energy production problem from drilling for oil to generating electricity.
There are plenty of methods to generate electricity cleanly (but perhaps still not efficiently) - solar, wind, hydro, and nuclear. Instead of the geo-socio-econo-political challenges that we have with oil, these methods have mostly technical challenges, which can be solved with much less strife.
(1) Turns out huge improvements in fabrication made this argument false and somethings like branch predictions just can't be done in software. Fast dedicated hardware is faster than software. Who came up with this idea of moving optimization to software?
and then you say:
(2) Transmeta had something interesting
I just want to point out that Transmeta did exactly what you are complaining about in #1... they moved optimizations to software, and proved that their solution was in the same performance ballpark while consuming substantially less power, because the underlying ISA was more power efficient.
Now, if one can come up ISA is better performing - which is basically a function of chip area, because one can do more things in silicon instead of uOps if one has more transistors - and usually a much smaller area is dedicated to a hardware compatibility unit, then implementing the compatibility in software running on the primary ISA is going to result in better performance.
At the processor level, hardware and software are pretty blurred. This is why CISC processors scream these days - the instructions are translated and optimized into a different ISA, which the processor actually executes well.
$11 billion? No problem. Now he has a reason to respond to the several debt consolidation services that advertise "a low monthly payment", and the home equity loan lines that "guarantee a loan despite bad credit".
-- Hiten
If you enjoy working on robots or being around robots, we have several openings listed on our careers web page. You don't have to be an engineer - it takes a lot of different types to operate any real company.
In addition to the open positions, we generally have room for interns, especially if you are a hands-on type.
All of our openings for both the Consumer and Government divisions are in our Burlington, MA headquarters located about 20 minutes from downtown Boston. The Government division also has a small facility in San Luis Obispo, CA about 4 hours from LA, where we make rare hires when we find the right people.
Drop me a note at "hsonpal at our domain name" when you apply - I'll let HR know that I'm referring you.
Apple could be dominating the I.T. world like Microsoft now does, if not for the poor business decision they made when they got started of pricing their computers above IBM's crappy PCs. Giving more clout to smart business men at that time could have changed things.
You're making an assumption that pricing decisions were made by engineers at the time. More likely than not, they were made by business people then just as they are made by business people now. The difference is that this time, the organization understands how to price the product and maket it so that price > perceived value and not the other way around.
They should try the Spartan3Es. Much cheaper, with support for cheap PROMs. 1.2M gates for $9 isn't bad at all. Heck, if they call and let Xilinx know that their vision is to have every Linux server use a video card with a Xilinx on it, Xilinx might design it for them for free. And a sale price point of $99 is easily achievable - after all, the Spartan3E Starter Kit is only $149 and has a lot more stuff!
So a decent first cut at filtering bulk spam (and recall that both Yahoo and Hotmail use "bulk mail" folders) would be to take an MD5 sum of each email (not including the "To" address header lines, of course), stick the sum in hash table or other database, and increment a counter for each email with that MD5 sum. Once the counter reached some arbitrary large-ish number, you'd mark all copies of that emails spam.
You are already two steps behind the spammers. Have you not received e-mails where the text of the message contains a number or a random-ish looking string in the e-mail text that is out of place? The string or number is different for every e-mail that is sent.
These days what you need to combat spam is a content analyzer, which is not nowhere near as easy to implement as a md5sum.
Andrea did some cool stuff. How come our high schools don't teach us how to:
Live off the land
Modify our cars
Hack computers
Understand personal finance
Write contracts
Defend ourselves in court
Defend ourselves physically
Handle a gun safely
Think critically
Change our government for the better
It seems to me that too much focus is given to understanding the past and not enough to understanding the present. Don't get me wrong, knowing the past is valuable, but I think that if we teach people about the present, people are naturally going to be interested in the past.
In general, people don't need to know how to calculate the area under a curve. But everyone needs to know how to think critically and not be manipulated.
... those horrendous semi-proprietary mini-PCI cards...
Mini-PCI is an open standard, just like PCI. You even buy the specifications from the same place, the PCI SIG. What nVidia is doing is pre-empting what the PCI SIG will eventually come out with, perhaps in the hopes that the PCI SIG will adopt their standard as the official PCI graphics standard for laptops.
If a couple of big players like Dell and Toshiba adopt it for their notebooks, this will most likely become the standard.
Someone else on this thread added that CardBus does exactly what you are asking for. I would also like to add that audio capabilities should be absorbed into the video card anyway so that you can send it a single datastream and get both video and audio out, so you don't need to add audio capabilities to the bus.
No that's not entirely accurate, because we have pass in city driving as well, and probably have to do it more often than on the highway.
The real reason why mileage is lower on the highway is because air resistance is proportional to the square of the velocity, and since highway speeds are usually twice as high, you have to fight four times the air resistance.
This also means that the Prius is one amazing car... close to 100% of the energy that used to be wasted braking (typical of city driving) is now recovered.
As always, we have to choose the lesser of two evils. There are good candidates on both sides, but somehow they don't make it to the oval office. In particular, Wesley Clark on the Democratic side and John McCain on the Republican side could do this country a lot of good.
I think that the party lines no longer break up between true conservative and liberal any more. For example, I think not going to war is the true conservative thing to do... and strangely enough, on that issue, Howard Dean ends up looking like the most conservative of any of the candidates who were running for President.
Re:Gary Gygax
on
D&D Is 30
·
· Score: 2, Informative
Slashdot Mirror
... the business model where I am the product to be sold to advertisers. I prefer a business model where I am a customer of the content. -- Hiten
Efficient price discovery is the purpose of the market. Liquidity is essential to that end. Traders provide that liquidity and, overall, the market benefits because of it.
Perhaps the problem is that what we really want is efficient value discovery. Price and value tend to diverge if you have a whole bunch of players betting on what the value will be in the future, because the underlying mechanisms creating value (manufacturing, development, etc.) can't change that quickly, but the market perception ("price") can change very quickly.
I would not want to rent my books, because I want to keep them for reference in the career that they are supposedly providing us!
I felt the same way in college but realized that I couldn't afford to keep the latest editions for reference. So I sold all textbooks where the same edition was being used by the next class, and either purchased an older edition for $5 or waited a few years until a new edition came out and then purchased it for $5. In instances where I had purchased a used book and then sold it, I was out a total of $5.
I saved a little more than a grand over 4 years and still have my reference copies. :-)
I've hired 40+ engineers over the last 4 years, and here's my take on a Masters degree.
Best option: combined 5-year Bachelor's + Master's program. You get more technical depth and a Master's on your resume for very little additional money. Your starting pay will be higher, and you can expect to break even in 3 years.
Next best option: 2 year Master's program at a top 5 or top 10 school in your field of interest. If it is not a top ranked program, or you're not changing your field of study (e.g. EE to CS or CS to Robotics) it's not worth the money.
Otherwise, get a job and work on your Masters part-time. Either negotiate an accelerated career track while you're working on it, negotiate a pay increase after you get it, or switch jobs for more pay / more relevance to your new field of study afterward.
-- Hiten
The bug is an out-of-order sequencing issue. The application sequence is CreateFile, WriteData, RenameFile. What is actually happening on disk is CreateFile, RenameFile, WriteData. If the crash happens between RenameFile and WriteData, you lose the data written to disk and have a zero length file. This is a filesystem / kernel issue.
The length of time between disk writes exacerbates the problem. sync() forces a write and reduces the window when the filesystem is susceptible, but the bug is still there.
This is a common bug when designing caches, because the sequence of writes of interdependent data must be in-order to maintain integrity.
-- Hiten
Sounds great in theory. In practice, it's like getting Ma Bell back, except with less innovation. Nationalizing and merging is the easy part, retaining the talent to keep it effective is next to impossible. -- Hiten
Wal-Mart made a net profit of $12.8B on revenues of $374.5B for the YEAR ending 1/31/08, or roughly 3.4%.
Exxon made a net profit of $11.7B on revenues of $138.1B or 8.4% for the QUARTER ending 3/31/08, with a percentage profit more than twice that of Wal-Mart, which is one of the most efficiently run businesses today.
Exxon's sells roughly 21.2B gallons a quarter (5.6M barrels/day * 91 days/quarter * 42 gallons/barrel). So their profits are roughly 55 cents a gallon, and that is the real "markup" (markup is a real business term and you and I are both using it loosely). I bet a lot of us would love to have 50 cents back per gallon of gas that you bought in Q1 of this year.
And yes, this is price gouging because these are not GROSS margins but NET margins - after every cost has been recouped, including all product cost, depreciation of assets, R&D, marketing, executives salaries and bonuses and perks, and shareholders dividends. Moreover, because of GAAP rules and Wall Street expectations, profit money can only be spent on assets - buying other companies, property, etc. So this money is not going to yield any substantial R&D investments in alternative energy either.
At the same time, General Motors is making huge losses, partly because of their lack of innovation but also because of the high price of gas, resulting in plant closures. Plant closures in a particular business in and of themselves are not bad as long as they happen over time. When they happen quickly, the people who have been working in the auto industry cannot simply start working in the oil industry because their skills don't translate. That is one of the many *real* tragedies of the windfall profits of Exxon.
-- Hiten
One could say the same about filesystems - but we figured out how to abstract the filesystem API in UNIX a long time ago. This led to a lot of innovation in filesystems - ext2, ext3, ReiserFS, AFS, ZFS, etc. I think we might see similar innovation in schedulers if the scheduler was pluggable. At the very least, I suspect that Con Klivas would still be a kernel developer had we supported pluggable schedulers, and that alone might justify making the scheduler pluggable.
I expect that there would be a performance impact if the scheduler were pluggable - modular and optimized do not generally go together. However, the worst case performance of any scheduler dominates the user experience, so IMHO, it is worth accepting a small performance penalty to enables competition and innovation toward reducing the worst case performance.
Your idea is pretty interesting, but it needs to be developed further. The two examples that you cited are moral improvements that occurred as a result of scientists to work to improve something. What also needs to be considered is why the scientists were working on those tasks. Assuming that the research was not being done in academia, in both instances the motivation would be economics - the first one is convenience -> productivity -> faster research -> lower costs, and the second one is low pollution -> tax incentives for car companies or revenues for alternative energy companies -> profits. While improvements over current paradigms can in certain instances address moral issues, we must also remember that using rats in the first place, or designing gas guzzlers would not be possible without scientists / engineers. Perhaps there were moral isues that we solved in the past that created what we consider moral problems today - e.g. you could argue that using rats is more moral than using prisoners on death row, or that using cars was more moral than using horses, but I would argue that it's economics and convenience (which is also economics). Cars are cheaper than horses and require less maintenance, and rats are easier to obtain than prisoners (also economics in a twisted way). I think we end up in the same place that all technology related moral arguments end up - technology itself is amoral, but can certainly be directed at solving or creating moral problems. I do agree that scientists / engineers are the ones building the technology that solves or creates these problems, but generally the motivation for this is economic unless it is done in a non-profit environment such as academia. Now if we could align ethics and economics, everyone could be a moral crusader. Although at that point, either ethics or ecnomics would redudant, and therefore one of them would have to be obsolete (no economics in Star Trek). -- Hiten
Agreed. To drive cars on kinetic energy, three things must happen:
1. Energy production
2. Energy transportation
3. Energy conversion
As you pointed out, if we use Hydrogen instead of oil as a fuel, we only solve the the energy transportation problem. The big win, of course, is that it changes the energy production problem from drilling for oil to generating electricity.
There are plenty of methods to generate electricity cleanly (but perhaps still not efficiently) - solar, wind, hydro, and nuclear. Instead of the geo-socio-econo-political challenges that we have with oil, these methods have mostly technical challenges, which can be solved with much less strife.
-- Hiten
(1) Turns out huge improvements in fabrication made this argument false and somethings like branch predictions just can't be done in software. Fast dedicated hardware is faster than software. Who came up with this idea of moving optimization to software?
and then you say:
(2) Transmeta had something interesting
I just want to point out that Transmeta did exactly what you are complaining about in #1... they moved optimizations to software, and proved that their solution was in the same performance ballpark while consuming substantially less power, because the underlying ISA was more power efficient.
Now, if one can come up ISA is better performing - which is basically a function of chip area, because one can do more things in silicon instead of uOps if one has more transistors - and usually a much smaller area is dedicated to a hardware compatibility unit, then implementing the compatibility in software running on the primary ISA is going to result in better performance.
At the processor level, hardware and software are pretty blurred. This is why CISC processors scream these days - the instructions are translated and optimized into a different ISA, which the processor actually executes well.
$11 billion? No problem. Now he has a reason to respond to the several debt consolidation services that advertise "a low monthly payment", and the home equity loan lines that "guarantee a loan despite bad credit". -- Hiten
In addition to the open positions, we generally have room for interns, especially if you are a hands-on type.
All of our openings for both the Consumer and Government divisions are in our Burlington, MA headquarters located about 20 minutes from downtown Boston. The Government division also has a small facility in San Luis Obispo, CA about 4 hours from LA, where we make rare hires when we find the right people.
Drop me a note at "hsonpal at our domain name" when you apply - I'll let HR know that I'm referring you.
-- Hiten
You're making an assumption that pricing decisions were made by engineers at the time. More likely than not, they were made by business people then just as they are made by business people now. The difference is that this time, the organization understands how to price the product and maket it so that price > perceived value and not the other way around.
-- Hiten
They should try the Spartan3Es. Much cheaper, with support for cheap PROMs. 1.2M gates for $9 isn't bad at all. Heck, if they call and let Xilinx know that their vision is to have every Linux server use a video card with a Xilinx on it, Xilinx might design it for them for free. And a sale price point of $99 is easily achievable - after all, the Spartan3E Starter Kit is only $149 and has a lot more stuff!
You are already two steps behind the spammers. Have you not received e-mails where the text of the message contains a number or a random-ish looking string in the e-mail text that is out of place? The string or number is different for every e-mail that is sent.
These days what you need to combat spam is a content analyzer, which is not nowhere near as easy to implement as a md5sum.
IMHO, this post is OTT with all the TLAs and FLAs.
Live off the land
Modify our cars
Hack computers
Understand personal finance
Write contracts
Defend ourselves in court
Defend ourselves physically
Handle a gun safely
Think critically
Change our government for the better
It seems to me that too much focus is given to understanding the past and not enough to understanding the present. Don't get me wrong, knowing the past is valuable, but I think that if we teach people about the present, people are naturally going to be interested in the past.
In general, people don't need to know how to calculate the area under a curve. But everyone needs to know how to think critically and not be manipulated.
Mini-PCI is an open standard, just like PCI. You even buy the specifications from the same place, the PCI SIG. What nVidia is doing is pre-empting what the PCI SIG will eventually come out with, perhaps in the hopes that the PCI SIG will adopt their standard as the official PCI graphics standard for laptops.
If a couple of big players like Dell and Toshiba adopt it for their notebooks, this will most likely become the standard.
Someone else on this thread added that CardBus does exactly what you are asking for. I would also like to add that audio capabilities should be absorbed into the video card anyway so that you can send it a single datastream and get both video and audio out, so you don't need to add audio capabilities to the bus.
What you're looking for is the median, which is sort of a mathematic definition for "most". Averages are skewed by very large or very small numbers.
The real reason why mileage is lower on the highway is because air resistance is proportional to the square of the velocity, and since highway speeds are usually twice as high, you have to fight four times the air resistance.
This also means that the Prius is one amazing car... close to 100% of the energy that used to be wasted braking (typical of city driving) is now recovered.
I think that the party lines no longer break up between true conservative and liberal any more. For example, I think not going to war is the true conservative thing to do... and strangely enough, on that issue, Howard Dean ends up looking like the most conservative of any of the candidates who were running for President.
Also, a more recent interview.
Why is it that all the greats are destined to be screwed over? Here is a summary of what happened to Gygax and TSR. Such a shame.
What about linux on a coaster?