The answer is no, Samba4 is not a good idea for admining a network of linux desktops. The point of Samba is to admin a windows network with a linux server.
The poster never mentioned windows, and is asking about a tool like Active Directory for linux. He likely just means distributed authentication management. The answer is likely openldap (with or without kerberos) For all the other functions, there are tools, like chef, puppet, dsh, etc... that are better than anything in the Windows world.
You're on slashdot, you know technology evolves. The tech on satellites is essentially CMOS cameras and computers to manage sending the data to ground stations. Satellite lifetime: 20 years from design to end of life... "Current" satellite designs are 20 years old. Launch costs are relatively the same. So the choice is to spend a tens (or hundreds) of millions of dollars to put up a platform, and operate for fifteen years or so. So do you put one up with technology that will be forty years old by the end of it's service life? It's going to be the same money whether it is 40 year old tech or twenty year old tech. Do you have a twenty year old CCD camera? Overall, Is that likely good value for money?
Democratic Society already ignores religion when it is inconvenient, and that is a good thing.
Divorce, adultery, abortion, and contraception are all legal, but against pretty much all religions. Ham and Alcohol are against some major ones. What makes you think that there is some form of religion in the laws or government today? When religion becomes wrong because society has changed, the laws follow society, not religion. In court swearing to tell the truth is enough. http://answers.yahoo.com/question/index?qid=20080218134708AAgMfKQ. Until 1956, Americans didn't trust God, but eachother: http://en.wikipedia.org/wiki/In_God_we_trust
I loved many of your earlier works that were very positive in explaining, in myriad ways, and with extraordinary beauty, how evolution gave rise to complexity and the wide variety of life we see today. Later works, like the God Delusion, are less about the beauty of nature, and more about the problem of the religious. You have made the point that religion is a sort of collective mental illness. It strikes me that the later, more strident work is either preaching to the converted, or scolding of a sick person in a hospital bed. So I guess the question is two fold.
1)
Is there something that precipitated the change in focus from uplifting exposition to intellectual bouncer? Did you meet a particularly condescending cleric in a dark alley and suddenly snap?
2)
If we accept as a hypothesis that religion is a mental disease, I'm not sanguine that saying "you crazy fools!" and explaining why is going to snap them out of it. If there any rational plan that could have any hope of actually working? What sort of meme could be established that would have better evolutionary stability, and gradually supplant religion in the pool. That's the goal, right?
Good point. I think the post was saying two things: 1) radiation shielding & repair strategies are not a big deal (we already use Fail-in-place to not touch systems for X year life cycles, for things like containerized data centres, or supercomputers.) 2) the unique environment allows some new choices...
the article says it is reasonable to expect such computers to run one hundred times faster, and/or consume 100x less power than conventional systems. We are talking about supercomputers, so those order of magnitude turn into large numbers, in watts, in tonnes. Considering launch costs, it might be worthwhile to incur the technical risk. So yeah, it harder, but if you are going to go to the trouble of putting it on the moon, wouldn't you want one that went 100x faster for the same power envelope?
Those are not real difficulties. The computing centre would be underground, that provides excellent radiation shielding. Computer just needs to survive transportation (when it will not be running) once. much simpler than the shuttle. You don't repair anything, just send a bit extra and apply fail-in-place maintenance strategy... What would be really cool is if they plan to operate at a natural temp... they could be designed for exploit superconduction... maybe the computer would be completely different from earthbound designs.
Robotics and autonomous systems are progressing to the point vehicles will be self-guiding... not just cars, but trucks, farm equipment, tractor trailers. Expert systems, such as Siri and Watson, will become skilled enough to replace most white collar work. Productivity of current technology is such that material needs are satisfied event with 30-50% unemployment today (look at Spain, Greece) China will not escape, the next iphone will be assembled robotically. We will be left with a natural employment level of perhaps 5%. If we have 95% unemployment... what does that mean? The next major revolution is the end of work as a meaningful way of dividing the spoils of the economy, defining our worth. If we dont figure out how to deal with that, then there will be roving bands of desperately poor people, who get hungry, being corralled by robotic soldiers (likely remotely supervised) to protect property. Desperately poor people and hungry people in large groups have a tendendcy, similar to large herds of farm animals, to incubate new diseases.
If this isn't just politial b.s. then the only way to address it would be to share the source and toolchain, so that the client can build and sign their own firmware... support then becomes interesting... Care to cite any published reports auditing networking gear? Router code is typically closed source firmware, every model being different, and with a new revision coming out every few months. Knock-knock protocols, where you send a message to one port, then to another port, etc... as a combination to open a service, are impossible to detect using black box techniques. The only way you are going to find them is by looking at the code. Looking at the code is tedious and expensive, to the point that it becomes cheaper to just buy the stuff from a source you trust.
I can say with quite some confidence that no-one is auditing most code running in the real world. Audits of proprietary code, in practice, are prohibitively expensive and totally impractical. It's actually a very strong argument for running code built from source, ideally where an international community with a broad diversity of interests continually reviews the code...
It would actually be smart for Google to do this... They could provide a virtual appliance that was a completely open g+, and the appliance would have ads in it. yet another way to get eyeballs in front of ads. Those worried about their personal info, would have their info on their servers, they could inspect the code, and be confident that Google did not have access to it. Companies and interested people would just get a hosting agreement with any provider, or host it on a VM in their home, and everybody wins. The pages served would have embedded links back to google directly, not related to the server itself. People could strip out the ad services, but then it's just the traditional cat & mouse, where the patch stream would only be provided for the service with ads, or it could be a premium service... On-premise Google+ Apps, or whatever. They could earn revenue from this, the main issue is whether the revenue will offset the support costs... can they make it good enough to be easy.
..except that I dont want my tax dollars going to cure tropical disease that very few americans get.
I am wondering what kind of bastard you would have to be to see a problem with people willingly paying for things that they benefit from, while simultaneously thinking its OK for people to be forced into unwillingly pay for things that they will never take advantage of.
In the case of ant-depressants, they have been on the market for fifty years. The new ones do not introduce dramatic improvements, but because the patent is still in force they charge 4-5x what a generic with essentially the same active ingredient costs. Those people would still have obtained treatment. They receive the new drug because drug companies spend more on marketing to physicians, than they do on R&D ( http://www.plosmedicine.org/article/info:doi/10.1371/journal.pmed.0050001?imageURI=info:doi/10.1371/journal.pmed.0050001.t001 ), so this is a pure effort to extract billions of dollars from people for no other reason than profit. Profit is perfectly legitimate, but what kind of sick bastard thinks that buying mercs and bimmers for advertising & pharma execs is a higher societal goal than preventing millions of deaths in the the rest of the world?
This isn't some socialist pipe dream, it is straight-forward reality that those poor people eventually figure out our priorities, and the result is the riots we have seen in recent weeks across the muslim world. It is far cheaper to help people, rather than bomb them first, and then "re-construct"...
yes. It is extremely expensive to create new forms of anti-depressants, and treatments for erectile dysfunction... meanwhile tropical diseases don't have a business case. If that's all patents cand fund, it would be more straightforward to fund merit-based research into worthwhile causes directly with taxes (NIH), rather than have the market invent more profitable problems to address and completely avoid the ones that would do the world the most good....
http://canadasworld.wordpress.com/2009/11/29/orphan-drugs-for-orphan-diseases-the-non-profit-pharmaceutical-model/
Sure it might work, but at a huge cost. You need to ship new ones for every security patch to any of the libraries you use. You end up duplicating the maintenance effort of the distro for every dependency. Plus your memory usage goes through the roof, pulling in a separate copy of each library for each binary. This is one of the central functions/benefits of using a distros, and even big companies do that properly in recent years. For example:
well I have ten year old Loki games, and only some of the binaries run. ELF/glibc changes, sound & graphic libraries requiring old versions, because the current ones didn't maintain compatibility. I don't have the option of re-compiling closed source games. The Freetards are going to say "don't use evil closed source!"... OK, but people seem to like closed source games, and this Valve thing is about bringing more of them onto the platform. I think this will be great if part of making games available is to give closed source games a sort of build environment, where they could easily re-compile games with new OS levels, and if we bought it once on Steam, then we can get the updated binaries forever... ie. maintained binaries, that get refreshed when a new stable distro release happens, rather than binary compatibility.
Same thing would likely be beneficial for people on android.
Newell was saying windows 8 is a trainwreck because Microsoft is closing the ecosystem down so that you need to use their app store to get anything sold. That the PC becomes a loss leader for the app store, is what will push OEMs out of the business. For Valve/Steam, it is the classic case where someone does something for a while and makes a reasonable living until MS notices and they move in and squeeze everyone else out. Whether windows 8 works well or sells well is completely beside Newell's point. He is a third party app store, who figures MS has him in his sights, even though it is really just collateral damage in it's competition with Apple. He's right.
bellcos can load custom firmware on the phones. The customization can be very slight, as in.. use the bellco's app shop for firefoxOS apps. Bellco takes a cut of each purchase, and provides ''support'' for the devices. if the firmware is modified, they don't support. so you have to use their app store. If they are smart, they hook it up so that it recurses to app vendor storelets, or a clearinghouse run by mozilla. App vendors only make their apps for one store and bellco's import from there and agree to pay charge a percentage on top of a wholesale price. Everybody happy. Bellco much happier than with Android & iPhone... whose stores they get 0 revenue for.
So bellco's just use contract manufacturers and make vanity models. the contract manufacturer neither knows nor cares about the software, QA is by the bellcos. Bellcos support their own widgets. If bellcos want to keep costs down, or there are smart widget makers, they all agree on models and buy the same model in higher volume... and have the branding stay with the manufacturer so that commercials work in any country... but the model is that the bell cos do very light branding/app store redirecting to the firmware, so that the phone stays "theirs."
Rather than google using phones as a loss leader for ad revenue, the bellcos adopt the apple model and use phones as a loss leader for the app stores. It will only work if they keep the OS compatible so that there a many vendors out there and critical mass for app developers to target. Otherwise, it will be just like the walled gardens of smart phones of yesteryear...
uh... not caring if it is profitable? Support from all the carriers who feel that the use of their network entitles them to a share of the app store pie, after seeing their potted plants (weren't big enough to get to walled garden status) wre not compelling, they can see that if they work together, perhaps they will share more of the pie than Apple or Google do.
on the ship, setup up a linux or bsd pc as the local vpn end point. Rent a VPS at any of hundreds of such providers in the US.
for one household to do this, you can get a US server for 8$/month or less. You need to pay more for network capacity, but not a huge amount.
You set up 1 and only 1 VPN connection... NAT through it. The people on the ship just set their default routes (you provide a DHCP service.)
I would use a pair of Debians for this, but whatever works for you.
1992 - http://en.wikipedia.org/wiki/Compaq_Concerto . Palm could have grown their devices but didn't, apple tried with newton, windows tablet edition 2005 was originally a demo by MS in 2000. There was a lot of marketing effort around newton and the XP tablet edition, but they didn't sell. Anyone had been looking at history would have given it the thumbs down. Indeed, that's what the analysts were saying.
The point is that what has been done before isn't a good indicator of the future because technology advances (much more capable devices & software) and people understand a lot more. I don't think the ipad would have succeeded without the masses being prepared by the ipod and iphone... The real risk was the first ipod touch in 2007, which was just one type of ipod, and represented little risk in the overall scheme of things. After that, the interface was proven, and it was pretty low risk to do variations on that. Not so for everyone else, because android has never been as polished and consistent an experience, and started years later.
Looking for lessons in what Apple did as an example for others to follow is useless, because no-one else is in Apple's position, and doesn't have the same choices open to it. Example, It was great that Jobs fought against DRM and won, but all the other MP3 players were... well MP3 players, one can't say that the other actors were cheering on the media companies. It's just that Jobs, via ITunes, had a far better negotiating position than anyone else.
I don't think the way our lord Jobs did things provides useful examples for others. As an example, just replace IPAD by any other tablet, and see how it sounds... ie. if RIM had not introuced the Playbook... if you take a probabilistic case, no-one should be launching tablet products, because apple's are the only ones that sell.
The US government loans were repaid as well. Americans decry jobs going to foreigners when their government "saves" the company. they fail to realize that foreign governments do similar things, in this case putting Canadian citizens on the hook for, per capita, twice as much as Americans.
Slashdot Mirror
The answer is no, Samba4 is not a good idea for admining a network of linux desktops. The point of Samba is to admin a windows network with a linux server. The poster never mentioned windows, and is asking about a tool like Active Directory for linux. He likely just means distributed authentication management. The answer is likely openldap (with or without kerberos) For all the other functions, there are tools, like chef, puppet, dsh, etc... that are better than anything in the Windows world.
Where is this house? can you get FTTH? if not, Cable? If you are stuck on DSL... well, forget it, it will never be a real geek house.
You're on slashdot, you know technology evolves. The tech on satellites is essentially CMOS cameras and computers to manage sending the data to ground stations. Satellite lifetime: 20 years from design to end of life... "Current" satellite designs are 20 years old. Launch costs are relatively the same. So the choice is to spend a tens (or hundreds) of millions of dollars to put up a platform, and operate for fifteen years or so. So do you put one up with technology that will be forty years old by the end of it's service life? It's going to be the same money whether it is 40 year old tech or twenty year old tech. Do you have a twenty year old CCD camera? Overall, Is that likely good value for money?
Democratic Society already ignores religion when it is inconvenient, and that is a good thing. Divorce, adultery, abortion, and contraception are all legal, but against pretty much all religions. Ham and Alcohol are against some major ones. What makes you think that there is some form of religion in the laws or government today? When religion becomes wrong because society has changed, the laws follow society, not religion. In court swearing to tell the truth is enough. http://answers.yahoo.com/question/index?qid=20080218134708AAgMfKQ. Until 1956, Americans didn't trust God, but eachother: http://en.wikipedia.org/wiki/In_God_we_trust
I loved many of your earlier works that were very positive in explaining, in myriad ways, and with extraordinary beauty, how evolution gave rise to complexity and the wide variety of life we see today. Later works, like the God Delusion, are less about the beauty of nature, and more about the problem of the religious. You have made the point that religion is a sort of collective mental illness. It strikes me that the later, more strident work is either preaching to the converted, or scolding of a sick person in a hospital bed. So I guess the question is two fold.
1) Is there something that precipitated the change in focus from uplifting exposition to intellectual bouncer? Did you meet a particularly condescending cleric in a dark alley and suddenly snap?
2) If we accept as a hypothesis that religion is a mental disease, I'm not sanguine that saying "you crazy fools!" and explaining why is going to snap them out of it. If there any rational plan that could have any hope of actually working? What sort of meme could be established that would have better evolutionary stability, and gradually supplant religion in the pool. That's the goal, right?
Good point. I think the post was saying two things: 1) radiation shielding & repair strategies are not a big deal (we already use Fail-in-place to not touch systems for X year life cycles, for things like containerized data centres, or supercomputers.) 2) the unique environment allows some new choices...
I think 1) is pretty solid. 2) is more admittedly quite a bit more speculative... why bother with 2? well according to this: http://www.academia.edu/1328244/SuperGreen_Computing_Superconducting_Computers_as_Green_Technology
the article says it is reasonable to expect such computers to run one hundred times faster, and/or consume 100x less power than conventional systems. We are talking about supercomputers, so those order of magnitude turn into large numbers, in watts, in tonnes. Considering launch costs, it might be worthwhile to incur the technical risk. So yeah, it harder, but if you are going to go to the trouble of putting it on the moon, wouldn't you want one that went 100x faster for the same power envelope?
Those are not real difficulties. The computing centre would be underground, that provides excellent radiation shielding. Computer just needs to survive transportation (when it will not be running) once. much simpler than the shuttle. You don't repair anything, just send a bit extra and apply fail-in-place maintenance strategy... What would be really cool is if they plan to operate at a natural temp... they could be designed for exploit superconduction... maybe the computer would be completely different from earthbound designs.
Robotics and autonomous systems are progressing to the point vehicles will be self-guiding... not just cars, but trucks, farm equipment, tractor trailers. Expert systems, such as Siri and Watson, will become skilled enough to replace most white collar work. Productivity of current technology is such that material needs are satisfied event with 30-50% unemployment today (look at Spain, Greece) China will not escape, the next iphone will be assembled robotically. We will be left with a natural employment level of perhaps 5%. If we have 95% unemployment... what does that mean? The next major revolution is the end of work as a meaningful way of dividing the spoils of the economy, defining our worth. If we dont figure out how to deal with that, then there will be roving bands of desperately poor people, who get hungry, being corralled by robotic soldiers (likely remotely supervised) to protect property. Desperately poor people and hungry people in large groups have a tendendcy, similar to large herds of farm animals, to incubate new diseases.
I'd rather trust a western company... http://spectrum.ieee.org/riskfactor/telecom/security/nortel-penetrated-by-hackers-since-at-least-2000
If this isn't just politial b.s. then the only way to address it would be to share the source and toolchain, so that the client can build and sign their own firmware... support then becomes interesting... Care to cite any published reports auditing networking gear? Router code is typically closed source firmware, every model being different, and with a new revision coming out every few months. Knock-knock protocols, where you send a message to one port, then to another port, etc... as a combination to open a service, are impossible to detect using black box techniques. The only way you are going to find them is by looking at the code. Looking at the code is tedious and expensive, to the point that it becomes cheaper to just buy the stuff from a source you trust.
I can say with quite some confidence that no-one is auditing most code running in the real world. Audits of proprietary code, in practice, are prohibitively expensive and totally impractical. It's actually a very strong argument for running code built from source, ideally where an international community with a broad diversity of interests continually reviews the code...
It would actually be smart for Google to do this... They could provide a virtual appliance that was a completely open g+, and the appliance would have ads in it. yet another way to get eyeballs in front of ads. Those worried about their personal info, would have their info on their servers, they could inspect the code, and be confident that Google did not have access to it. Companies and interested people would just get a hosting agreement with any provider, or host it on a VM in their home, and everybody wins. The pages served would have embedded links back to google directly, not related to the server itself. People could strip out the ad services, but then it's just the traditional cat & mouse, where the patch stream would only be provided for the service with ads, or it could be a premium service... On-premise Google+ Apps, or whatever. They could earn revenue from this, the main issue is whether the revenue will offset the support costs... can they make it good enough to be easy.
In the case of ant-depressants, they have been on the market for fifty years. The new ones do not introduce dramatic improvements, but because the patent is still in force they charge 4-5x what a generic with essentially the same active ingredient costs. Those people would still have obtained treatment. They receive the new drug because drug companies spend more on marketing to physicians, than they do on R&D ( http://www.plosmedicine.org/article/info:doi/10.1371/journal.pmed.0050001?imageURI=info:doi/10.1371/journal.pmed.0050001.t001 ), so this is a pure effort to extract billions of dollars from people for no other reason than profit. Profit is perfectly legitimate, but what kind of sick bastard thinks that buying mercs and bimmers for advertising & pharma execs is a higher societal goal than preventing millions of deaths in the the rest of the world?
This isn't some socialist pipe dream, it is straight-forward reality that those poor people eventually figure out our priorities, and the result is the riots we have seen in recent weeks across the muslim world. It is far cheaper to help people, rather than bomb them first, and then "re-construct"...
yes. It is extremely expensive to create new forms of anti-depressants, and treatments for erectile dysfunction... meanwhile tropical diseases don't have a business case. If that's all patents cand fund, it would be more straightforward to fund merit-based research into worthwhile causes directly with taxes (NIH), rather than have the market invent more profitable problems to address and completely avoid the ones that would do the world the most good. ...
http://canadasworld.wordpress.com/2009/11/29/orphan-drugs-for-orphan-diseases-the-non-profit-pharmaceutical-model/
zenzen% file /usr/lib/chromium-browser/chromium-browser /usr/lib/chromium-browser/chromium-browser: ELF 64-bit LSB shared object, x86-64, version 1 (SYSV), dynamically linked (uses shared libs), for GNU/Linux 2.6.24, BuildID[sha1]=0xd411a190a5169a1e304b61bdf356cb8a62dc4e89, stripped
Unless there is a super important reason, using the distro supplied version (not pulling things directly from a developer) is the best way to go.
well I have ten year old Loki games, and only some of the binaries run. ELF/glibc changes, sound & graphic libraries requiring old versions, because the current ones didn't maintain compatibility. I don't have the option of re-compiling closed source games. The Freetards are going to say "don't use evil closed source!" ... OK, but people seem to like closed source games, and this Valve thing is about bringing more of them onto the platform. I think this will be great if part of making games available is to give closed source games a sort of build environment, where they could easily re-compile games with new OS levels, and if we bought it once on Steam, then we can get the updated binaries forever... ie. maintained binaries, that get refreshed when a new stable distro release happens, rather than binary compatibility.
Same thing would likely be beneficial for people on android.
how about storing it in electric cars?
Newell was saying windows 8 is a trainwreck because Microsoft is closing the ecosystem down so that you need to use their app store to get anything sold. That the PC becomes a loss leader for the app store, is what will push OEMs out of the business. For Valve/Steam, it is the classic case where someone does something for a while and makes a reasonable living until MS notices and they move in and squeeze everyone else out. Whether windows 8 works well or sells well is completely beside Newell's point. He is a third party app store, who figures MS has him in his sights, even though it is really just collateral damage in it's competition with Apple. He's right.
So bellco's just use contract manufacturers and make vanity models. the contract manufacturer neither knows nor cares about the software, QA is by the bellcos. Bellcos support their own widgets. If bellcos want to keep costs down, or there are smart widget makers, they all agree on models and buy the same model in higher volume... and have the branding stay with the manufacturer so that commercials work in any country... but the model is that the bell cos do very light branding/app store redirecting to the firmware, so that the phone stays "theirs." Rather than google using phones as a loss leader for ad revenue, the bellcos adopt the apple model and use phones as a loss leader for the app stores. It will only work if they keep the OS compatible so that there a many vendors out there and critical mass for app developers to target. Otherwise, it will be just like the walled gardens of smart phones of yesteryear...
uh... not caring if it is profitable? Support from all the carriers who feel that the use of their network entitles them to a share of the app store pie, after seeing their potted plants (weren't big enough to get to walled garden status) wre not compelling, they can see that if they work together, perhaps they will share more of the pie than Apple or Google do.
on the ship, setup up a linux or bsd pc as the local vpn end point. Rent a VPS at any of hundreds of such providers in the US. for one household to do this, you can get a US server for 8$/month or less. You need to pay more for network capacity, but not a huge amount. You set up 1 and only 1 VPN connection... NAT through it. The people on the ship just set their default routes (you provide a DHCP service.) I would use a pair of Debians for this, but whatever works for you.
http://www.youtube.com/watch?v=kAG39jKi0lI
The point is that what has been done before isn't a good indicator of the future because technology advances (much more capable devices & software) and people understand a lot more. I don't think the ipad would have succeeded without the masses being prepared by the ipod and iphone... The real risk was the first ipod touch in 2007, which was just one type of ipod, and represented little risk in the overall scheme of things. After that, the interface was proven, and it was pretty low risk to do variations on that. Not so for everyone else, because android has never been as polished and consistent an experience, and started years later.
Looking for lessons in what Apple did as an example for others to follow is useless, because no-one else is in Apple's position, and doesn't have the same choices open to it. Example, It was great that Jobs fought against DRM and won, but all the other MP3 players were... well MP3 players, one can't say that the other actors were cheering on the media companies. It's just that Jobs, via ITunes, had a far better negotiating position than anyone else.
I don't think the way our lord Jobs did things provides useful examples for others. As an example, just replace IPAD by any other tablet, and see how it sounds... ie. if RIM had not introuced the Playbook... if you take a probabilistic case, no-one should be launching tablet products, because apple's are the only ones that sell.
The US government loans were repaid as well. Americans decry jobs going to foreigners when their government "saves" the company. they fail to realize that foreign governments do similar things, in this case putting Canadian citizens on the hook for, per capita, twice as much as Americans.
Crown Vic? St-Catherines, Ontario... Thats in Québec, right? http://en.wikipedia.org/wiki/St._Thomas_Assembly uh... Ford has never had any plant in Québec. Hyundai did: http://en.wikipedia.org/wiki/Bromont,_Quebec#Hyundai_plant (RIP: 1994), GM did ( http://en.wikipedia.org/wiki/Sainte-Th%C3%A9r%C3%A8se_Assembly ), (RIP 2002) but Ford/Linocln? never heard of it. With the Canadian Dollar trading at par with the US, that has meant a 30% increase in labour costs up here, and plants a closing left and right, and moving to the US (Caterpillar: http://www.bloomberg.com/news/2012-02-23/caterpillar-union-workers-ratify-severance-package-as-ontario-plant-shuts.html ) oh, and fwiw, for every dollar each American put into saving GM, a Canadian put in 2$. (http://www.cbc.ca/news/canada/windsor/story/2010/04/21/wdr-detroit-gm-government-loans-100421.html) oh, and the St-Catherines plant closed last year.