I'll second that is's a good electronics book. The one place it's lacking is it's coverage of the newer GaAs and InP transistors (mostly MESFETS used in communications). Then again I don't know of a book that DOES cover that stuff well....
Also if you need a general EE reference (circuits, filtering, fields, compatability, etc) I'd go with Kaiser's Electromagnetic Compatibility Handbook. It's not released yet but I have a pre-press edition for review and it is THE most thorough book on basic EE that I have ever seen.
It can't be targeted unless the vendor has both an unsecure BGP protocal and a crappy TCP implementaion.
This isn't a TCP problem, it's just being billed that way because a bunch of vendors have crappy implementations of the above protocals. Yes, in theory this could affect everyone, but the difficulty of doing this type of attack on a system with a good TCP implementation is next to nothing.
Basically, the attack takes advantage of certain predictable behaviors that arn't in the spec but that most of the TCP implementations have to reduce the possible space of packets to something that is reasonably tryable.
If he doesn't give you a license, you are stuck with the default rights under Copyright Law. That's DJB's "license".
As such you can't make changes and distribute them as part of a integrated set (in theory you could distribute just the patches but it's a grey area). Hence many people consider it "un-free". Some people see this as a problem.
Perhaps YOUR dealer doesn't bargin. Saturn has no such company policy. The price of the vehicle is entirely at the dealer's discresion. I personally never do business with someone that won't haggle with me, but then again I'm a stingy SoB.
Furthermore GM PAYS the dealer to keep the vehicles in stock (about 3% of the base MSRP). So if you buy the car as soon as it hits the lot, the dealer pockes the extra cash (because he's not using it to pay for intrest on the loan he took out to buy his stock).
I'd say that as a rule of thumb, if you are paying cash for the vehicle, you should get at least 7% off. I've always managed closer to 10% though.
Also now that you have a car you should keep it for as long as possible. Too many people fall prey to the "it costs more to fix it then the vehicle is worth" mentality and go finance a new car. The cost of financing a new car is much higher then the cost of repairs plus setting aside for a new vehicle in most cases (when figuring this out don't forget to factor in the extra money you make on intrest when you are saving)
except that each district has to have exactly the same number of people and be geographically contigous, and be "reasonably shaped". In other works you can't do what you are suggesting.
While the representatives themselves are no longer needed, the number of representatives awarded to each state was based on a delicate compromise that does not favor states with a large population or states with a small one. A popular vote or any other "new" voting system will disturb that balance one way or the other.
The initial idea was to elect on person for each congressional district and two "at-large" electors for the entire state. Then some states got the "brilliant" idea that if they made it so that whoever won a popular majority in the state got ALL of the electoral votes from that state (as opposed to the two at-large votes and which ever districts they won), candidates would spend more time in that state then in others (because a victory there means more). To compeate other states did the same thing; the result has been a stark reduction of the number of votes that actually matter (you only need to campain in a few key spots in a few key states to win).
The solution is obvious: fix the abuse of the electoral college system. We don't need to swap to any compilcated voting scheme or go to outright popular voting (as both of those would have unintended consequences as well). All we need to do is require electors to be assigned on a district-by-district basis. Then each district is equally valuable. The two at-large electors can be chosen based on who won the popular vote state wide.
Note that this will NEVER happen. It is however the only solution that uses the intended system without adjusting the compromise made by the founders.
And yes, we can get rid of the electors themselves if people really want. I tend look at it more as a historic tradition then an unneeded complication though.
Maybe I wasn't clear. You can install PHP on your personal machine and use make files to run it on the.php files giving you.html files that have the content generated at make time instead of at download time. That sounds like all the lbi files do anyway.
One thing I forgot: if you have loans, morgages, or credit card dept that carry intest, you want to pay that off ASAP before investing. (Obviously if you can get 7% per year investing and the intrest on your credit card is 15% it makes sense to pay it off first; but some people don't see this....)
What's listed above is great advise; I'll add one more suggestion: In addition to "savings", invest a portion of your income for your retirement NOW. $100 a month every month from here on out will let you be better off then someone who sets aside 10x that much for the 10 years just before they retire (assuming the stock market's performance of the past 60 years is reflective of the next 60).
How to invest(I'm not responsible for this advise, know your risks, etc.):
Option 1: Buy mutual funds; I'd stick to funds with low management overheads like spiders, and S&P 500 and Wilkshire 7000 tracking funds. These do better then 97% of managed funds anyway and the risk and thought are low, just put some portion of your income into the market every month. Possibly put some of this money in bonds as a way of reducing risk.
Option 2: Get a professional to manage your money. This involves learning a little bit and looking into the track record of the guys involved. There's also more cost involved. Typically I wouldn't expect them to be able to out perform the S&P but some do, and your portfolio is taylored to you.
Option 3: Learn to do it yourself. It's not as easy as above, but I wouldn't call it difficult; if that's what you want I'd start with Fisher(Common Stocks, Uncommon Profits) and Graham(The Intelligent Investor). Both of these were written with the non-professional in mind.
We have gotten WAY WAY off topic....
>there are good reasons Crays, SunFires, and mere (and affordable) dual Intels and AMDs exist
I'll agree that those systems exist for a reason, but there are cost with supporting them; and most systems sold are still UP.
There is only so much developer time in the day. There are other OSes that support SMP. Having OpenBSD support SMP, while it would be nice, isn't top priority because most of the developers and users don't need it and don't care. If it's that important to you, spend some of your money and pay one of the developers to work on it, or at least donate some hardware to inspire them.
I'd rather see more stuff like pf, BGP, CARP, W^X, and propolice then SMP.
>I doubt, this is your claim, BTW
Others have made it before; I didn't originate it; but in this context it was my base claim.
>SMP is a perfectly legitimate direction of scaling
I'll agree, but many of the people who use it just think they need it. Typically programs are more bound by memory bandwidth, network bandwidth, and disk speed.
As for the "hack" comment, SMP was hacked on to the i386 line of processors to squeeze extra performance out, it wasn't designed in like it was for other systems. Newer buses, storage, etc. all were added on as extensions, none of them required what basically amounts to a re-write of the entire OS, and a reworking of the underlying hardware. (although if you sum up all of those changes, then I guess the OS did almost get re-writen and the hardware changed compleately....)
> If you truly want OpenBSD to succeed as an operating system on a grand scale,
I could care less and so could the developers. I hate to sound harsh but: "It works for me!" That's all I really care about. Whether or not you or anyone else uses the same OS as I do, doesn't really matter. I suggest you use what works and does the job need it to.
>those shortcomings that are obviously bad
How is not having SMP "obviously bad"? SMP complicates the kernel code path and introduces entire catagories of bugs, and exploits that just weren't there before. All of this, just to support a few special cases where you do actually need more CPU power on the same box (as opposed to more RAM, faster disks, a better designed program, etc).
Think about what you are saying for a minute:
"I like how secure and stable and well documented this operating system is. It's so great! But I want to have these features that this other not so secure, stable, and well documented system has!"
OpenBSD isn't feature focused, it's focused on security, stability, and open licensing. The reason OpenBSD has the security and stability it does is because the developers focused on that instead of adding more buggy features. Furthermore when they do add features, they add stuff that helps them: e.g. IP source tracking in the firewall, BGP protocal, redundancy and failover support, etc. In my mind having this sort of stuff is far more important then SMP.
>You need to understand this...
No, I need to have a system that works and doesn't cause me any hassle. If OpenBSD isn't the tool for your job; use a different tool. You arn't going to hurt anyone's feelings.
>Until which time, the OpenBSD zealots will continue to deny the issue exists or is of any importance. I see...
I am not an "OpenBSD zelot"; I use OpenBSD on a few servers and hence am subscribed to the misc@ list. This was recently discussed so I thought I'd fill the poster in.
All of the software you mention, while important, isn't what's used on a majority of systems (and more importantly isn't what the developers use their systems for). Many many more systems operate web servers, routers, file servers, and mail servers: general work horse stuff. That's what OpenBSD likes to focus on (because that's what the developers use it for). Hence SMP isn't a huge priority. Most of the developers don't work on OpenBSD for a living; they just code to solve there own problems (which by necessity are in the domain of what they use the software for). If this happens to solve your problem too that's great; if it doesn't well you do have the code.
Furthermore, all of the software you mentioned is corner case, no where near common. How many bank computers and render farms are there compared to the number of web servers, mail servers, and routers?
As for databases, most databases are more in need of RAM and fast disks then CPU. Those that are CPU hogs are often poorly designed. For those that actually do need multiple CPUs, let me remind you that there is no such thing as an all-in-one-solves-every-problem-every-time tool: you proably don't want to use OpenBSD for that anyway.
None of this changes my claim that SMP is a (useful) hack to squeeze more performance out of today's technology for those who just can't wait (and are willing to pay through the nose).
It really only affects poor implementations of TCP (like Cisco's).
Any implementation that sets a reasonable window size and uses randomized sequence numbers is still near impossible to attack. Such correct implementations include Linux, and the various BSDs.
You are correct; OpenBSD doesn't support SMP right now. This is mostly because SMP is a way to get next year's computing power today (and paying a lot of money in the process). Combine that with the fact that almost all of the programs used on a modern UNIX-like system arn't CPU bound and it's easy to see why SMP took a back seat to more "interesting" issues.
However there is a developer being paid to work full time writing SMP support for OpenBSD. He expects to have a working implementation by 3.6 or 3.7.
The motor has 20% of the LOSS of a typical motor. This doesn't make it anywhere near 80% more efficient. (blame the slash-bots for that one)
Also he's measuring power with a simple multi-meter instead of a real watt-meter, hence failing to take into account the power factor (i.e. P=V*I*pf). That's why he's getting more out then he's putting in.
Like I said, see my comment
further down. This is sound like a reluctance motor (hence the 20% less loss) that is either controlable or gets torque at non-syncronus speed. In other words, for automotive uses, it's probably limited to HVAC stuff. Realistically though, syncronous motors are the most common, so he's got a big market.
Furthermore this motor will only sell well if the power factor is reasonable (as the advantage of syncronous motors is that they can be run at either lagging or leading, thus eliminating the need for extra capacitors). This likely doesn't have an adjustable pf, and will need a capacitor bank to make pf unity. So the extra efficiency has to be enough to offset the cost of buying high voltage caps.
Based off some back of the napkin calculations, the energy savings don't offset the cost here in my part of the States, but in places where power is more expensive, it probably would.
Read the fine print at the bottom.
The results are based on the NUMBER of hotspots, the NUMBER of students, and the NUMBER of computers. This biases the study against smaller schools.
For example Kettering University, a small engineering school (interesting, we graduate more engineers than any other University, and 1 out of 5 graduates becomes a business owner or fortune 500 exec, but that's a side point) which has great, almost 100% coverage (I live two blocks away and get signal) isn't on the list because it's are so small (there are 5 buildings on campus).
He's saying the GOVERNMENT (i.e. the tax office) won't be paying the tax. After all, it's silly to charge yourself for something. And in most places government agencies are exempt from most taxes.
Motors are efficient enough (75-80% is typical for variable speed applications). Most of the losses involved are in storing and supplying the power. If you reduce input power to the motor by 80%, you just ended up with better then 100% efficiency, something that can't happen. See my post further down for my extrapolation of what this motor could actually be on a really good day
To make a viable, sellable purely electric car, you need batteries that weight AT LEAST 1/4 as much and cost 1/10th as much as today's best.
While we're at it, inventing power transistors that are nearly lossless would be nice, since switching losses are the other major diffuculty.
Battery-powered cars arn't realistic because of power density issues with modern bateries.You can already make some damn efficient motors. If put a fully electric power train together, you get too much weight and cost to realistically market the vehicle. You can BUILD a working electric car with today's technology (and it will work well); it'll make a porsche look cheep though.....
Also industrial consumption won't go down much, there's ton's on manufacturing processes and techniques that we can't do because they take too much power, this just opens up the ability to use them.
All of this of course assumes that the inventor actually made something and isn't scamming people. While I'd like to think that someone finally made a realistically-working reluctance motor, this article reeks of violating the 2nd law of thermodynamics.
The article is fubar'ed beyond belief. Ignoring for a minute the obsurd thermodynamics claims, based off of the description of the motor, it seems like he's invented a very controlable reluctance motor.
Asside: For those who arn't EEs you can use magents to spin things in various ways: induction, rotating fields generated by coils, reluctance, etc. Reluctance motors are 10-20% more efficient they their syncronous counterparts, but tend to be limited in size, hard to manufacture, and difficult to control. A lot of research has gone in to the different ways to make the magnetic stator to make the motor easier to make, control, and scale up.
At best he's invented a particular rotor/stator combination that creates a really odd magnetic field that he can actually control. My guess is that the motor he has made runs syncronous after spinning up and that his particualar arangement of magents makes it possible for the motor to get enough torque to spin up at non-syncronous speed (i.e. start when you plug it in, and possibly give it a spin).
IF this does work, IF he can get the reliability to the level of syncronous motors, IF it runs at a reasonable power factor, IF its reasonably EMC, AND IF it doesn't require complicated or expensive control mechanisms, he will have a good product on his hands. This would likely be used in a lot of factories, and in HVAC systems in cars. It's probably not that useful for speed control based applications (if it's a reluctance based motor, it's running at syncrous speed) so that excludes it from replacing induction motors and DC motors, unless it's so much more efficient that adding a variable AC supply to the control equipment leaves it still more efficient.
Honestly though, I think the countless posts here are probably right: he invented something and only THINKS it works.
Slashdot Mirror
Also if you need a general EE reference (circuits, filtering, fields, compatability, etc) I'd go with Kaiser's Electromagnetic Compatibility Handbook. It's not released yet but I have a pre-press edition for review and it is THE most thorough book on basic EE that I have ever seen.
This isn't a TCP problem, it's just being billed that way because a bunch of vendors have crappy implementations of the above protocals. Yes, in theory this could affect everyone, but the difficulty of doing this type of attack on a system with a good TCP implementation is next to nothing.
Basically, the attack takes advantage of certain predictable behaviors that arn't in the spec but that most of the TCP implementations have to reduce the possible space of packets to something that is reasonably tryable.
As such you can't make changes and distribute them as part of a integrated set (in theory you could distribute just the patches but it's a grey area). Hence many people consider it "un-free". Some people see this as a problem.
Furthermore GM PAYS the dealer to keep the vehicles in stock (about 3% of the base MSRP). So if you buy the car as soon as it hits the lot, the dealer pockes the extra cash (because he's not using it to pay for intrest on the loan he took out to buy his stock).
I'd say that as a rule of thumb, if you are paying cash for the vehicle, you should get at least 7% off. I've always managed closer to 10% though.
Also now that you have a car you should keep it for as long as possible. Too many people fall prey to the "it costs more to fix it then the vehicle is worth" mentality and go finance a new car. The cost of financing a new car is much higher then the cost of repairs plus setting aside for a new vehicle in most cases (when figuring this out don't forget to factor in the extra money you make on intrest when you are saving)
except that each district has to have exactly the same number of people and be geographically contigous, and be "reasonably shaped". In other works you can't do what you are suggesting.
The initial idea was to elect on person for each congressional district and two "at-large" electors for the entire state. Then some states got the "brilliant" idea that if they made it so that whoever won a popular majority in the state got ALL of the electoral votes from that state (as opposed to the two at-large votes and which ever districts they won), candidates would spend more time in that state then in others (because a victory there means more). To compeate other states did the same thing; the result has been a stark reduction of the number of votes that actually matter (you only need to campain in a few key spots in a few key states to win).
The solution is obvious: fix the abuse of the electoral college system. We don't need to swap to any compilcated voting scheme or go to outright popular voting (as both of those would have unintended consequences as well). All we need to do is require electors to be assigned on a district-by-district basis. Then each district is equally valuable. The two at-large electors can be chosen based on who won the popular vote state wide.
Note that this will NEVER happen. It is however the only solution that uses the intended system without adjusting the compromise made by the founders.
And yes, we can get rid of the electors themselves if people really want. I tend look at it more as a historic tradition then an unneeded complication though.
One thing I forgot: if you have loans, morgages, or credit card dept that carry intest, you want to pay that off ASAP before investing. (Obviously if you can get 7% per year investing and the intrest on your credit card is 15% it makes sense to pay it off first; but some people don't see this....)
The PHP has include capabilities and other processing ability. The make files should be able to run PHP and generate the resultant pages.
How to invest(I'm not responsible for this advise, know your risks, etc.):
Option 1: Buy mutual funds; I'd stick to funds with low management overheads like spiders, and S&P 500 and Wilkshire 7000 tracking funds. These do better then 97% of managed funds anyway and the risk and thought are low, just put some portion of your income into the market every month. Possibly put some of this money in bonds as a way of reducing risk.
Option 2: Get a professional to manage your money. This involves learning a little bit and looking into the track record of the guys involved. There's also more cost involved. Typically I wouldn't expect them to be able to out perform the S&P but some do, and your portfolio is taylored to you.
Option 3: Learn to do it yourself. It's not as easy as above, but I wouldn't call it difficult; if that's what you want I'd start with Fisher(Common Stocks, Uncommon Profits) and Graham(The Intelligent Investor). Both of these were written with the non-professional in mind.
Good luck!
I'll agree that those systems exist for a reason, but there are cost with supporting them; and most systems sold are still UP.
There is only so much developer time in the day. There are other OSes that support SMP. Having OpenBSD support SMP, while it would be nice, isn't top priority because most of the developers and users don't need it and don't care. If it's that important to you, spend some of your money and pay one of the developers to work on it, or at least donate some hardware to inspire them.
I'd rather see more stuff like pf, BGP, CARP, W^X, and propolice then SMP.
>I doubt, this is your claim, BTW
Others have made it before; I didn't originate it; but in this context it was my base claim.
>SMP is a perfectly legitimate direction of scaling
I'll agree, but many of the people who use it just think they need it. Typically programs are more bound by memory bandwidth, network bandwidth, and disk speed.
As for the "hack" comment, SMP was hacked on to the i386 line of processors to squeeze extra performance out, it wasn't designed in like it was for other systems. Newer buses, storage, etc. all were added on as extensions, none of them required what basically amounts to a re-write of the entire OS, and a reworking of the underlying hardware. (although if you sum up all of those changes, then I guess the OS did almost get re-writen and the hardware changed compleately....)
BTW, what's with the SIG?
> If you truly want OpenBSD to succeed as an operating system on a grand scale,
I could care less and so could the developers. I hate to sound harsh but: "It works for me!" That's all I really care about. Whether or not you or anyone else uses the same OS as I do, doesn't really matter. I suggest you use what works and does the job need it to.
>those shortcomings that are obviously bad
How is not having SMP "obviously bad"? SMP complicates the kernel code path and introduces entire catagories of bugs, and exploits that just weren't there before. All of this, just to support a few special cases where you do actually need more CPU power on the same box (as opposed to more RAM, faster disks, a better designed program, etc).
Think about what you are saying for a minute: "I like how secure and stable and well documented this operating system is. It's so great! But I want to have these features that this other not so secure, stable, and well documented system has!"
OpenBSD isn't feature focused, it's focused on security, stability, and open licensing. The reason OpenBSD has the security and stability it does is because the developers focused on that instead of adding more buggy features. Furthermore when they do add features, they add stuff that helps them: e.g. IP source tracking in the firewall, BGP protocal, redundancy and failover support, etc. In my mind having this sort of stuff is far more important then SMP.
>You need to understand this...
No, I need to have a system that works and doesn't cause me any hassle. If OpenBSD isn't the tool for your job; use a different tool. You arn't going to hurt anyone's feelings.
I am not an "OpenBSD zelot"; I use OpenBSD on a few servers and hence am subscribed to the misc@ list. This was recently discussed so I thought I'd fill the poster in.
All of the software you mention, while important, isn't what's used on a majority of systems (and more importantly isn't what the developers use their systems for). Many many more systems operate web servers, routers, file servers, and mail servers: general work horse stuff. That's what OpenBSD likes to focus on (because that's what the developers use it for). Hence SMP isn't a huge priority. Most of the developers don't work on OpenBSD for a living; they just code to solve there own problems (which by necessity are in the domain of what they use the software for). If this happens to solve your problem too that's great; if it doesn't well you do have the code.
Furthermore, all of the software you mentioned is corner case, no where near common. How many bank computers and render farms are there compared to the number of web servers, mail servers, and routers?
As for databases, most databases are more in need of RAM and fast disks then CPU. Those that are CPU hogs are often poorly designed. For those that actually do need multiple CPUs, let me remind you that there is no such thing as an all-in-one-solves-every-problem-every-time tool: you proably don't want to use OpenBSD for that anyway.
None of this changes my claim that SMP is a (useful) hack to squeeze more performance out of today's technology for those who just can't wait (and are willing to pay through the nose).
OpenBSD works great for me as a router, webserver, and mail server. Admittedly, I'm not running a huge corporate network here.
If it doesn't work for you, either use what does, or change it so it works. OpenBSD is just a tool, you should use the right one for the job.
Any implementation that sets a reasonable window size and uses randomized sequence numbers is still near impossible to attack. Such correct implementations include Linux, and the various BSDs.
This "TCP exploit" is less a function of the protocal and more a function of poor implementations there of.
I'm pretty sure that the other BSDs and Linux arn't affected either.
However there is a developer being paid to work full time writing SMP support for OpenBSD. He expects to have a working implementation by 3.6 or 3.7.
The motor has 20% of the LOSS of a typical motor. This doesn't make it anywhere near 80% more efficient. (blame the slash-bots for that one)
Also he's measuring power with a simple multi-meter instead of a real watt-meter, hence failing to take into account the power factor (i.e. P=V*I*pf). That's why he's getting more out then he's putting in.
Like I said, see my comment further down. This is sound like a reluctance motor (hence the 20% less loss) that is either controlable or gets torque at non-syncronus speed. In other words, for automotive uses, it's probably limited to HVAC stuff. Realistically though, syncronous motors are the most common, so he's got a big market.
Furthermore this motor will only sell well if the power factor is reasonable (as the advantage of syncronous motors is that they can be run at either lagging or leading, thus eliminating the need for extra capacitors). This likely doesn't have an adjustable pf, and will need a capacitor bank to make pf unity. So the extra efficiency has to be enough to offset the cost of buying high voltage caps.
Based off some back of the napkin calculations, the energy savings don't offset the cost here in my part of the States, but in places where power is more expensive, it probably would.
One word: COST
For example Kettering University, a small engineering school (interesting, we graduate more engineers than any other University, and 1 out of 5 graduates becomes a business owner or fortune 500 exec, but that's a side point) which has great, almost 100% coverage (I live two blocks away and get signal) isn't on the list because it's are so small (there are 5 buildings on campus).
He's saying the GOVERNMENT (i.e. the tax office) won't be paying the tax. After all, it's silly to charge yourself for something. And in most places government agencies are exempt from most taxes.
To make a viable, sellable purely electric car, you need batteries that weight AT LEAST 1/4 as much and cost 1/10th as much as today's best.
While we're at it, inventing power transistors that are nearly lossless would be nice, since switching losses are the other major diffuculty.
Battery-powered cars arn't realistic because of power density issues with modern bateries.You can already make some damn efficient motors. If put a fully electric power train together, you get too much weight and cost to realistically market the vehicle. You can BUILD a working electric car with today's technology (and it will work well); it'll make a porsche look cheep though.....
Also industrial consumption won't go down much, there's ton's on manufacturing processes and techniques that we can't do because they take too much power, this just opens up the ability to use them.
All of this of course assumes that the inventor actually made something and isn't scamming people. While I'd like to think that someone finally made a realistically-working reluctance motor, this article reeks of violating the 2nd law of thermodynamics.
Asside: For those who arn't EEs you can use magents to spin things in various ways: induction, rotating fields generated by coils, reluctance, etc. Reluctance motors are 10-20% more efficient they their syncronous counterparts, but tend to be limited in size, hard to manufacture, and difficult to control. A lot of research has gone in to the different ways to make the magnetic stator to make the motor easier to make, control, and scale up.
At best he's invented a particular rotor/stator combination that creates a really odd magnetic field that he can actually control. My guess is that the motor he has made runs syncronous after spinning up and that his particualar arangement of magents makes it possible for the motor to get enough torque to spin up at non-syncronous speed (i.e. start when you plug it in, and possibly give it a spin).
IF this does work, IF he can get the reliability to the level of syncronous motors, IF it runs at a reasonable power factor, IF its reasonably EMC, AND IF it doesn't require complicated or expensive control mechanisms, he will have a good product on his hands. This would likely be used in a lot of factories, and in HVAC systems in cars. It's probably not that useful for speed control based applications (if it's a reluctance based motor, it's running at syncrous speed) so that excludes it from replacing induction motors and DC motors, unless it's so much more efficient that adding a variable AC supply to the control equipment leaves it still more efficient.
Honestly though, I think the countless posts here are probably right: he invented something and only THINKS it works.
copywrite doesn't go "poof" if you don't enforce it. You are thinking of trademarks.