Tesla is doing everything wrong, it's almost like they're trying to fail.
Tesla actually makes money on each car sold, and they have 17000 on waiting list so far. Tesla does everything right: They make an electric car that actually works, damn the price. Everyone else cuts corners and end up with a lousy car with miserable performance that runs out of battery all the time, and after all that they are not actually cheap. Tesla is for rich people who either just want something cool or like being environmentally friendly -- and that is actually a real market. Not an extremely large market, but it exists. All other electric cars are for... Well who exactly?
There are 50 preorders for Tesla Model S for taxi use in Denmark. Again, perfect for the purpose, large enough to be useful, fits with the environmental profile many companies demand from their suppliers, luxurious like a Mercedes -- and the battery pack can handle a day of city driving. The only electric competitor for the taxi market is the Renault Fluence with its battery swap system, but the boot space is too small for luggage and it is not a luxury car.
Of course the Model S is way too nice-looking to have a chance in the US, but the Model X attempts to fix that.
IDS will not help protect you from a DDOS. The closed it might come to offering any kind of DDOS protect is it may help your firewall thwart scanning and information gathering in preparation for a DDOS.
I would have agreed with you until recently, but today you can get IPS boxes which will do TCP SYN proxy (with cookies) and similar at 10Gbps. Now you can obviously get hit by more than 10Gbps of traffic, but in most cases that means you need to ask your provider for help anyway, since your own Internet connection is full. Some providers offer that you can pass dynamic blacklists to them which they will then install at their end of the connection, and some IDS boxes know how to provide such blacklists.
If people generally run more than one program on their FPGAs, then they are general purpose. Can you download apps for your FPGA?
But I could be wrong, perhaps Richard Stallman just means that they are computing devices which are popular. Then FPGAs would certainly count as generally used. I would certainly be in favour of that. Either way, case law would settle the question in no time at all.
I believe the "generally used" means the same as "general purpose" here. I could be wrong of course, but it would take only a single court case to get that one settled anyway. FPGA's, once they are programmed, are rarely general purpose, they do one specific thing only.
That wouldn't protect the armies of software developers in the world working on industrial or imbedded applications...
That is one reason I don't believe the proposal goes far enough. Richard Stallman is being too moderate.
Besides, how do DSPs escape the "generally used computing hardware" category? They're above your line, too.
Same as FPGA's: The program never changes. Possibly a little bit with firmware upgrades, but you never download an app for a DSP.
If the microcomputer is "generally used computing hardware", then it is not patent infringement. FPGA's are not at this time "generally used computing hardware", so the patent still applies to that line and above.
I would prefer a different split, but I think Stallman's proposal is clear and easy to understand. There will be a gray area -- is an iPhone "generally used computing hardware"? Is a Raspberry Pi in general? Is the Raspberry Pi GPU? Case law should make that gray area quite small in no time at all though.
It is fine to have machines counting the hand-marked slips of paper, so you can announce a preliminary result on election night -- although many places manage that perfectly well by hand.
The vote counters bring accountability to the table. You can never be truly sure that a machine is not compromised. If humans are compromised, we catch them and prosecute them, and a conspiracy needs to involve at least hundreds of people. With machines, a few people can compromise an entire election.
If you allow the machines into the voting booth, anonymous voting is in danger and voter mistakes become impossible to detect. If you allow them to actually record votes, the whole process becomes a joke.
I'm making the whole plane rectangular and you are complaining that electric motors are inefficient? Your concerns are valid, but even my ridiculous plane only has 20% of the required power.
The project itself is obviously within the realm of physical possibility as someone already made a plane seating two which can fly on solar power. Scaling it up can be done by simply building more of those and making them fly in formation. Physically possible, completely impractical.
I do not know how you got that a 737 has half of 3500 square feet per second. Let us try to do this in sensible units.
Solar insolation is 1366W/m2 and it is always noon. Solar cells are 100% efficient. Electric engines 100% efficient. Plane uses 89 M BTUs/h, which is 26MW. Now, jet engines are pretty good, but I bet they aren't 50% efficient, let us say 40%, so 10MW. 7300m2. Boeing 737, longest version, is 42m by 34m. Let us assume we make the plane square and cover it with solar cells (this may cause a slight degradation of flight performance, but since everything is perfect in this exercise, that is not a problem). That is 1400m2. Sadly the plane falls down.
The only way to win is to fly much slower and make the wings a lot larger. At which point trains overtake you and night falls and the plane drops out of the sky.
My experience with HP Procurve gear is that it does exactly what it says it does in the spec sheets. The software is rarely buggy, at least not uselessly so, and the hardware is very reliable -- and if it does have problems, they tend to show up from the start, not a year later. The challenge is that there are so few features. I haven't tried the 3com stuff that HP bought yet; back when it was 3com it was quite lousy.
Suddenly, you need IT guys trained ina bunch of vendors' network equipment, you need three different management and monitoring tools, and your op-ex is through the roof.
The challenge with Cisco is that you still need three different management and monitoring tools... IOS and IOS-XR and ASA-OS and NX-OS are similar in that you type commands and things happen...
They wanted a router which could handle both a T1 (so a slow legacy software router with obsolete interfaces) and a fiber (most likely 1Gbps ethernet). The 3945 is not a particularly stupid choice if those are the requirements and you like Cisco.
Now, split those requirements in two, and you can use a) any old T1 router and b) a random 1Gbps ethernet CPE. a) costs practically nothing refurbished. If you go with Cisco and you're a government agency, you can probably get an ME3400 with advanced license for under $1000. Juniper EX2200C + license is a bit cheaper than that, OneAccess is a lot cheaper.
Extra costs for the 2 router solution: A power bar with 2 outlets and an ethernet cable...
All joking aside, for most applications, we don't mind energy loss.
Algal biofuel was supposed to be energy positive. It will take up a whole lot of space which could have been covered with solar cells or farmed instead. It is a bit of a bummer to have to waste land on something which isn't energy positive. Methane synthesis doesn't take up much space.
I don't have a hardware DVD player anymore. XBMC's DVD support is surprisingly abysmal, at least under Linux. It can play about a third of all disks I try.
If is difficult to tell how much of that is caused by the rather non-standard things modern DVD's do to avoid being copied, and how is just caused by no one caring enough to fix the bugs. There are after all much nicer file formats which just provide the actual content and not all those awkward menus and unskippable adverts.
Is the government wrong? Probably, but you should focus on changing the government to change the laws rather than just breaking the law and hoping for the best.
Should you? It is unlikely that you will succeed, copyrights and patents have support from both sides of the spectrum. Right wingers think "people are taking stuff for free" and left wingers think "people are ripping off artists".
Meanwhile, no one is likely to prosecute you for watching a DVD you own. What exactly do you gain from not doing that?
You can't install a rootkit without it being written to the audit daemon (and you would have to reload the kernel anyways so it would require an unauthorized restart).
Of course you can install a rootkit wihout rebooting or writing to the audit daemon, if you have sufficient privileges. Just write to kernel memory. If you want to prevent that, you use mandatory access control. What you don't do is rely on obscure filesystem features. They won't save you.
Once you enable mandatory access control, you CANNOT truncate the access log and you CANNOT rename anything over it. Unless the security policy permits it, of course.
Really, I cannot believe that you are claiming that a silly limitation is a security feature.
No different in Windows. You can clear the audit log in Windows too, with sufficient privileges. And with sufficient privileges, you can just do a rootkit and hide everything. If you enable mandatory access control, none of those games are possible -- apart from bugs of course, but Windows has had tons of local privilege escalation bugs.
Do you really believe that when someone has unconstrained administrator access, the thing that will stop him from hiding his tracks is "oh, the audit daemon keeps its log file open, so I can't replace the audit log file"? If so, good luck with that.
If you want mandatory access control, use that. Linux has both SELinux and AppArmor for that purpose. Misusing file semantics to protect against a rootkit does not work. If you have kernel access, you can play all the games you want, including breaking promised file semantics.
Replacing open log files leave a log trail if you set that up in the audit system. That gets you right back to "explain that to your security administrator".
If you watch the "Fork Yeah!" video the impression I get is that it looks like they wanted to open source as much possible but was held back by legal.
So what if certain engineers wanted to open source things? They didn't get to make that decision.
The quotes are implying that the GPL does not work and that you can combine CDDL-licensed code with GPL'd code and distribute the combination. That position is rather weird, but then again Sun did suffer from a reality distortion field when it came to legal issues. The only other person I have heard of with the same view is Jörg Schilling.
You are going to come up with better arguments than that. Your quotes do not support that statement.
Sun was about as Linux-hostile as any company could get, basically from 1995 and forwards. They tried to do as much as they could to make sure that Linux did not benefit in any way from any Solaris or Sun technology.
Of course it makes sense that they tried to fight against the OS which was destined to make them obsolete. Luckily they did not have a particularly competent legal team.
If you have access to replace log files, you can probably install a proper root kit instead... But yes, a malicious root user in Unix can pull all sorts of tricks on unsuspecting users (even if those users believe they have root access themselves), and that can be a problem.
Nothing prevents the updater from forcing a reboot after finishing installing patches. Users hate it, so that doesn't happen. With Windows semantics, you end up replacing some of the files that a particular app needs but not others (because they happened to be open at the time). That is quite likely to cause mischief, so Windows tries to force reboots. Again, users hate it and click "reboot later" and go on to install more patches.
When some patches need to update in-use files, the patcher has no good way to identify that those files are already patched (because they're waiting as temporary files to be renamed over an in-use file).
If Windows had Unix file semantics, the updater could just do the renaming. The next patch along will then see the new contents, while whatever is keeping the file in-use will keep seeing the old contents. Thus it would be fairly trivial to just apply all the patches in sequence and reboot at the end.
Note that e.g. rpm/dpkg cannot work sensibly on a Windows system due to these awkward file semantics. Mandatory locks were introduced to Unix a long time ago, but happily practically no software uses them and you can safely just keep them turned off. In Windows they are used by practically everything (every executable locks its binary when executing).
A workaround is to have a standard placement and naming convention for patched in-use files. That way the next update could check that location first, before checking if the real destination is locked. Locking would be fun of course, and third-party updates better learn about that convention too.
Hmm, traditionally the only places people describe workarounds for Windows misfeatures is in patent documents. Perhaps I am missing an opportunity. It cannot be obvious, because Microsoft surely has a whole team working on Windows Update with at least one member counting as a person of ordinary skill in the art, and they did not seem to think of this fantastic idea.
Slashdot Mirror
Tesla is doing everything wrong, it's almost like they're trying to fail.
Tesla actually makes money on each car sold, and they have 17000 on waiting list so far. Tesla does everything right: They make an electric car that actually works, damn the price. Everyone else cuts corners and end up with a lousy car with miserable performance that runs out of battery all the time, and after all that they are not actually cheap. Tesla is for rich people who either just want something cool or like being environmentally friendly -- and that is actually a real market. Not an extremely large market, but it exists. All other electric cars are for... Well who exactly?
There are 50 preorders for Tesla Model S for taxi use in Denmark. Again, perfect for the purpose, large enough to be useful, fits with the environmental profile many companies demand from their suppliers, luxurious like a Mercedes -- and the battery pack can handle a day of city driving. The only electric competitor for the taxi market is the Renault Fluence with its battery swap system, but the boot space is too small for luggage and it is not a luxury car.
Of course the Model S is way too nice-looking to have a chance in the US, but the Model X attempts to fix that.
IDS will not help protect you from a DDOS. The closed it might come to offering any kind of DDOS protect is it may help your firewall thwart scanning and information gathering in preparation for a DDOS.
I would have agreed with you until recently, but today you can get IPS boxes which will do TCP SYN proxy (with cookies) and similar at 10Gbps. Now you can obviously get hit by more than 10Gbps of traffic, but in most cases that means you need to ask your provider for help anyway, since your own Internet connection is full. Some providers offer that you can pass dynamic blacklists to them which they will then install at their end of the connection, and some IDS boxes know how to provide such blacklists.
If people generally run more than one program on their FPGAs, then they are general purpose. Can you download apps for your FPGA?
But I could be wrong, perhaps Richard Stallman just means that they are computing devices which are popular. Then FPGAs would certainly count as generally used. I would certainly be in favour of that. Either way, case law would settle the question in no time at all.
I believe the "generally used" means the same as "general purpose" here. I could be wrong of course, but it would take only a single court case to get that one settled anyway. FPGA's, once they are programmed, are rarely general purpose, they do one specific thing only.
That wouldn't protect the armies of software developers in the world working on industrial or imbedded applications...
That is one reason I don't believe the proposal goes far enough. Richard Stallman is being too moderate.
Besides, how do DSPs escape the "generally used computing hardware" category? They're above your line, too.
Same as FPGA's: The program never changes. Possibly a little bit with firmware upgrades, but you never download an app for a DSP.
If the microcomputer is "generally used computing hardware", then it is not patent infringement. FPGA's are not at this time "generally used computing hardware", so the patent still applies to that line and above.
I would prefer a different split, but I think Stallman's proposal is clear and easy to understand. There will be a gray area -- is an iPhone "generally used computing hardware"? Is a Raspberry Pi in general? Is the Raspberry Pi GPU? Case law should make that gray area quite small in no time at all though.
It is fine to have machines counting the hand-marked slips of paper, so you can announce a preliminary result on election night -- although many places manage that perfectly well by hand.
The vote counters bring accountability to the table. You can never be truly sure that a machine is not compromised. If humans are compromised, we catch them and prosecute them, and a conspiracy needs to involve at least hundreds of people. With machines, a few people can compromise an entire election.
If you allow the machines into the voting booth, anonymous voting is in danger and voter mistakes become impossible to detect. If you allow them to actually record votes, the whole process becomes a joke.
I'm making the whole plane rectangular and you are complaining that electric motors are inefficient? Your concerns are valid, but even my ridiculous plane only has 20% of the required power.
The project itself is obviously within the realm of physical possibility as someone already made a plane seating two which can fly on solar power. Scaling it up can be done by simply building more of those and making them fly in formation. Physically possible, completely impractical.
I do not know how you got that a 737 has half of 3500 square feet per second. Let us try to do this in sensible units.
Solar insolation is 1366W/m2 and it is always noon. Solar cells are 100% efficient. Electric engines 100% efficient. Plane uses 89 M BTUs/h, which is 26MW. Now, jet engines are pretty good, but I bet they aren't 50% efficient, let us say 40%, so 10MW. 7300m2. Boeing 737, longest version, is 42m by 34m. Let us assume we make the plane square and cover it with solar cells (this may cause a slight degradation of flight performance, but since everything is perfect in this exercise, that is not a problem). That is 1400m2. Sadly the plane falls down.
The only way to win is to fly much slower and make the wings a lot larger. At which point trains overtake you and night falls and the plane drops out of the sky.
It's UTC-0600 or GMT-0600 if you don't believe in leap-seconds.
Only if you suffer from POSIXitis and think plus is minus and minus is plus. Plusungood.
My experience with HP Procurve gear is that it does exactly what it says it does in the spec sheets. The software is rarely buggy, at least not uselessly so, and the hardware is very reliable -- and if it does have problems, they tend to show up from the start, not a year later. The challenge is that there are so few features. I haven't tried the 3com stuff that HP bought yet; back when it was 3com it was quite lousy.
Suddenly, you need IT guys trained ina bunch of vendors' network equipment, you need three different management and monitoring tools, and your op-ex is through the roof.
The challenge with Cisco is that you still need three different management and monitoring tools... IOS and IOS-XR and ASA-OS and NX-OS are similar in that you type commands and things happen...
Disclaimer: I'm a Juniper fan boy.
They wanted a router which could handle both a T1 (so a slow legacy software router with obsolete interfaces) and a fiber (most likely 1Gbps ethernet). The 3945 is not a particularly stupid choice if those are the requirements and you like Cisco.
Now, split those requirements in two, and you can use a) any old T1 router and b) a random 1Gbps ethernet CPE. a) costs practically nothing refurbished. If you go with Cisco and you're a government agency, you can probably get an ME3400 with advanced license for under $1000. Juniper EX2200C + license is a bit cheaper than that, OneAccess is a lot cheaper.
Extra costs for the 2 router solution: A power bar with 2 outlets and an ethernet cable...
Trust me, even if you accept 35 C, the temperature goes well beyond that in a big hurry when the chillers cut out.
Only because the chillers going out kills the ventilation at the same time. THAT is unhealthy. Cooling a datacenter through radiation is adventurous.
All joking aside, for most applications, we don't mind energy loss.
Algal biofuel was supposed to be energy positive. It will take up a whole lot of space which could have been covered with solar cells or farmed instead. It is a bit of a bummer to have to waste land on something which isn't energy positive. Methane synthesis doesn't take up much space.
Only as long as you have analog outputs on the cable/satellite box.
I don't have a hardware DVD player anymore. XBMC's DVD support is surprisingly abysmal, at least under Linux. It can play about a third of all disks I try.
If is difficult to tell how much of that is caused by the rather non-standard things modern DVD's do to avoid being copied, and how is just caused by no one caring enough to fix the bugs. There are after all much nicer file formats which just provide the actual content and not all those awkward menus and unskippable adverts.
Is the government wrong? Probably, but you should focus on changing the government to change the laws rather than just breaking the law and hoping for the best.
Should you? It is unlikely that you will succeed, copyrights and patents have support from both sides of the spectrum. Right wingers think "people are taking stuff for free" and left wingers think "people are ripping off artists".
Meanwhile, no one is likely to prosecute you for watching a DVD you own. What exactly do you gain from not doing that?
You can't install a rootkit without it being written to the audit daemon (and you would have to reload the kernel anyways so it would require an unauthorized restart).
Of course you can install a rootkit wihout rebooting or writing to the audit daemon, if you have sufficient privileges. Just write to kernel memory. If you want to prevent that, you use mandatory access control. What you don't do is rely on obscure filesystem features. They won't save you.
Once you enable mandatory access control, you CANNOT truncate the access log and you CANNOT rename anything over it. Unless the security policy permits it, of course.
Really, I cannot believe that you are claiming that a silly limitation is a security feature.
No different in Windows. You can clear the audit log in Windows too, with sufficient privileges. And with sufficient privileges, you can just do a rootkit and hide everything. If you enable mandatory access control, none of those games are possible -- apart from bugs of course, but Windows has had tons of local privilege escalation bugs.
Do you really believe that when someone has unconstrained administrator access, the thing that will stop him from hiding his tracks is "oh, the audit daemon keeps its log file open, so I can't replace the audit log file"? If so, good luck with that.
If you want mandatory access control, use that. Linux has both SELinux and AppArmor for that purpose. Misusing file semantics to protect against a rootkit does not work. If you have kernel access, you can play all the games you want, including breaking promised file semantics.
Replacing open log files leave a log trail if you set that up in the audit system. That gets you right back to "explain that to your security administrator".
If you watch the "Fork Yeah!" video the impression I get is that it looks like they wanted to open source as much possible but was held back by legal.
So what if certain engineers wanted to open source things? They didn't get to make that decision.
The quotes are implying that the GPL does not work and that you can combine CDDL-licensed code with GPL'd code and distribute the combination. That position is rather weird, but then again Sun did suffer from a reality distortion field when it came to legal issues. The only other person I have heard of with the same view is Jörg Schilling.
That's a myth / blatant lie.
You are going to come up with better arguments than that. Your quotes do not support that statement.
Sun was about as Linux-hostile as any company could get, basically from 1995 and forwards. They tried to do as much as they could to make sure that Linux did not benefit in any way from any Solaris or Sun technology.
Of course it makes sense that they tried to fight against the OS which was destined to make them obsolete. Luckily they did not have a particularly competent legal team.
If you have access to replace log files, you can probably install a proper root kit instead... But yes, a malicious root user in Unix can pull all sorts of tricks on unsuspecting users (even if those users believe they have root access themselves), and that can be a problem.
Nothing prevents the updater from forcing a reboot after finishing installing patches. Users hate it, so that doesn't happen. With Windows semantics, you end up replacing some of the files that a particular app needs but not others (because they happened to be open at the time). That is quite likely to cause mischief, so Windows tries to force reboots. Again, users hate it and click "reboot later" and go on to install more patches.
When some patches need to update in-use files, the patcher has no good way to identify that those files are already patched (because they're waiting as temporary files to be renamed over an in-use file).
If Windows had Unix file semantics, the updater could just do the renaming. The next patch along will then see the new contents, while whatever is keeping the file in-use will keep seeing the old contents. Thus it would be fairly trivial to just apply all the patches in sequence and reboot at the end.
Note that e.g. rpm/dpkg cannot work sensibly on a Windows system due to these awkward file semantics. Mandatory locks were introduced to Unix a long time ago, but happily practically no software uses them and you can safely just keep them turned off. In Windows they are used by practically everything (every executable locks its binary when executing).
A workaround is to have a standard placement and naming convention for patched in-use files. That way the next update could check that location first, before checking if the real destination is locked. Locking would be fun of course, and third-party updates better learn about that convention too.
Hmm, traditionally the only places people describe workarounds for Windows misfeatures is in patent documents. Perhaps I am missing an opportunity. It cannot be obvious, because Microsoft surely has a whole team working on Windows Update with at least one member counting as a person of ordinary skill in the art, and they did not seem to think of this fantastic idea.
Why? so I can get less light output for the same watts used but instead of spending $8.95 per bulb I get to spend $39.99?
LED can come much closer to proper full-spectrum light than CFL ever will. CFL is just a stopgap technology we have to deal with until LED gets there.
Also, it is possible to make LED spotlights to handle that strange modern trend of building lots of spotlights into ceilings. CFL cannot do that.