I've always thought of data projection queries in terms of Z-buffer processing. It would be interesting to see what a GPGPU could do for such queries.
For example, pricing, products, and services often have start and end date-times. Given a particular date, the effective pricing is most recently started data set that hasn't expired yet. It sounds easy in english, but it tends to be a rather bruatal set of unions and hierarchical-join queries to implement.
But given the I/O intensive nature of such processing, I'm not sure hardware accelerating the query evaluation would pay off.
Perhaps a better approach would be to use the power of such a processor to provide a hardware-based data widget implementation, such as topo-maps used to visualize geographic data sets. Where reporting tools pull in a data set then massage it, the GPGPU would be fed the data set and just display it directly.
It's rare these days to find a public site that depends only on IE.
I guess you don't download updates for video games. Several such sites mandate IE, and look like absolute crap in any other browser. Some won't let you download updates with anything other than their IE plugins.
I don't think that's true any more. This time it would be reasonable for Microsoft to rewrite their browser in C#.Net, which theoretically provides the kind of sandboxing protection that prevents buffer overflows.
But would that address evil Java/J/Ecma Scripts? Image file exploits? Any of the vulnerabilities that are actually rooted in the Win32 APIs and the NT kernel?
Go do some searches on "ADD ADHD meth cocaine speed" and see what comes up for research reports. Deny it all you want, there is a correlation between the addictions and the use of the ADD/ADHD drugs earlier in life.
Of all the kids I've known who were "diagnosed" ADD/ADHD, only two respond well to Ritalin. One of them clearly does have a disorder that is managed by the drugs. The other 10-20 kids that I know who were "diagnosed" either showed no behavioral changes, had additional problems apparently caused by the drugs, and the majority were taken off the ADD/ADHD drugs when their parents realized the drugs made no difference in their child's attention span or behaviour.
You may despise postings that "spread these myths". I despise a "medical" community that pushes unnecessary drugs on children rather than telling the parents and schools to deal with the children and their problems.
The increase in ADD, ADHD, Asperger's, and Autism would seem to indicate that children are being "revved" beyond their abilities.
I don't think it's the "fault" of electronic entertainment, but rather the incessant push to not merely succeed, but to excel. Those children with a variety of educational/entertainment/sport activities end up more balanced, but are still stressed.
Another part of the problem is that parents and authorities would rather push pills for ADD/ADHD than punish a child. When we twitched around in our seats in school, we got punished and learned to pay attention (sort of.) Now they flag a "problem" and stuff the kid full of pills.
The truly scary thing is that statistics are now showing that the ADD/ADHD "patients" grow up to suffer an increase in cocaine and meth addiction problems. Not surprising when you realize that ADD/ADHD medications are speed, so they're just trying to maintain the addiction developed by the educational and medical systems that would rather drug children than deal with the problems.
The vulnerabilities are there for anyone to find, so not disclosing the results in a reasonably short time frame so they can be fixed would be irresponsible. Hiding vulnerability reports is only advantageous to closed source, where the crackers can't see the problematic code.
I haven't seen a BSOD in over 5 years, but I have made it a habit to shut down every night and restart the machine in the morning. Otherwise the machine gets slower and slower, and starts exhibiting odd behaviour.
My guess is there are still too many WinXP/Win32 programs that allocate resources which don't get properly released. I can't see Vista fixing that, even if every single bit of the entire OS and all it's applications were recoded in C#.net. As long as apps need to communicate, one of them needs to create the shared objects, which means they can leak.
The most important change I hope to see with WinVista is security improvements, not stability. Running a post-Win2K server with one service is stable enough for the reboot-once-a-week datacenter, and every box in those datacenters gets rebooted including the mainframes, AIX, Solaris, Linux, and HPUX boxen.
If nothing else, a reboot is a precautionary measure to make sure that all config and services is running as intended. Even experienced SA's will periodically forget to do something like a/etc/rc.d/some-service restart after changing config files.
The biggest push I've heard given to corps over the years is not that OSS can be modified, enhanced, integrated, or reused, but that it can be inspected, reviewed, and fixed.
If there is anyone working in OSS who doesn't appreciate receiving such an analysis of potential bugs, then they shouldn't be programming anywhere. Whether for fun or profit, fixing the bugs and adding features is what the "job" is.
When the patent system was first set up, people had to provide prototypes and/or plans for the physical items being patented. There were clear ways to see how different designs worked, and how they might overlap on the functionality of their designs.
Software patents are a word game worse than any case law has been in millenia.
I guarantee such abuses and bullshit are not what the creators of the patent system envisioned. But the creators are long dead and those who object are not the ones with the money to buy lobbyists and laws.
Lead poisoning occurs regardless of the size of the lead particles.
It seems the article poster has a reputation, based on the grandparent comment. If they can try to spin lead poisoning as proof that nano-tech is safe and keep a straight face, they must have spent part of their career working for the tobacco industry.
Doesn't Lucas theatre digital system have a lower resolution than the HD disk formats? If so, there may be some truth to grainy or blurred image scaling. The same would be true of any older digital media remastered to a higher resolution format.
New features are all well and good, except when you consider where Microsoft actually is on their roadmap to features promised since Chicago.
The one thing I look forward to is seeing how many exploits there are that can bypass the core OS, both with and without AV software. If the focus has been on security as it was publicly announced to be, there should be a dramatic reduction in the number of exploits.
That is the one thing that could pique my interest in Vista.
But it doesn't get US$300 worth of interest. I have better uses for the money.
The GPL is effectively a poison-pill license, not a viral license. It's saying that you can only write GPL code using a GPL base; a "viral" license would force the GPL on other code.
For example, people use Apache code modules to support GPL projects and products. If the GPL were viral, the Apache code would have to be redistributed under GPL. The GPL is a poison-pill in that you couldn't take that GPL code and merge it in to the Apache code base; only the original author or their legal representative could submit that code under an Apache license.
Of course if my understanding is wrong, then every web application out there running on Apache or using Apache libraries needs to rip out all GPL code or rip out all Apache code. Never mind all the other relevant licenses like Mozilla or BSD.
Getting vaguely back on topic, the US federal government has imposed an anti-R&D approach for decades.
Ask anyone in the US trying to do medical cannabis research if they've had any luck obtaining research materials, permits, or approval to do useful studies. In the meantime, the federal government denies the validity of all "foreign" research in Canada, the UK, Israel, Australia, etc.
What was the purpose of the IBM breakup a few decades ago, if not to stop a company from leveraging their own investment in R&D to continue growing their business? In theory it was because IBM had grown to a near monopoly, yet no action is taken against Microsoft when they are far closer to a monopoly than IBM ever was. Obviously market dominance was not the reason for the breakup.
Pharmaceutical research is often forced offshore because US regulations don't permit the kind of testing that would be needed to determine the efficacy of some drugs. Plus that means the US government and US pharmacorps don't have the embarassment of another national Thalidomide debacle -- future mistakes will be kept out of sight in foreign nations.
Bottom line is the US government has done a great deal to ensure that true R&D doesn't happen, because what is a great new product/service line to the owner is a huge threat to the status quo that pays the lobbyists and thereby the government's members. R&D is profitable for new companies, but it's a loss for the ineffective and staid "competition" that cuts R&D budgets in favour of short-term profits to satisfy the stock market.
Therein lies the crux of the matter: The US corporations and federal government, or rather their management, will happily let anything crumble and die, provided they can turn a profit now.
Patents are supposed to be for specific implementations, not general ideas.
Check out the variety of automatic transmission designs, each under it's own patent. Yet clearly they do the same "obvious" task of shifting.
The only reason that isn't the case for software patents is that the USPTO and legal system haven't got a clue how to do anything but follow the money. And the money is in the hands of those who benefit from misinterpreting the law.
One likely incentive for that migration will be reduced storage costs.
I read that as "one likely media spin" rather than "likely incentive."
The main "incentive" is to migrate all your documentation to a new format that none of your older systems or employee's home machines can access -- unless you pay for upgrades. Paying around $300 for an upgrade to Office2K would pay for about 480GB of IDE PATA drives -- per user.
That's a lot of storage, even with bloated.DOC files of embedded images and spreadsheets.
Those sites which have "millions" of documents probably hang on to them for legal reasons, such as the financial services sector's 7-year history requirements. They're not allowed to "migrate" those documents, because that would change the archived data.
I also can't fathom why anyone would want to migrate millions of documents to a new format if the new version of Office can read the older document format. It's not like MS Office tools can directly scan the document repositories used by large corporations -- they're stored with search index databases and accessed through specialized tools, not mounted as network drives.
I think you misunderstand the impact on networked systems.
I build a software package that has a client-side key to verify that it's unaltered software, and use DRM to ensure that it can't have an exploit inserted. That client-side key is a "signature" enabling communications with the server, so the server can only "hear" unaltered clients.
While you could run the altered software without the key, it would be useless because it's not recognized by the server.
Even games aren't standalone programs anymore. I see DRM as part of a security solution incorporating Kerberos, physical token ids, encryption, etc. DRM's part is to protect the installed client software and provide client-side verification that the software has not been altered.
If forced to release the keys so that your modified client can "run normally" against the server, the entire security chain may as well be discarded.
openSuSE 10.1 actually makes it sickeningly easy to configure a firewall, subnet masquerading, DNS merging, and port forwarding. It took less than an two hours to get it all working (including dial-up and DHCP network alteration of the DNS forwarding.) IIRC it took almost two days to get it working with RedHat 5.2.
I realize it's not a fair comparison, as there is over 5 years of dev work in between the two, but the point is you don't need much knowledge, just a spare dual-nic box that'll run one of the more recent distros.
A friend of mine is a bit annoyed. It was faster and easier to set up SuSE's firewall and have it working reliably than his WinXP dial-up node.:P
Sure you can use a hardware packet firewall for the essential functionality, but what of detailed logs, intrusion detection software, VPN solutions, SSH proxying, etc.?
Aside from that, the hardware firewalls can be cracked, and have been in the past. But they're harder to upgrade and repair when an OSS patch is released a few hours later.
People place too much faith in firewalls anyhow. Worse, a lot of them enable uPNP functionality, which becomes a gaping hole in the security because anything connected to the internal network can temporarily enable a port forward/masquerade. i.e. A uPNP exploit on the internal network means your external firewall is enslaved by the exploit, the same way admin rights allow an exploit to disable a software firewall on a Wintendo box.
If I'm signing software to certify that it's been clean-built and tested, I do not want someone modifying the code and then distributing copies under my key. What's the point of PKCS if you give away the signing key?
If you want to run modified versions, build them, sign them, and add their signatures to your internal approved signatures. It won't talk to anyone else's version of the code, but that isn't their problem, it's your problem for not implementing the changes by sharing them with the community through the relevant project.
The only DRM-related point that I see as valid is that hardware manufacturers must make DRM keys available so people can build code for their hardware. Security through obscurity (restricting access to the hardware keys) always fails. Everyone with a functioning brain cell already knows that.
What would be the point of signatures if they no longer identify the signer? What is the point of signing code if some idiotic license mandates that it's no longer your signature?
Tied in with this whole mess is encrypted data/media delivery and storage. How the hell is someone supposed to be able to guarantee information privacy if you're forced to guarantee there is a back door accessing the information?
People just don't think any more. They pick up a pet cause and forget about the side-effects of their quick and dirty "solutions".
Rick Stallman has lost the original vision of shared source and confused it with shared access to data. There are segments of GPLv3 which try to restrict uses of the software, including the restrictions on encryption (which is all DRM is.) That goes directly against the original GPL, and it's not a change that can be blown off as "keeping up" with newer technologies like ASP and web delivery.
Slashdot Mirror
Where did I say the game required IE?
It's the game websites that require IE for the D/L manager/queuing systems in order to download updates and add-ons.
I've always thought of data projection queries in terms of Z-buffer processing. It would be interesting to see what a GPGPU could do for such queries.
For example, pricing, products, and services often have start and end date-times. Given a particular date, the effective pricing is most recently started data set that hasn't expired yet. It sounds easy in english, but it tends to be a rather bruatal set of unions and hierarchical-join queries to implement.
But given the I/O intensive nature of such processing, I'm not sure hardware accelerating the query evaluation would pay off.
Perhaps a better approach would be to use the power of such a processor to provide a hardware-based data widget implementation, such as topo-maps used to visualize geographic data sets. Where reporting tools pull in a data set then massage it, the GPGPU would be fed the data set and just display it directly.
I don't think that's true any more. This time it would be reasonable for Microsoft to rewrite their browser in C#.Net, which theoretically provides the kind of sandboxing protection that prevents buffer overflows.
But would that address evil Java/J/Ecma Scripts? Image file exploits? Any of the vulnerabilities that are actually rooted in the Win32 APIs and the NT kernel?
Go do some searches on "ADD ADHD meth cocaine speed" and see what comes up for research reports. Deny it all you want, there is a correlation between the addictions and the use of the ADD/ADHD drugs earlier in life.
Of all the kids I've known who were "diagnosed" ADD/ADHD, only two respond well to Ritalin. One of them clearly does have a disorder that is managed by the drugs. The other 10-20 kids that I know who were "diagnosed" either showed no behavioral changes, had additional problems apparently caused by the drugs, and the majority were taken off the ADD/ADHD drugs when their parents realized the drugs made no difference in their child's attention span or behaviour.
You may despise postings that "spread these myths". I despise a "medical" community that pushes unnecessary drugs on children rather than telling the parents and schools to deal with the children and their problems.
Once Bob was done with his dramatic thrashing, flailing, sweating, and panic, the IT department decided for him: no email.
One way or the other, Bob would be forced to speak to a human being.
"Better unplug the fax machine, too."
The increase in ADD, ADHD, Asperger's, and Autism would seem to indicate that children are being "revved" beyond their abilities.
I don't think it's the "fault" of electronic entertainment, but rather the incessant push to not merely succeed, but to excel. Those children with a variety of educational/entertainment/sport activities end up more balanced, but are still stressed.
Another part of the problem is that parents and authorities would rather push pills for ADD/ADHD than punish a child. When we twitched around in our seats in school, we got punished and learned to pay attention (sort of.) Now they flag a "problem" and stuff the kid full of pills.
The truly scary thing is that statistics are now showing that the ADD/ADHD "patients" grow up to suffer an increase in cocaine and meth addiction problems. Not surprising when you realize that ADD/ADHD medications are speed, so they're just trying to maintain the addiction developed by the educational and medical systems that would rather drug children than deal with the problems.
It's OPEN SOURCE.
The vulnerabilities are there for anyone to find, so not disclosing the results in a reasonably short time frame so they can be fixed would be irresponsible. Hiding vulnerability reports is only advantageous to closed source, where the crackers can't see the problematic code.
I haven't seen a BSOD in over 5 years, but I have made it a habit to shut down every night and restart the machine in the morning. Otherwise the machine gets slower and slower, and starts exhibiting odd behaviour.
My guess is there are still too many WinXP/Win32 programs that allocate resources which don't get properly released. I can't see Vista fixing that, even if every single bit of the entire OS and all it's applications were recoded in C#.net. As long as apps need to communicate, one of them needs to create the shared objects, which means they can leak.
The most important change I hope to see with WinVista is security improvements, not stability. Running a post-Win2K server with one service is stable enough for the reboot-once-a-week datacenter, and every box in those datacenters gets rebooted including the mainframes, AIX, Solaris, Linux, and HPUX boxen.
If nothing else, a reboot is a precautionary measure to make sure that all config and services is running as intended. Even experienced SA's will periodically forget to do something like a /etc/rc.d/some-service restart after changing config files.
The biggest push I've heard given to corps over the years is not that OSS can be modified, enhanced, integrated, or reused, but that it can be inspected, reviewed, and fixed.
If there is anyone working in OSS who doesn't appreciate receiving such an analysis of potential bugs, then they shouldn't be programming anywhere. Whether for fun or profit, fixing the bugs and adding features is what the "job" is.
I can't agree.
When the patent system was first set up, people had to provide prototypes and/or plans for the physical items being patented. There were clear ways to see how different designs worked, and how they might overlap on the functionality of their designs.
Software patents are a word game worse than any case law has been in millenia.
I guarantee such abuses and bullshit are not what the creators of the patent system envisioned. But the creators are long dead and those who object are not the ones with the money to buy lobbyists and laws.
FFS, is /. so desperate for articles that it's now a means for someone's BLOG to build traffic?
Rob, this is really sinking to a new low for content... :(
Lead poisoning occurs regardless of the size of the lead particles.
It seems the article poster has a reputation, based on the grandparent comment. If they can try to spin lead poisoning as proof that nano-tech is safe and keep a straight face, they must have spent part of their career working for the tobacco industry.
Doesn't Lucas theatre digital system have a lower resolution than the HD disk formats? If so, there may be some truth to grainy or blurred image scaling. The same would be true of any older digital media remastered to a higher resolution format.
New features are all well and good, except when you consider where Microsoft actually is on their roadmap to features promised since Chicago.
The one thing I look forward to is seeing how many exploits there are that can bypass the core OS, both with and without AV software. If the focus has been on security as it was publicly announced to be, there should be a dramatic reduction in the number of exploits.
That is the one thing that could pique my interest in Vista.
But it doesn't get US$300 worth of interest. I have better uses for the money.
If you were in Canada, they gave you zero hours so you'd quit.
If you quit, they don't have to pay severance. If they fire you or lay you off, they have to pay severance.
The "zero hour" tactic is very, very common in retail and food services industries.
The GPL is effectively a poison-pill license, not a viral license. It's saying that you can only write GPL code using a GPL base; a "viral" license would force the GPL on other code.
For example, people use Apache code modules to support GPL projects and products. If the GPL were viral, the Apache code would have to be redistributed under GPL. The GPL is a poison-pill in that you couldn't take that GPL code and merge it in to the Apache code base; only the original author or their legal representative could submit that code under an Apache license.
Of course if my understanding is wrong, then every web application out there running on Apache or using Apache libraries needs to rip out all GPL code or rip out all Apache code. Never mind all the other relevant licenses like Mozilla or BSD.
Getting vaguely back on topic, the US federal government has imposed an anti-R&D approach for decades.
Ask anyone in the US trying to do medical cannabis research if they've had any luck obtaining research materials, permits, or approval to do useful studies. In the meantime, the federal government denies the validity of all "foreign" research in Canada, the UK, Israel, Australia, etc.
What was the purpose of the IBM breakup a few decades ago, if not to stop a company from leveraging their own investment in R&D to continue growing their business? In theory it was because IBM had grown to a near monopoly, yet no action is taken against Microsoft when they are far closer to a monopoly than IBM ever was. Obviously market dominance was not the reason for the breakup.
Pharmaceutical research is often forced offshore because US regulations don't permit the kind of testing that would be needed to determine the efficacy of some drugs. Plus that means the US government and US pharmacorps don't have the embarassment of another national Thalidomide debacle -- future mistakes will be kept out of sight in foreign nations.
Bottom line is the US government has done a great deal to ensure that true R&D doesn't happen, because what is a great new product/service line to the owner is a huge threat to the status quo that pays the lobbyists and thereby the government's members. R&D is profitable for new companies, but it's a loss for the ineffective and staid "competition" that cuts R&D budgets in favour of short-term profits to satisfy the stock market.
Therein lies the crux of the matter: The US corporations and federal government, or rather their management, will happily let anything crumble and die, provided they can turn a profit now.
The USPTO is hiring. :p
Patents are supposed to be for specific implementations, not general ideas.
Check out the variety of automatic transmission designs, each under it's own patent. Yet clearly they do the same "obvious" task of shifting.
The only reason that isn't the case for software patents is that the USPTO and legal system haven't got a clue how to do anything but follow the money. And the money is in the hands of those who benefit from misinterpreting the law.
I read that as "one likely media spin" rather than "likely incentive."
The main "incentive" is to migrate all your documentation to a new format that none of your older systems or employee's home machines can access -- unless you pay for upgrades. Paying around $300 for an upgrade to Office2K would pay for about 480GB of IDE PATA drives -- per user.
That's a lot of storage, even with bloated .DOC files of embedded images and spreadsheets.
Those sites which have "millions" of documents probably hang on to them for legal reasons, such as the financial services sector's 7-year history requirements. They're not allowed to "migrate" those documents, because that would change the archived data.
I also can't fathom why anyone would want to migrate millions of documents to a new format if the new version of Office can read the older document format. It's not like MS Office tools can directly scan the document repositories used by large corporations -- they're stored with search index databases and accessed through specialized tools, not mounted as network drives.
I think you misunderstand the impact on networked systems.
I build a software package that has a client-side key to verify that it's unaltered software, and use DRM to ensure that it can't have an exploit inserted. That client-side key is a "signature" enabling communications with the server, so the server can only "hear" unaltered clients.
While you could run the altered software without the key, it would be useless because it's not recognized by the server.
Even games aren't standalone programs anymore. I see DRM as part of a security solution incorporating Kerberos, physical token ids, encryption, etc. DRM's part is to protect the installed client software and provide client-side verification that the software has not been altered.
If forced to release the keys so that your modified client can "run normally" against the server, the entire security chain may as well be discarded.
openSuSE 10.1 actually makes it sickeningly easy to configure a firewall, subnet masquerading, DNS merging, and port forwarding. It took less than an two hours to get it all working (including dial-up and DHCP network alteration of the DNS forwarding.) IIRC it took almost two days to get it working with RedHat 5.2.
I realize it's not a fair comparison, as there is over 5 years of dev work in between the two, but the point is you don't need much knowledge, just a spare dual-nic box that'll run one of the more recent distros.
A friend of mine is a bit annoyed. It was faster and easier to set up SuSE's firewall and have it working reliably than his WinXP dial-up node. :P
Sure you can use a hardware packet firewall for the essential functionality, but what of detailed logs, intrusion detection software, VPN solutions, SSH proxying, etc.?
Aside from that, the hardware firewalls can be cracked, and have been in the past. But they're harder to upgrade and repair when an OSS patch is released a few hours later.
People place too much faith in firewalls anyhow. Worse, a lot of them enable uPNP functionality, which becomes a gaping hole in the security because anything connected to the internal network can temporarily enable a port forward/masquerade. i.e. A uPNP exploit on the internal network means your external firewall is enslaved by the exploit, the same way admin rights allow an exploit to disable a software firewall on a Wintendo box.
If I'm signing software to certify that it's been clean-built and tested, I do not want someone modifying the code and then distributing copies under my key. What's the point of PKCS if you give away the signing key?
If you want to run modified versions, build them, sign them, and add their signatures to your internal approved signatures. It won't talk to anyone else's version of the code, but that isn't their problem, it's your problem for not implementing the changes by sharing them with the community through the relevant project.
The only DRM-related point that I see as valid is that hardware manufacturers must make DRM keys available so people can build code for their hardware. Security through obscurity (restricting access to the hardware keys) always fails. Everyone with a functioning brain cell already knows that.
What would be the point of signatures if they no longer identify the signer? What is the point of signing code if some idiotic license mandates that it's no longer your signature?
Tied in with this whole mess is encrypted data/media delivery and storage. How the hell is someone supposed to be able to guarantee information privacy if you're forced to guarantee there is a back door accessing the information?
People just don't think any more. They pick up a pet cause and forget about the side-effects of their quick and dirty "solutions".
Rick Stallman has lost the original vision of shared source and confused it with shared access to data. There are segments of GPLv3 which try to restrict uses of the software, including the restrictions on encryption (which is all DRM is.) That goes directly against the original GPL, and it's not a change that can be blown off as "keeping up" with newer technologies like ASP and web delivery.