I don't find a reference now, but wasn't there a lawsuit about Popeye where the court decided that if the copyright is expired a trademark can't prevent copying of the work? Isn't this a similar situation?
I mean, if Internet Brands adds text containing the trademark "WikiTravel" to a CC work, they gave permission to use the trademark in the work under the CC license. If they use their own trademark as the title of the work, I don't see how they can prevent people to refer to the work using the trademark
I agree that the maintenace required by a diesel plant is low, but it is still much higher than that of solar. You will spent more man hours refilling the fuel than the total maintenance for the solar plant.
The rest of the post tries to compare apples to oranges. Either we compare technology currently available on the market. In that case the comparison I made is accurate.
Or you compared anticipated technolgies. In this case you have on the nuclear side proposals for inexpensive fuel cycles with greatly reduced risk. All of these designs will use more concrete per Watt electrical output so the energy paypack time will be greater than for current designs. But still these advanced reactors would be great improvements compared to current designs.
On the solar side there are concepts for designs that use much less material and less exotic materials. Thin film and metal based cells are on the verge of beeing market ready and there might be a breakthrough in polymere cells any time (so it is not guaranteed). By the time generation 4 reactors will be market ready solar power might be almost free.
Essentially it comes down to a bet that most of the industry currently is not willing to make: Gen 3 and Gen 4 reactors are designed to operate 65 years after a 15 year design time. Currently solar has a 20% to 30% cost improvement every year for 10 years now. Anybody putting his money on nuclear is betting that this progress suddenly stops before solar passes all other technologies. This might very well happen, but it might be smart to wait a few more years before commiting to an 80 year project competing with solar.
I read this a few year ago and current can't find it again. It might be that this number ist the energy for producing the plant divided by the average electrical power. This would mean that the energy for running the plant and for producing the fuel is not considerd. Sorry I don't have a better source. What I do find a lot are sources on the average amount of CO2 produced by nuclear power. It is a lot worse than wind and hydro but somewhat better than solar if the uranium does not come from south africa.
> Also this BS argument I constantly see without facts to back it up that some how solar cells release so much CO2 in their manufacture that they can't possibly offset the CO2 over their lives.
You here the same for energy saving light bulbs, etc. As a first ballpark estimate you can assume that even manufacturers have to pay for their energy. So if it is economical feasible the CO2 balance can't be that bad.
Actual there is data on that available. It is called energy pay back time. It is less than a year for solar thermal and wind energy. It is about 2 years for current nuclear plants. It is about 3 years for solar cells rapidly going down to 1 year anticipated for thin film technology. Of course there are big variations (about a faktor of two in both directions) depending on many parameters.
> Solar panels do require a lot of maintenance if you want good performance. Compared to a small diesel plan? Definitely not.
> Look up how much energy is used to produce one square centimetre of a solar panel. Energy parity is achieved quicker for solar panels than for nuclear power plants in some cases. Accoring to this source by the German government, CO2 emissions of a nuclear plant driven by Uranium from South Africa were higher than those of solar panels in 2007. http://www.bundestag.de/dokumente/analysen/2007/CO2-Bilanzen_verschiedener_Energietraeger_im_Vergleich.pdf Don't underestimate the CO2 emissions for creating concrete (6% of world CO2 production) and of Uranium enrichment and Uranium mining.
Solar has an annual improvement in efficicency in the order of 20% per year.
Authentication is something different, but the thread is about DNS. BIND, probably the same DNS that your IP-Hoster is running so it is likely proven that your laptop is playing nicely with it, is available for windows. The DHCP server in your router can tell all local machines that they should ask your local DNS to resolve addresses.
Or, you could add all your local devices IP numbers to the hosts file on all machines. For a small number of machines this should be feasible.
I am not an expert on authetification and file sharing, but Samba, LDAP, Kerberos, etc. are all available for windows.
Solar power also cools the earth (technically wind power is Solar too)..
Only if you store the electricity infinitely. Most of the electricity is converted to heat pretty quickly. The solar panel has a lower reflectivity as most materials so actually placing a solar panel might increase Earths temperature.
> Why after 7+ years is it all of a sudden an urgent matter that needs to be resolved immediately?
That is exactly the main question that a judge ruling an injunction has to decide and I am pretty sure that the German court will decline the injunction on this ground.
But I still don't see on which ground a foreign court shall deny the right to have this question checked by a competent court. (No, the US court is not competent to decide this.) The questions really are independent. The result of the US law suit will have no legal consequences for the relationship of the two companies in Germany.
How can doing something legal be a threat? They had to file the patent in dozens of countries, pay patent fees in dozens of countries and they can sue in each and every of these countries.
The result of the dispute will be resolved in each country seperately and may differ from country to country and will only be binding for that country. The lawsuit will be irrelevant for the use of the patented technology in europe and vice versa. So why should that wait?
Also: I agree that the injunction will not be granted in Germany, so it is even less a threat. We are talking Microsoft here, not private persons who are scared of courts or are forvce to settle because they can't pay the lawyers.
The problem is that Motorola is trying to circumvent the US legal system by filing a case in Germany about the exact same patents
But the same patent might be valid in one jurisdiction and invalid in another one. So essentially it is necessary to argue seperately in both jurisdictions.
Once the case here is settled either way, the judge will allow Motorola to file whatever it wants again in other countries.
Yeah, and by that time the injunction in Germany will be denied because Motorola did not take immediate action. Injunctions must be filed within a certain timeframe after detecting the violation. So the US judge is trying to overrule the German court which clearly is illegal.
B) Guy's not a pedophile Do nothing: Child is OK. Do something: Child is traumatized by 4 months of people talking about his dad beeing a criminal without getting the usual level of comforting from its dad to cope with this abnormal situation.
Fixed that for you.
Also please note: There is evidence that access to pornography (not necessarily child pornography) reduces the risk of someone molesting a child. (James D. Weinberg: Sexual Landscapes. S. 397ff)
The way to go is to use lossy compression formats based on 24 bit raw data with at least 96kHz sample rate. Reducing the file size drastically from that starting point is possible without any reduction in perceived quality. But doing that by the way the CD does (e.g. removing half the samples and cutting of the lower bits) does a really bad job of distributing the error.
Especially a dynamic of more than 16 bits is important for classical music or movie audio tracks. If you have a 60dB dynamic in a track, the silent parts will be quantized to 6 bits on a CD. A dolby audio stream will at a medium data rate will have much better signal quality than the CD in cases like that.
Of course the worst thing to do is to convert it to CD format first and then add lossy compression later, as you get the worst of both worlds.
> Are you really claiming that there are more researchers legitimately investigating porn websites than there are horny frat boys who just want to jerk off in their dorm rooms No. But I am claiming, that 1000 frat boys jerking of in their dorm do less harm, than one legitimiate research who can't do his work. If it were just porns, things would be easy. The point is, that in all sorts of areas administration tries to seperate useful from useless sites, but the people doing that have no way of knowing what will be required by their staff. (It might be porn for researcher X, a shopping site for researcher Y, slashdot for another one. I could imagine that access to warez websites has been important to the research of Lawrence Lessig.) So instead they should concentrate on detectign and filtering malware and not filter based on content.
> More software companies who have not figured out a better way to deliver their product than emailing it to random employees than random employees who would install every "screensaver" emailed to them by a criminal? Really? Because that sure sounds pretty implausible to me.
This is a misrepresentation of the case I made. The E-Mails we trying to send are to engineers specifically designated to develop driver installers in cooperation with us. Yet, there was no official approved way of getting.exe files to them, they had to use trial and error to figure out a way to get past their system. The random employee your are talking about should not have permission to install software in his system. But maybe they should have permission to receive.exe files to forward them to another employee with the correct permissions?
It is a long time that I have been to university, but I have similar trouble with customers. Our Engineers waste a lot of time trying to get software we developed for a customer to the customers engineers because any of the following occur frequently:
* dropbox is blocked *.exe and.dll are not allowed in e-mail * our hoster is in a class A net blacklisted by customers spam-filter * we chose a file name that matches some regular expression deemed dangerous by their IT staff * sftp is blocked and so on, and so on
This is fine, if there is a clear procedure handling these exceptions. (e.g. if a researcher writing a paper on porn site can walk up to the IT appartment and get a list of sites opened for his computer within in minutes.) But ultimately these restrictions serve no real purpose and just waste a lot of money in the form of time lost by both IT, administrative and research staff.
Also, I wonder if research really works, if researches have to convince a censoring body that there request to access a site is legitimate before they can proceed with their research. (Yes sir, gamesexpert.com is not a sex site!)
You can patent the "push e-mail to device instead of polling" part, because it does not relate to software. You could implement it with manual labor and pidgeons and would still be violating the patent.
What you can't patent in Germany is algorithm, so a claim that says "push e-mail where the e-mail headers are processed via regular expressions" would no be valid.
what is stranger is that if the patent is very broad, it could be invalid because regular letters, faxes, teletypes, etc. have all been based on pushing the message to the receiver so there should be prior art.
> It really gets absurd in the cases where you have 100,000 lines of code that's your own but including 2 lines of GPL code means including 2 lines of ANY code makes it a derivative work. That's copyright law.
> you have to give away your code as it's a derivative work. The GPL allows you to automatically obtain a license to the code you included, without getting a specific prior permission from the author. All you have to do is open up the code.
Of course all other ways to get a copyright license for those two lines are still open to you. (like, you know, paying the author)
So, the GPL is not taking anything away, it is just giving you an additional option to normal copyright law. It might be true that in your scenario this additional option is almost worthless, but you are still better off than with code that is not GPL licensed.
> Maybe you should stop being a child and trying to insinuate that the only way someone can have a monopoly is by being the only actor. What do you think "mono" means in the word?
The definition on Wikipedia is: "A monopoly (from Greek monos (alone or single) + polein (to sell)) exists when a specific person or enterprise is the only supplier of a particular commodity."
German anti trust laws therefore don't talk about monopolies but about illegal market domination which is assumed to happen above 30% market share.
>No, that is entirely incorrect. This has nothing to do with different works.
I was talking about one work, too.
>What the other poster was claiming was that there might be multiple instances of work X on your server, and some of those might be authorized. That is the scenario that I describe. Music distribution rights usually are held on a regional basis. It is very common that multiple entities hold the rights for the same work.
Multiple identical uploads of the work to youtube will end up in the same files (You can bet that google is using deduplication at least on a block level.)
A takedown notice for a legal copy will cause a deletion of all illegal copies under your reasoning.
Youtube might not even be the best example: Think of amazon market place: If a company asks Amazon to remove image of their product (or even their logo) to be removed from the website because a competitor is using it without permission, Amazon would have to remove the image also from the product offerings of the original company.
> However, they are required to remove the file, not just > one link. If some other user had an AUTHORIZED link to that file, they can file a counter-notice, and the host would restore the file.
This means at least the damage is not permanent, but the damage is caused.
While I agree that Megaupload are fraudsters, their reasoning in this topic actually makes a lot of sense.
> The law does not say 'remove a single link to the material', or 'remove one copy of the material', it says 'remove the material'.
So, say, some music company (A) is using youtube to promote their music and upload a music video (X) that they have online disctribution rights for. Someone else (B) without these rights is uploading an illegal copy (Y) to youtube. Another rights holder (C) - say another music company from another country - issues a takedown notice for URL (Y).
What you are saying is that the DMCA requires youtube to remove the legitimate video (X) alongside with the illegal copy (Y)?
Wow.
Your legal system is so fucked up. I can't understand why anybody is still doing business with the US. It's just to risky.
What is the scope of the paragraph? Must a web hoster do this for all servers? So my data is at risk if I am using the same hoster as some illegitimite user?
This is not just about putting DRAM and a CPU on the same chip while keeping the architecture of both unchanged. This is about how computer architecture is effected by the possibility of implementing both on the same chip.
Dave Patterson noted in the nineties that the number of DRAM chips per computer went down with time. He predicted that DRAM will become large enough soon that at least the memory for a single process will fit into one chip soon. At that point it is unecessary slow and power consuming to move the data to the CPU and back for every computation (or alternativly spend 90% of the CPU chip area for cache to reduce the number of transports)
When you do put CPU and DRAM on the same chip the cost functions change and different architectures become optimal. Patterson noted that when you have a CPU and DRAM on the same chip the relative architectural cost functions will be similar to the technologies of the 70ies, just a few orders of magnitude smaller. Therefore he revisited architectures of that time and suggested to put a vector computer on a DRAM chip called the IRAM. http://www.cs.berkeley.edu/~pattrsn/talks/iram.html
Vector computers do not benefit much from cache. Latency is not a big issue for vector computers but they really benefit from bandwidth. On chip you can connect the DRAM to the CPU with 2048 bits bus width or more. (And the latency would be much smaller than the latency of a CPU going through a big cache hierachy and an external bus to the RAM)
If more memory is needed than fits on one chip he suggested to minimize data transports between chips. Instead the register state of the process would be migrated to the DRAM where the desired data resides.
Slashdot Mirror
I don't find a reference now, but wasn't there a lawsuit about Popeye where the court decided that if the copyright is expired a trademark can't prevent copying of the work? Isn't this a similar situation?
I mean, if Internet Brands adds text containing the trademark "WikiTravel" to a CC work, they gave permission to use the trademark in the work under the CC license.
If they use their own trademark as the title of the work, I don't see how they can prevent people to refer to the work using the trademark
The study is about kids who grow up in housholds where the parents speak different languages.
Not about kids who are taught a second language on purpose by a teacher. No discipline involved.
I agree that the maintenace required by a diesel plant is low, but it is still much higher than that of solar. You will spent more man hours refilling the fuel than the total maintenance for the solar plant.
The rest of the post tries to compare apples to oranges. Either we compare technology currently available on the market. In that case the comparison I made is accurate.
Or you compared anticipated technolgies. In this case you have on the nuclear side proposals for inexpensive fuel cycles with greatly reduced risk. All of these designs will use more concrete per Watt electrical output so the energy paypack time will be greater than for current designs. But still these advanced reactors would be great improvements compared to current designs.
On the solar side there are concepts for designs that use much less material and less exotic materials. Thin film and metal based cells are on the verge of beeing market ready and there might be a breakthrough in polymere cells any time (so it is not guaranteed). By the time generation 4 reactors will be market ready solar power might be almost free.
Essentially it comes down to a bet that most of the industry currently is not willing to make: Gen 3 and Gen 4 reactors are designed to operate 65 years after a 15 year design time. Currently solar has a 20% to 30% cost improvement every year for 10 years now. Anybody putting his money on nuclear is betting that this progress suddenly stops before solar passes all other technologies. This might very well happen, but it might be smart to wait a few more years before commiting to an 80 year project competing with solar.
I read this a few year ago and current can't find it again.
It might be that this number ist the energy for producing the plant divided by the average electrical power. This would mean that the energy for running the plant and for producing the fuel is not considerd. Sorry I don't have a better source.
What I do find a lot are sources on the average amount of CO2 produced by nuclear power. It is a lot worse than wind and hydro but somewhat better than solar if the uranium does not come from south africa.
> Also this BS argument I constantly see without facts to back it up that some how solar cells release so much CO2 in their manufacture that they can't possibly offset the CO2 over their lives.
You here the same for energy saving light bulbs, etc.
As a first ballpark estimate you can assume that even manufacturers have to pay for their energy. So if it is economical feasible the CO2 balance can't be that bad.
Actual there is data on that available. It is called energy pay back time. It is less than a year for solar thermal and wind energy. It is about 2 years for current nuclear plants. It is about 3 years for solar cells rapidly going down to 1 year anticipated for thin film technology.
Of course there are big variations (about a faktor of two in both directions) depending on many parameters.
> Solar panels do require a lot of maintenance if you want good performance.
Compared to a small diesel plan? Definitely not.
> Look up how much energy is used to produce one square centimetre of a solar panel.
Energy parity is achieved quicker for solar panels than for nuclear power plants in some cases.
Accoring to this source by the German government, CO2 emissions of a nuclear plant driven by Uranium from South Africa
were higher than those of solar panels in 2007.
http://www.bundestag.de/dokumente/analysen/2007/CO2-Bilanzen_verschiedener_Energietraeger_im_Vergleich.pdf
Don't underestimate the CO2 emissions for creating concrete (6% of world CO2 production) and of Uranium enrichment
and Uranium mining.
Solar has an annual improvement in efficicency in the order of 20% per year.
http://en.wikipedia.org/wiki/File:US_Trade_Balance_1980_2011.svg
Authentication is something different, but the thread is about DNS. BIND, probably the same DNS that your IP-Hoster is running so it is likely proven that your laptop is playing nicely with it, is available for windows. The DHCP server in your router can tell all local machines that they should ask your local DNS to resolve addresses.
Or, you could add all your local devices IP numbers to the hosts file on all machines. For a small number of machines this should be feasible.
I am not an expert on authetification and file sharing, but Samba, LDAP, Kerberos, etc. are all available for windows.
Are you aware that you can run software on windows machines that is not provided by microsoft but by other vendors?
Solar power also cools the earth (technically wind power is Solar too). .
Only if you store the electricity infinitely. Most of the electricity is converted to heat pretty quickly. The solar panel has a lower reflectivity as most materials so actually placing a solar panel might increase Earths temperature.
> Why after 7+ years is it all of a sudden an urgent matter that needs to be resolved immediately?
That is exactly the main question that a judge ruling an injunction has to decide and I am pretty sure that the German court will decline the injunction on this ground.
But I still don't see on which ground a foreign court shall deny the right to have this question checked by a competent court. (No, the US court is not competent to decide this.) The questions really are independent. The result of the US law suit will have no legal consequences for the relationship of the two companies in Germany.
How can doing something legal be a threat? They had to file the patent in dozens of countries, pay patent fees in dozens of countries and they can sue in each and every of these countries.
The result of the dispute will be resolved in each country seperately and may differ from country to country and will only be binding for that country. The lawsuit will be irrelevant for the use of the patented technology in europe and vice versa. So why should that wait?
Also: I agree that the injunction will not be granted in Germany, so it is even less a threat. We are talking Microsoft here, not private persons who are scared of courts or are forvce to settle because they can't pay the lawyers.
The problem is that Motorola is trying to circumvent the US legal system by filing a case in Germany about the exact same patents
But the same patent might be valid in one jurisdiction and invalid in another one. So essentially it is necessary to argue seperately in both jurisdictions.
Once the case here is settled either way, the judge will allow Motorola to file whatever it wants again in other countries.
Yeah, and by that time the injunction in Germany will be denied because Motorola did not take immediate action. Injunctions must be filed within a certain timeframe after detecting the violation. So the US judge is trying to overrule the German court which clearly is illegal.
You mean, like this:
http://i.dailymail.co.uk/i/pix/2010/02/25/article-1253724-0876E885000005DC-28_233x368.jpg
?
B) Guy's not a pedophile
Do nothing: Child is OK.
Do something: Child is traumatized by 4 months of people talking about his dad beeing a criminal without getting the usual level of comforting from its dad to cope with this abnormal situation.
Fixed that for you.
Also please note: There is evidence that access to pornography (not necessarily child pornography) reduces the risk of someone molesting a child. (James D. Weinberg: Sexual Landscapes. S. 397ff)
The way to go is to use lossy compression formats based on 24 bit raw data with at least 96kHz sample rate.
Reducing the file size drastically from that starting point is possible without any reduction in perceived quality. But doing that by the way the CD does (e.g. removing half the samples and cutting of the lower bits) does a really bad job of distributing the error.
Especially a dynamic of more than 16 bits is important for classical music or movie audio tracks. If you have a 60dB dynamic in a track, the silent parts will be quantized to 6 bits on a CD. A dolby audio stream will at a medium data rate will have much better signal quality than the CD in cases like that.
Of course the worst thing to do is to convert it to CD format first and then add lossy compression later, as you get the worst of both worlds.
> Are you really claiming that there are more researchers legitimately investigating porn websites than there are horny frat boys who just want to jerk off in their dorm rooms
No. But I am claiming, that 1000 frat boys jerking of in their dorm do less harm, than one legitimiate research who can't do his work. If it were just porns, things would be easy. The point is, that in all sorts of areas administration tries to seperate useful from useless sites, but the people doing that have no way of knowing what will be required by their staff. (It might be porn for researcher X, a shopping site for researcher Y, slashdot for another one. I could imagine that access to warez websites has been important to the research of Lawrence Lessig.)
So instead they should concentrate on detectign and filtering malware and not filter based on content.
> More software companies who have not figured out a better way to deliver their product than emailing it to random employees than random employees who would install every "screensaver" emailed to them by a criminal? Really? Because that sure sounds pretty implausible to me.
This is a misrepresentation of the case I made. The E-Mails we trying to send are to engineers specifically designated to develop driver installers in cooperation with us. Yet, there was no official approved way of getting .exe files to them, they had to use trial and error to figure out a way to get past their system. .exe files to forward them to another employee with the correct permissions?
The random employee your are talking about should not have permission to install software in his system. But maybe they should have permission to receive
Yes. But there also is research on porn.
It is a long time that I have been to university, but I have similar trouble with customers. Our Engineers waste a lot of time trying to get software we developed for a customer to the customers engineers because any of the following occur frequently:
* dropbox is blocked .exe and .dll are not allowed in e-mail
*
* our hoster is in a class A net blacklisted by customers spam-filter
* we chose a file name that matches some regular expression deemed dangerous by their IT staff
* sftp is blocked
and so on, and so on
This is fine, if there is a clear procedure handling these exceptions. (e.g. if a researcher writing a paper on porn site can walk up to the IT appartment and get a list of sites opened for his computer within in minutes.) But ultimately these restrictions serve no real purpose and just waste a lot of money in the form of time lost by both IT, administrative and research staff.
Also, I wonder if research really works, if researches have to convince a censoring body that there request to access a site is legitimate before they can proceed with their research. (Yes sir, gamesexpert.com is not a sex site!)
You can patent the "push e-mail to device instead of polling" part, because it does not relate to software. You could implement it with manual labor and pidgeons and would still be violating the patent.
What you can't patent in Germany is algorithm, so a claim that says "push e-mail where the e-mail headers are processed via regular expressions" would no be valid.
what is stranger is that if the patent is very broad, it could be invalid because regular letters, faxes, teletypes, etc. have all been based on pushing the message to the receiver so there should be prior art.
> It really gets absurd in the cases where you have 100,000 lines of code that's your own but including 2 lines of GPL code means
including 2 lines of ANY code makes it a derivative work. That's copyright law.
> you have to give away your code as it's a derivative work.
The GPL allows you to automatically obtain a license to the code you included, without getting a specific prior permission from the author. All you have to do is open up the code.
Of course all other ways to get a copyright license for those two lines are still open to you. (like, you know, paying the author)
So, the GPL is not taking anything away, it is just giving you an additional option to normal copyright law. It might be true that in your scenario this additional option is almost worthless, but you are still better off than with code that is not GPL licensed.
> Maybe you should stop being a child and trying to insinuate that the only way someone can have a monopoly is by being the only actor.
What do you think "mono" means in the word?
The definition on Wikipedia is:
"A monopoly (from Greek monos (alone or single) + polein (to sell)) exists when a specific person or enterprise is the only supplier of a particular commodity."
German anti trust laws therefore don't talk about monopolies but about illegal market domination which is assumed to happen above 30% market share.
> A takedown notice for a legal copy will cause a deletion of all illegal copies under your reasoning.
Sorry, the other way around...
>No, that is entirely incorrect. This has nothing to do with different works.
I was talking about one work, too.
>What the other poster was claiming was that there might be multiple instances of work X on your server, and some of those might be authorized.
That is the scenario that I describe.
Music distribution rights usually are held on a regional basis. It is very common that multiple entities hold the rights for the same work.
Multiple identical uploads of the work to youtube will end up in the same files (You can bet that google is using deduplication at least on a block level.)
A takedown notice for a legal copy will cause a deletion of all illegal copies under your reasoning.
Youtube might not even be the best example: Think of amazon market place: If a company asks Amazon to remove image of their product (or even their logo) to be removed from the website because a competitor is using it without permission, Amazon would have to remove the image also from the product offerings of the original company.
> However, they are required to remove the file, not just
> one link. If some other user had an AUTHORIZED link to that file, they can file a counter-notice, and the host would restore the file.
This means at least the damage is not permanent, but the damage is caused.
While I agree that Megaupload are fraudsters, their reasoning in this topic actually makes a lot of sense.
> The law does not say 'remove a single link to the material', or 'remove one copy of the material', it says 'remove the material'.
So, say, some music company (A) is using youtube to promote their music and upload a music video (X) that they have online disctribution rights for.
Someone else (B) without these rights is uploading an illegal copy (Y) to youtube.
Another rights holder (C) - say another music company from another country - issues a takedown notice for URL (Y).
What you are saying is that the DMCA requires youtube to remove the legitimate video (X) alongside with the illegal copy (Y)?
Wow.
Your legal system is so fucked up. I can't understand why anybody is still doing business with the US.
It's just to risky.
What is the scope of the paragraph? Must a web hoster do this for all servers? So my data is at risk if I am using the same hoster as some illegitimite user?
This is not just about putting DRAM and a CPU on the same chip while keeping the architecture of both unchanged.
This is about how computer architecture is effected by the possibility of implementing both on the same chip.
Dave Patterson noted in the nineties that the number of DRAM chips per computer went down with time. He predicted that DRAM
will become large enough soon that at least the memory for a single process will fit into one chip soon. At that point it is unecessary
slow and power consuming to move the data to the CPU and back for every computation (or alternativly spend 90% of the CPU chip
area for cache to reduce the number of transports)
When you do put CPU and DRAM on the same chip the cost functions change and different architectures become optimal.
Patterson noted that when you have a CPU and DRAM on the same chip the relative architectural cost functions will be similar to the
technologies of the 70ies, just a few orders of magnitude smaller. Therefore he revisited architectures of that time and suggested to
put a vector computer on a DRAM chip called the IRAM.
http://www.cs.berkeley.edu/~pattrsn/talks/iram.html
Vector computers do not benefit much from cache. Latency is not a big issue for vector computers but they really benefit from bandwidth.
On chip you can connect the DRAM to the CPU with 2048 bits bus width or more. (And the latency would be much smaller than the latency
of a CPU going through a big cache hierachy and an external bus to the RAM)
If more memory is needed than fits on one chip he suggested to minimize data transports between chips. Instead the register state of the
process would be migrated to the DRAM where the desired data resides.