Sure, call me in two years from now, when they come out with $5 laser-etched holographic 3d memory cubes, which store an unlimited amount of data in a space the size of a few cubic inches...
If he does call you, I think he'll be laughing. I can remember hearing about 3D holographic memory being the "wave of the future" something like 8 or 10 years ago. Wake me up if it ever actually shows up.
About a year ago, I built an office file server and a second server for point-in-time snapshots. The two use some relatively cheap CodeGen cases that had 8 5.25" bays in which to mount tons of drives. The file server got 4x200 gig drives in RAID 5, the backup machine got 9x300 gig drives in RAID 5.
The backup machine keeps a regular snapshot of the data for a relatively quick worst-case cold-standby, by using rsync once per night - and after rsync is done, the resultant snapshot gets tar/gzipped and rotated.
By your arguments, I'd suggest just buying cheap ones and letting them break. Buy a second as backup, if you want. At $300 each, it'll take a looong time to break even on that GL1. In fact, as time goes on, price comes down, and features go up - a digital camcorder that fits the basic requirements will probably be in the $150 range in a couple of years. At an attrition rate of one per year, it could take a decade before he'd have layed out what a GL1 would cost.
Maybe you don't understand: They were benchmarking a CPU. Not a graphics card, a CPU. The more they turn up the resolution and detail, the more the video card will be a factor, and mask the benefits of the CPU. Even if they used the same video card, as the card becomes more of a limitting factor, the more all of the CPUs will look the same.
Now, that's not to say that it wouldn't have been interesting to have some 1600x1200 benchmarks, but in and of itself, the choice of 640x480 is not a bad one.
Where I live, you get charged a disposal fee if you buy a tire. If they don't actually dispose of an old tire, too bad, they're required by law to charge you anyway. Then if *YOU* take that tire in to be recycled, then you get charged a recycling fee on top of it. It's similar with other products like batteries.
If it weren't for idiocies like that, then I'd be much more open to these sorts of fees.
There are tons of different apps and threads competing all of the time - you have network code, file system code, network drivers, GUI code, messaging code, disk drivers, video drivers, DirectX, blah, blah, blah.
Shoot, NT even has two hidden threads sitting around doing nothing but watching whether two registry entries get changed - the entries that turn Workstation into Server.
If you look at the hidden apps and threads, there are a LOT of them. But here's the catch: All in all, they tend to use relatively little compared to the programs you run, and so going to dual processers means very little (if any) speedup in most normal apps. Unless you know that your app will benefit, don't count on a speedup.
On the other hand, you can gain quite a bit of responsiveness under heavy load with dual processers, but there are plenty of other places (such as slow/cheap disks or controllers) that will kill your responsiveness as well.
Even my daul-CPU rigs with GF6600 cards get along just fine with smaller CPUs. It's very entertaining to watch people put these things in single-proc computers with a single video card.
Why, did cars stop working? Did elevators break and drop to the ground? Is that sewage somehow worse than everything else that gets dumped into the rivers in Russia? Biological waste can (and will) be broken down very quickly by the life in the river, but all of the heavy metals and other pollutants that get dumpes aren't the same...
Because once the computer is infected, that can easily be disabled. There are many worms/virii/spyware/etc. that already bypass, disable, or even delete various scanners and firewalls.
It's easy to think that the firewall will protect them from being infected, but it won't protect them from things that they tell the computer to retrieve - if they go to a web page with the IE exploit du jour, then the firewall doesn't do a whole lot of good.
By exponentially dropping the available bandwidth to that port as traffic on it
That would mean that a single large attachment would penalize other mail traffic. A better way would be to throttle the numer of new connections which can be established.
With port 25 blocked, zombie owners are forced to use the ISP's outgoing mail servers. If throttling is intelligently applied to all port 25 traffic on a per-host basis, the feasibility of zombie spamming drops off.
No, it just means that they need to infect more machines. It's easier for them to get around limitting and throttling if they have a large number of infected hosts than if they have to pass it through a single or few servers.
The OpenBSD team is working on a transparent traffic shaping proxy that will make magic like this trivial for the pf priesthood
While it isn't exponential, you can still limit the connection rate with iptables or tcplimit without much trouble. I'm not a big BSD guy, but I would expect there would already be provision to do something similar there...
The catch is that then email sits on their lame SMTP server for x number of hours if it doesn't go out,
That sounds like it's working as designed. Depending on the error involved, mail servers are designed to either bounce it immediately, or queue it and keep trying.
If your mail server bounces email for every single transient error, you're going to bounce a very significant portion of your mail that really shouldn't be bounced.
The real solution is to filter spam using something like Spamassassin
No, the real solution is for every country to approve the death penalty for spammers, and allow extradition to any country with a warrant for them. Most anything less, and spam will continue. With filtering, they just increase the number they send by a few orders of magnitude, and they still get enough by to make a profit. Since they aren't paying for the vast majority of the resources used, then it doesn't matter to them if 99.9% of the mail gets filtered, they'll just send 1000 times more.
However, for the tinfoil hat crowd and hardcore geek types, this can be a problem.
I'm not so sure that it is. Most of the folks that I know who are running a mail server on their DSL or cable line really aren't all that qualified to be running a mail server - despite the fact that they always, always, always think that they are.
Yes, I've run across a few that are qualified, but they're vastly in the minority. As much as I'd love to run my own mail server at home, I just don't do it. Enough of my mail would be blocked anyway just because it's coming from a cable network.
As to blocking mail coming from a cable network, I don't blame people, either. I do it all the time on my mail servers. If I block mail coming from cable modems (NOT the cable company's mail servers, but the modem pools themselves), I block at least 20,000,000 pieces of spam for every one legitimate message that I lose. No, those numbers are not exagerations.
You're confusing constitutionally-guaranteed rights with priveliges.
You have the right to be free from unreasonable search and seizure.
You do NOT have the right to do whatever you want with my private property. My mail server is my private property.
If you argue that, then I should be free to walk over to your house and use your car whenever you're not, because as long as I fill it up with gas, no damage has been done to you.
There's your mistake. While those sorts of runtimes get claimed pretty often, I don't know a single person with a PC-based laptop that gets that sort of runtime - nearly all say that they get two hours or less.
Also, if you're splitting hairs here, going from watt-hours to watts and hours doesn't work out exactly in most situations, as the watt-hour rating is measured under the most ideal current draw and temperature, which you won't often achieve - and that's *IF* they were actually honest about the watt-hour rating, the batteries that actually measure up to the manufacturer's claims are very few and far between.
Are you sure that it's 1/3 of the power budget? Western Digital's 2.5" drives only use 2.5 watts under read/write activity, and 2 watts idle - and they're nothing special.
So, for your claim to be true, then the CPU, northbridge, southbridge, memory, display, and everything else would have to consume no more than 4 watts...
The CF+ and Compact Flash specification 3.0 includes UDMA 33 and UDMA 66 support. I've seen references to certain cards and CF->IDE adapters that support DMA, so that problem is partially solved, and will get better.
As for the problem of sustained speeds, there's always RAID 0...
If you have 140 boys and 100 girls, then the percentage of boys is 58.3%, just 8.3% above the expected 50%.
If I were a more sarcastic person, I'd question whether a comment on a statistical phenominon could be trusted if the author didn't grasp simple percentage problems... but I wouldn't want to hurt any feelings. =)
It is true that the sperm determine the gender. However, your assumption that geeks will cause being female to become a recessive trait is a bit flawed, because (a) geeks are a relatively small portion of the population, and (b) it could be argued that their contribution to the next generation is relatively small, as well.
If anything, I would argue that persons of lower intelligence are more likely to breed early and breed often, and that intelligence itself is slowly becoming a much less common trait...
You're assuming that the first sperm to the egg causes conception, when that's not true. It takes a very good number just to wear down and break through the protective coating on the egg, and *then* one lucky little swimmer becomes the star, so to speak. You could also theorize that the faster Y-based sperm got their first, broke down the barrier, and a somewhat-late-coming X-based sperm was the "winner". There is a lot that can happen. : )
The research in the article seems to be nothing but statistical. As with any but the most simple biological processes, there are hundreds or thousands (or TENS of thousands) of intermediate steps along the way, a change in any of which can have large effects.
One of the pathways I once studied had to do with Vul/Muv defects in flatworms - either the lack of a vulva in a female, or the presence of many vulvas. The developmental process involved a very good number of genes, each coding for proteins that had to interact with their surroundings in a certain way, over a certain chain of events in order to trigger the production of a vulva. An anomoly at almost any one of these stages could produce one of the two mentioned defects, and if I recall, certain anomolies could mask previous anomolies - it wasn't as simple as "Vul defect, must have a defective {XXX} gene." It was more like "Vul defect, could be any of twenty different defective genes or any of a complete set of combinations."
You're right, it's the presence of a Y chromosome that ends up causing an embryo to become male - without it, the embryo becomes (by default) a girl. There are genetic conditions whereby someone with a Y chromosome can end up becoming a girl anyway, but that's a lecture for another day.
To get back on track, the first guess at an explanation would be that a man's "gender strength" (for lack of a better term) would not only determine what his personality turned out to be (and hence, what job he is in), but also what percentage of his sperm that reach the egg would have a Y chromosome as opposed to an X.
Off of the top of my head, I would imagine that a significant difference in the X:Y ratio in sperm would have been noticed before now - it could have to do with other factors, such as their Y-sperm having a greater ability to dissolve the protective barrier around the egg, being a little faster, or many other things - conception is an awfully complicated business, and there are plenty of small, subtle changes that can lead to relatively large differences in outcome.
Slashdot Mirror
Sure, call me in two years from now, when they come out with $5 laser-etched holographic 3d memory cubes, which store an unlimited amount of data in a space the size of a few cubic inches...
If he does call you, I think he'll be laughing. I can remember hearing about 3D holographic memory being the "wave of the future" something like 8 or 10 years ago. Wake me up if it ever actually shows up.
steve
About a year ago, I built an office file server and a second server for point-in-time snapshots. The two use some relatively cheap CodeGen cases that had 8 5.25" bays in which to mount tons of drives. The file server got 4x200 gig drives in RAID 5, the backup machine got 9x300 gig drives in RAID 5.
The backup machine keeps a regular snapshot of the data for a relatively quick worst-case cold-standby, by using rsync once per night - and after rsync is done, the resultant snapshot gets tar/gzipped and rotated.
steve
By your arguments, I'd suggest just buying cheap ones and letting them break. Buy a second as backup, if you want. At $300 each, it'll take a looong time to break even on that GL1. In fact, as time goes on, price comes down, and features go up - a digital camcorder that fits the basic requirements will probably be in the $150 range in a couple of years. At an attrition rate of one per year, it could take a decade before he'd have layed out what a GL1 would cost.
steve
Maybe you don't understand: They were benchmarking a CPU. Not a graphics card, a CPU. The more they turn up the resolution and detail, the more the video card will be a factor, and mask the benefits of the CPU. Even if they used the same video card, as the card becomes more of a limitting factor, the more all of the CPUs will look the same.
Now, that's not to say that it wouldn't have been interesting to have some 1600x1200 benchmarks, but in and of itself, the choice of 640x480 is not a bad one.
steve
It's a long-standing practice from Apple, just like the boot ROM that you had to have on a motherboard to get MacOS to run.
Apple knows that people will pay a premium for their hardware because (A) it looks fancy, and (B) it's the only way to get their OS.
So, just because they're switching chips is no reason for them to stop locking in the hardware market, too...
I'm actually saddened that they're switching - not because I like Macs, but because it means less overall competition in the CPU market.
steve
How many does he get right - something like 1 out of 100? With as many as he makes, he's bound to get at least a few right.
steve
"Unfortunately there are still some (...) Jews around who have problems with the idea of Jews being armed"
Certainly not these jews.
steve
Where I live, you get charged a disposal fee if you buy a tire. If they don't actually dispose of an old tire, too bad, they're required by law to charge you anyway. Then if *YOU* take that tire in to be recycled, then you get charged a recycling fee on top of it. It's similar with other products like batteries.
If it weren't for idiocies like that, then I'd be much more open to these sorts of fees.
steve
There are tons of different apps and threads competing all of the time - you have network code, file system code, network drivers, GUI code, messaging code, disk drivers, video drivers, DirectX, blah, blah, blah.
Shoot, NT even has two hidden threads sitting around doing nothing but watching whether two registry entries get changed - the entries that turn Workstation into Server.
If you look at the hidden apps and threads, there are a LOT of them. But here's the catch: All in all, they tend to use relatively little compared to the programs you run, and so going to dual processers means very little (if any) speedup in most normal apps. Unless you know that your app will benefit, don't count on a speedup.
On the other hand, you can gain quite a bit of responsiveness under heavy load with dual processers, but there are plenty of other places (such as slow/cheap disks or controllers) that will kill your responsiveness as well.
steve
Even my daul-CPU rigs with GF6600 cards get along just fine with smaller CPUs.
That should be:
Even my daul-CPU rigs with GF6600 cards get along just fine with smaller PSUs .
Even my daul-CPU rigs with GF6600 cards get along just fine with smaller CPUs. It's very entertaining to watch people put these things in single-proc computers with a single video card.
steve
Why, did cars stop working? Did elevators break and drop to the ground? Is that sewage somehow worse than everything else that gets dumped into the rivers in Russia? Biological waste can (and will) be broken down very quickly by the life in the river, but all of the heavy metals and other pollutants that get dumpes aren't the same...
Because once the computer is infected, that can easily be disabled. There are many worms/virii/spyware/etc. that already bypass, disable, or even delete various scanners and firewalls.
It's easy to think that the firewall will protect them from being infected, but it won't protect them from things that they tell the computer to retrieve - if they go to a web page with the IE exploit du jour, then the firewall doesn't do a whole lot of good.
steve
By exponentially dropping the available bandwidth to that port as traffic on it
That would mean that a single large attachment would penalize other mail traffic. A better way would be to throttle the numer of new connections which can be established.
With port 25 blocked, zombie owners are forced to use the ISP's outgoing mail servers. If throttling is intelligently applied to all port 25 traffic on a per-host basis, the feasibility of zombie spamming drops off.
No, it just means that they need to infect more machines. It's easier for them to get around limitting and throttling if they have a large number of infected hosts than if they have to pass it through a single or few servers.
The OpenBSD team is working on a transparent traffic shaping proxy that will make magic like this trivial for the pf priesthood
While it isn't exponential, you can still limit the connection rate with iptables or tcplimit without much trouble. I'm not a big BSD guy, but I would expect there would already be provision to do something similar there...
steve
The catch is that then email sits on their lame SMTP server for x number of hours if it doesn't go out,
That sounds like it's working as designed. Depending on the error involved, mail servers are designed to either bounce it immediately, or queue it and keep trying.
If your mail server bounces email for every single transient error, you're going to bounce a very significant portion of your mail that really shouldn't be bounced.
The real solution is to filter spam using something like Spamassassin
No, the real solution is for every country to approve the death penalty for spammers, and allow extradition to any country with a warrant for them. Most anything less, and spam will continue. With filtering, they just increase the number they send by a few orders of magnitude, and they still get enough by to make a profit. Since they aren't paying for the vast majority of the resources used, then it doesn't matter to them if 99.9% of the mail gets filtered, they'll just send 1000 times more.
steve
However, for the tinfoil hat crowd and hardcore geek types, this can be a problem.
I'm not so sure that it is. Most of the folks that I know who are running a mail server on their DSL or cable line really aren't all that qualified to be running a mail server - despite the fact that they always, always, always think that they are.
Yes, I've run across a few that are qualified, but they're vastly in the minority. As much as I'd love to run my own mail server at home, I just don't do it. Enough of my mail would be blocked anyway just because it's coming from a cable network.
As to blocking mail coming from a cable network, I don't blame people, either. I do it all the time on my mail servers. If I block mail coming from cable modems (NOT the cable company's mail servers, but the modem pools themselves), I block at least 20,000,000 pieces of spam for every one legitimate message that I lose. No, those numbers are not exagerations.
steve
You're confusing constitutionally-guaranteed rights with priveliges.
You have the right to be free from unreasonable search and seizure.
You do NOT have the right to do whatever you want with my private property. My mail server is my private property.
If you argue that, then I should be free to walk over to your house and use your car whenever you're not, because as long as I fill it up with gas, no damage has been done to you.
steve
steve
Assume a laptop gets a 5hr runtime
There's your mistake. While those sorts of runtimes get claimed pretty often, I don't know a single person with a PC-based laptop that gets that sort of runtime - nearly all say that they get two hours or less.
Also, if you're splitting hairs here, going from watt-hours to watts and hours doesn't work out exactly in most situations, as the watt-hour rating is measured under the most ideal current draw and temperature, which you won't often achieve - and that's *IF* they were actually honest about the watt-hour rating, the batteries that actually measure up to the manufacturer's claims are very few and far between.
steve
Are you sure that it's 1/3 of the power budget? Western Digital's 2.5" drives only use 2.5 watts under read/write activity, and 2 watts idle - and they're nothing special.
So, for your claim to be true, then the CPU, northbridge, southbridge, memory, display, and everything else would have to consume no more than 4 watts...
steve
The CF+ and Compact Flash specification 3.0 includes UDMA 33 and UDMA 66 support. I've seen references to certain cards and CF->IDE adapters that support DMA, so that problem is partially solved, and will get better.
As for the problem of sustained speeds, there's always RAID 0...
steve
If you have 140 boys and 100 girls, then the percentage of boys is 58.3%, just 8.3% above the expected 50%.
If I were a more sarcastic person, I'd question whether a comment on a statistical phenominon could be trusted if the author didn't grasp simple percentage problems... but I wouldn't want to hurt any feelings. =)
steve
It is true that the sperm determine the gender. However, your assumption that geeks will cause being female to become a recessive trait is a bit flawed, because (a) geeks are a relatively small portion of the population, and (b) it could be argued that their contribution to the next generation is relatively small, as well.
If anything, I would argue that persons of lower intelligence are more likely to breed early and breed often, and that intelligence itself is slowly becoming a much less common trait...
steve
You're assuming that the first sperm to the egg causes conception, when that's not true. It takes a very good number just to wear down and break through the protective coating on the egg, and *then* one lucky little swimmer becomes the star, so to speak. You could also theorize that the faster Y-based sperm got their first, broke down the barrier, and a somewhat-late-coming X-based sperm was the "winner". There is a lot that can happen. : )
The research in the article seems to be nothing but statistical. As with any but the most simple biological processes, there are hundreds or thousands (or TENS of thousands) of intermediate steps along the way, a change in any of which can have large effects.
One of the pathways I once studied had to do with Vul/Muv defects in flatworms - either the lack of a vulva in a female, or the presence of many vulvas. The developmental process involved a very good number of genes, each coding for proteins that had to interact with their surroundings in a certain way, over a certain chain of events in order to trigger the production of a vulva. An anomoly at almost any one of these stages could produce one of the two mentioned defects, and if I recall, certain anomolies could mask previous anomolies - it wasn't as simple as "Vul defect, must have a defective {XXX} gene." It was more like "Vul defect, could be any of twenty different defective genes or any of a complete set of combinations."
steve
steve
You're right, it's the presence of a Y chromosome that ends up causing an embryo to become male - without it, the embryo becomes (by default) a girl. There are genetic conditions whereby someone with a Y chromosome can end up becoming a girl anyway, but that's a lecture for another day.
To get back on track, the first guess at an explanation would be that a man's "gender strength" (for lack of a better term) would not only determine what his personality turned out to be (and hence, what job he is in), but also what percentage of his sperm that reach the egg would have a Y chromosome as opposed to an X.
Off of the top of my head, I would imagine that a significant difference in the X:Y ratio in sperm would have been noticed before now - it could have to do with other factors, such as their Y-sperm having a greater ability to dissolve the protective barrier around the egg, being a little faster, or many other things - conception is an awfully complicated business, and there are plenty of small, subtle changes that can lead to relatively large differences in outcome.
steve