of course, at the end of the day, by the numbers, google is an advertising sales company not a software company. that's probably a little unfair, but it's necessary to point out that there are many many projects at google that do not really earn significant income, and only a few that really sustain the vast majority of their cash flow. that is not to say that many of the other things they work on are not cool and useful; quite the opposite in fact. we should not discourage anyone from building something of incredible usefulness (which I think facebook has de facto just based on the amount of traffic it gets) just because we don't yet know where to attach the meter.
but at the end of the day most users dont care about the technology behind the products that they use, only the direct utility value to them. the miracle of facebook (don't get me wrong, I barely use the damn thing) is that it taught people a new mode of communication that resonated well with them. more importantly, the developers of the service listened to the users and expanded it in ways that accelerated its growth. when the service first launched, it did not even support global group memberships (memberships in groups outside of your individual facebook 'site,' which was of course back when it was still a closed system).
wave is certainly cool but google knows very well that it will be the ability of the developers to find novel applications that everyday users will find fun and exciting to use; applications with strong network effects and possibly ones that do not compete directly with facebook. IMHO, that task (finding something the users will like and then getting them to use it in large numbers) is much more difficult than designing the middle-ware.
Fair enough. I'm not anti-progressive, and I agree that risks should be taken, I was just taken aback by how cavalier everyone was about this particular project. Something that can potentially cause millions of dollars in damages (even based on the 3.4 quake resulting from the previous project) is a very high level of risk for any project to take place on public land. As long as there is sufficient oversight I really don't care; but people's tendency in these posts to dismiss risk as irrelevant really rubs me the wrong way.
and, once those rocks are fractured, water is then supposed to be injected so that it can be turned into steam (the mode through which the power will actually be extracted) which I would imagine would increase the pressure. There is the section of the article where they describe a similar process taking place close to the surface in another part of the state where it has caused a significant increase in seismic activity, all of which is apparently close to the surface. the article suggests that more destructive seismic activity and larger quakes emerge from deeper faults.
yeah, i feel like the pressure is the point of the project and it's not the pressure added by injecting the water, but by the heat coming through the rock turning that water into steam. essentially creating a steam explosion several miles underground in under conditions that probably have more unknowns than knowns, all in an area with known levels of significant instability seems like a prospect deserving of extreme scrutiny.
I'm shocked at how good people feel about this, and how much license many of the commenters are willing to give to this venture even if it is known to possible trigger a significant seismic event. What if this were being funded by Exxon instead of KP + Google? I doubt people would be as dismissive of the risks involved. also, filing for patents on improving the process is the whole point of funding something like this, not building a single power station, but gaining the know-how and experience necessary to scale it and the IP protections necessary to prevent others firms from exploiting the knowledge gained. That's just _at best_ spin; because clearly, if something is patented it has to be safer.
This makes perfect sense from a business perspective and is therefore somewhat inevitable. I think the biggest challenge yet to be solved is how you can easily give online distribution games as gifts. The winter season is by far the period of greatest sales in the games industry for obvious reasons. However, I think that there is a slight hurdle between what we have now, and something that parents can confidently use to buy games for their kids.
In the end, everyone just makes more money using digital distribution. Not only do you not have to finance, press, package, and ship the game; but there is no secondary market to speak of.
I remember when the first WiMax technical specifications were published back in like, 2006. I read them as part of a research project associated with a securities firm to help compile research on the companies involved. Even back then, the prospect of having a broadband network operational at least two to three years in the future with deployments probably (at least, hopefully for the wimax people) continuing for at least five, that could only pump 10MB at a 10kM range seemed absurd. This is a system that at its unrealistic ideal pushes sub-ethernet levels of connectivity.
I know it's a mobile network, and at that, it's very very impressive compared to a lot of what's out there. But honestly the only reason to build a network this large is the hope that you will capture at home users and compete with traditional broadband services (which they fully hope to do in metro areas). How can WiMax possibly hope to compete in those settings against GPON and DOCSIS 3.0/4? Does anyone else think that this has already been cornered by the competition into a niche before it even gets off of the drawing board?
Just like it happened for manufacturing, in our modern globalized economy, the jobs will shift to where they can be most inexpensively performed (ceteris paribus). If someone somewhere else can do your job as well as you can at a price equivalent or lower to your wage when factoring in added expenses for communication and a less efficient team, then your job will leave and you will be unemployed at some point in the future unless you act appropriately.
Thus, the true function of trade unions is revealed. By limiting knowledge that can be obtained by people functioning outside the union, you prevent the circumstance of someone else being able to do your job from arising in the first place. All the tech community has to do is learn not to share knowledge with 'outsiders' (smother the open source movement in the middle of the night with a pillow or something) and it will be set forever with its high wages and superhero attitude. Sadly, there are plenty of people out there who would give quite a lot for the opportunity to work endless amounts of unpaid overtime in a skilled profession (particular types of manufacturing are skilled, yes) if it would mean that they were getting paid for the rest of their work week. From one perspective, fighting back might end up like quicksand and result in the entire community sinking even faster than it would otherwise.
Then again, from another perspective, there is presently a massive shortage of IT workers in the US which is something that many employers are trying to rectify using H1-B visas. Indeed, fear of outsourcing has created a paradox (in that there is now a surplus of jobs, because less people chose a career that they thought would be quickly sent overseas). So, maybe there is room for a union, a professional organization, or something. In the end, there really is no substitute for getting all the team members sitting in the same room at the same location to work together. However,it is a globalized economy, and more and more high tech firms are being started outside the borders of the western world (perhaps, in response to this very pressure) which is an even worse situation as the new jobs are being created overseas as opposed to just being sent there.
But all of this is really irrelevant. The tech community as a whole, as the article points out, has an attitude which is infamously contrary to cooperative endeavors, especially ones aimed at collective benefit in the face of potential individual sacrifice. Oh, and as one last thought, unions probably give preference to seniority, as senior members of unions are employees of companies that cannot be easily dispatched (they often have the greatest amount of accumulated knowledge). If the senior members of the union left the union, it would be very easy to fire the rest of them and replace them with newly trained workers. In the end, there is sadly no substitute for experience, especially experience in a particular company or institution. No amount of self taught-ness will allow you to divine how an undocumented system that was written several years ago by an intern works. But that old guy over there probably knows.
Having watched the video on the chrome page at google, it seems to me that they are blaming the current browsers for effectively holding them and the rest of the web development community back. This is probably fair in a lot of respects, however, google used to do amazing things despite these limitations. Further, there are better ways of doing this (working through standards organizations to push the extensions of specifications which would then be built into mature platforms) as opposed to creating a new platform which will have its own (by design i might add) incompatibilities (or everything else will gain incompatibilities) and require extended maintenance throughout its indefinite browser lifespan. Except for the "we're going to open source it" rhetoric, the launch of this product is being marketed/pitched in the same way that I would imagine microsoft first pushed Internet explorer. Netscape/Java are bad at developing applications/services that run within web browsers that do nifty things. / Come check out our new and improved VBScript, iFrames, and the mother of all "application in a web browser" technologies, ActiveX / Haven't we been through this already and found the results rather hard to swallow?
something i've always had some degree of trouble understanding is the venom with which a lot of people hold onto a particular FOSS license perspective. I have to admit that at certain times it seems as though developers license their code under the GPL at least partially out of spite, as in, "look at this nice code that you totally can't use with your business model." However, this is somewhat fitting, as it seems to me that RMS himself wanted to spite the proprietary operating systems and tools developers of the 70s who forced his brothers in code to sign NDAs and stop sharing their tweaks and improvements (I remember reading somewhere about a printer control program that they couldn't get the source for really frustrating him).
The bottom line is that with the right FOSS license you can really spite commercial interests from 'exploiting' your code to their own ends. Congrats. You're now in an interesting situation. If the commercial interest is well funded, they will develop their own proprietary version of your gadget, and depending on the strength of their marketing and sales teams, they could potentially make their proprietary solution the standard for the market and actually end up cutting out your free solution purely because you forbade them from using it. Or, if they're not well funded, they could just rip you off anyway, committing a crime, and justifying you staying up late at night unpacking firmware binaries and running 'strings' just to ensure that someone is not cheating you out of your hard spent time and effort (like the effort you're then spending to ensure that they're not ripping you off for the effort you made earlier). In the first case, you potentially hurt the community by forcing the introduction of competing proprietary standards. In the latter, you just hurt yourself and small commercial interests who may simply not have any other easy way to become profitable.
I'll admit this is all kind of dramatic. In the end, I too am an avid open source advocate, and have released more software under the GPL than I have under the MIT. But the real deal is that the majority of open source projects are so small and are abandoned so early out of the gate that they are like grains of sand on a beach somewhere. The only thing that their perceived significance might dwarf is their author's litigation budget which matters more than the actual COPYING file included with the package. Unless copyright shifts such that the government takes up the baton in terms of enforcing licenses directly through federal administration (as the record, and tv, and film companies want) that situation wont change, and this discussion will remain largely a moot point.
This might be a stupid question, but does it really actually make sense to run big number crunching applications on top of WINE? I know WINE performance is really quite amazing considering what it is, but at the same time, if you have a job that you expect to run for a long time doing a lot of work (especially a threaded one) repeatedly you should probably re-implement it if you can natively. OpenMP is kind of fun
you're way overstating this. Nobody will deny that true end-to-end QOS is the best solution. However, saying that you can't get any 'true success' with traffic shaping is just bunk! As a network engineer you should appreciate that so many problems are caused by misconfiguration and not any fault with (or the pricetag on) the hardware.
exactly! it's really necessary to do some outgoing traffic shaping just to ensure that you dont ever fill that buffer cause the results really really suck! so much of ensuring a good connection ends up coming down to ensuring that you dont step on any of the grumpy stuff in the middle's toes.
This isnt really true though. TCP/IP as it's designed sends packets at a faster and faster rate until it ends up with loss, and then scales back. You can't control how fast data comes in, but you can control how fast you forward it. Therefore, QOS/Scheduling on the internal interface of any router can be used to scale the incoming traffic on the other side by keeping the connections from becoming too hot in the first place. Can you truly limit the incoming bandwidth of your router? not really. But, you can truly limit the incoming bandwidth of all the hosts on the internal interface of the router, which amounts to the same thing in the end. I would suggest reading through the advanced linux router and traffic control howto. it's a very good document.
Actually, as I stop to think about it, it occurs to me that there are far larger problems here. Benjamin (as in Walter) makes clear in his critique of violence, law is used to replace violence by the state. That is, the state moves to obtaining a monopoly on the use of violence.
Now that we're that far, let's think about how this applies to internet governance. Large scale DOS attacks have been sort of illegal since the 90s (if my memory/understanding serves me correctly). However, it begs the question, can the new Cyber Command legally execute DOS attacks against hosts, domestic or foreign, for purposes that serve to defend the people (state)? Further, CAN-SPAM makes it illegal to forge mail headers, but can the federal government legally forge mail headers in the execution of an investigation (such as sting operations)? How much authority have we already given to the government without our consent?
Finally, if legislation passes that massively increases the scope of the government's monitoring abilities, then we are suggesting that the government has a right to monitor, and that we do not have a right to monitor (as invasion of privacy is a form of violence) potentially making the use of things like packet sniffers illegal under all but the most unusual of circumstances. Finally, we have to ask the question, if ISPs push back against government monitoring efforts, what recourse does the government then have? If the government needed to put a police station in a 'dangerous neighborhood,' could they use eminent domain to acquire property to do so? Could the government use eminent domain to... lets say... buy a peering facility to facilitate traffic monitoring?
What annoys me about this article is the glorification of the system that they have created. P2P networks use open protocols which are exceedingly well documented. Almost every bittorrent client that I've ever seen has a specific window with which to view the peers you are connected to. Literally, someone could copy and paste code from an OSS project, couple it with a dictionary of keywords, and then end up with this system. It's infuriating that someone could be taking so much credit for it. Finally, and most egregiously, they fail to point out that this entire system is ineffective in monitoring people with even the least bit of intelligence (you would imagine if you were transferring one of the only types of illegal content on the internet, that you would use an encrypted session), further indicating that this system is probably just a totally behind the scenes funded starting point for the enforcement of copyright through active federal monitoring of open P2P systems.
Sometimes it's just really not a good idea to push a piece of software out to hundreds of millions of people on its first release just because they use/update your other products. This is the real way that it could come back and bite them, and it certainly seems to have.
Slashdot Mirror
of course, at the end of the day, by the numbers, google is an advertising sales company not a software company. that's probably a little unfair, but it's necessary to point out that there are many many projects at google that do not really earn significant income, and only a few that really sustain the vast majority of their cash flow. that is not to say that many of the other things they work on are not cool and useful; quite the opposite in fact. we should not discourage anyone from building something of incredible usefulness (which I think facebook has de facto just based on the amount of traffic it gets) just because we don't yet know where to attach the meter.
but at the end of the day most users dont care about the technology behind the products that they use, only the direct utility value to them. the miracle of facebook (don't get me wrong, I barely use the damn thing) is that it taught people a new mode of communication that resonated well with them. more importantly, the developers of the service listened to the users and expanded it in ways that accelerated its growth. when the service first launched, it did not even support global group memberships (memberships in groups outside of your individual facebook 'site,' which was of course back when it was still a closed system).
wave is certainly cool but google knows very well that it will be the ability of the developers to find novel applications that everyday users will find fun and exciting to use; applications with strong network effects and possibly ones that do not compete directly with facebook. IMHO, that task (finding something the users will like and then getting them to use it in large numbers) is much more difficult than designing the middle-ware.
"one vision to rule them all... one vision to find them... one vision to bring them all and in the ddddaaaarrrrkkkneeesssss BIND THEM"
Fair enough. I'm not anti-progressive, and I agree that risks should be taken, I was just taken aback by how cavalier everyone was about this particular project. Something that can potentially cause millions of dollars in damages (even based on the 3.4 quake resulting from the previous project) is a very high level of risk for any project to take place on public land. As long as there is sufficient oversight I really don't care; but people's tendency in these posts to dismiss risk as irrelevant really rubs me the wrong way.
Well reasoned.
and, once those rocks are fractured, water is then supposed to be injected so that it can be turned into steam (the mode through which the power will actually be extracted) which I would imagine would increase the pressure. There is the section of the article where they describe a similar process taking place close to the surface in another part of the state where it has caused a significant increase in seismic activity, all of which is apparently close to the surface. the article suggests that more destructive seismic activity and larger quakes emerge from deeper faults.
yeah, i feel like the pressure is the point of the project and it's not the pressure added by injecting the water, but by the heat coming through the rock turning that water into steam. essentially creating a steam explosion several miles underground in under conditions that probably have more unknowns than knowns, all in an area with known levels of significant instability seems like a prospect deserving of extreme scrutiny.
I'm shocked at how good people feel about this, and how much license many of the commenters are willing to give to this venture even if it is known to possible trigger a significant seismic event. What if this were being funded by Exxon instead of KP + Google? I doubt people would be as dismissive of the risks involved. also, filing for patents on improving the process is the whole point of funding something like this, not building a single power station, but gaining the know-how and experience necessary to scale it and the IP protections necessary to prevent others firms from exploiting the knowledge gained. That's just _at best_ spin; because clearly, if something is patented it has to be safer.
This makes perfect sense from a business perspective and is therefore somewhat inevitable. I think the biggest challenge yet to be solved is how you can easily give online distribution games as gifts. The winter season is by far the period of greatest sales in the games industry for obvious reasons. However, I think that there is a slight hurdle between what we have now, and something that parents can confidently use to buy games for their kids.
In the end, everyone just makes more money using digital distribution. Not only do you not have to finance, press, package, and ship the game; but there is no secondary market to speak of.
I remember when the first WiMax technical specifications were published back in like, 2006. I read them as part of a research project associated with a securities firm to help compile research on the companies involved. Even back then, the prospect of having a broadband network operational at least two to three years in the future with deployments probably (at least, hopefully for the wimax people) continuing for at least five, that could only pump 10MB at a 10kM range seemed absurd. This is a system that at its unrealistic ideal pushes sub-ethernet levels of connectivity.
I know it's a mobile network, and at that, it's very very impressive compared to a lot of what's out there. But honestly the only reason to build a network this large is the hope that you will capture at home users and compete with traditional broadband services (which they fully hope to do in metro areas). How can WiMax possibly hope to compete in those settings against GPON and DOCSIS 3.0/4? Does anyone else think that this has already been cornered by the competition into a niche before it even gets off of the drawing board?
"government warehouse!"
I see them in gas stations all over the place these days
Thus, the true function of trade unions is revealed. By limiting knowledge that can be obtained by people functioning outside the union, you prevent the circumstance of someone else being able to do your job from arising in the first place. All the tech community has to do is learn not to share knowledge with 'outsiders' (smother the open source movement in the middle of the night with a pillow or something) and it will be set forever with its high wages and superhero attitude. Sadly, there are plenty of people out there who would give quite a lot for the opportunity to work endless amounts of unpaid overtime in a skilled profession (particular types of manufacturing are skilled, yes) if it would mean that they were getting paid for the rest of their work week. From one perspective, fighting back might end up like quicksand and result in the entire community sinking even faster than it would otherwise.
Then again, from another perspective, there is presently a massive shortage of IT workers in the US which is something that many employers are trying to rectify using H1-B visas. Indeed, fear of outsourcing has created a paradox (in that there is now a surplus of jobs, because less people chose a career that they thought would be quickly sent overseas). So, maybe there is room for a union, a professional organization, or something. In the end, there really is no substitute for getting all the team members sitting in the same room at the same location to work together. However,it is a globalized economy, and more and more high tech firms are being started outside the borders of the western world (perhaps, in response to this very pressure) which is an even worse situation as the new jobs are being created overseas as opposed to just being sent there.
But all of this is really irrelevant. The tech community as a whole, as the article points out, has an attitude which is infamously contrary to cooperative endeavors, especially ones aimed at collective benefit in the face of potential individual sacrifice. Oh, and as one last thought, unions probably give preference to seniority, as senior members of unions are employees of companies that cannot be easily dispatched (they often have the greatest amount of accumulated knowledge). If the senior members of the union left the union, it would be very easy to fire the rest of them and replace them with newly trained workers. In the end, there is sadly no substitute for experience, especially experience in a particular company or institution. No amount of self taught-ness will allow you to divine how an undocumented system that was written several years ago by an intern works. But that old guy over there probably knows.
Having watched the video on the chrome page at google, it seems to me that they are blaming the current browsers for effectively holding them and the rest of the web development community back. This is probably fair in a lot of respects, however, google used to do amazing things despite these limitations. Further, there are better ways of doing this (working through standards organizations to push the extensions of specifications which would then be built into mature platforms) as opposed to creating a new platform which will have its own (by design i might add) incompatibilities (or everything else will gain incompatibilities) and require extended maintenance throughout its indefinite browser lifespan. Except for the "we're going to open source it" rhetoric, the launch of this product is being marketed/pitched in the same way that I would imagine microsoft first pushed Internet explorer. Netscape/Java are bad at developing applications/services that run within web browsers that do nifty things. / Come check out our new and improved VBScript, iFrames, and the mother of all "application in a web browser" technologies, ActiveX / Haven't we been through this already and found the results rather hard to swallow?
something i've always had some degree of trouble understanding is the venom with which a lot of people hold onto a particular FOSS license perspective. I have to admit that at certain times it seems as though developers license their code under the GPL at least partially out of spite, as in, "look at this nice code that you totally can't use with your business model." However, this is somewhat fitting, as it seems to me that RMS himself wanted to spite the proprietary operating systems and tools developers of the 70s who forced his brothers in code to sign NDAs and stop sharing their tweaks and improvements (I remember reading somewhere about a printer control program that they couldn't get the source for really frustrating him).
The bottom line is that with the right FOSS license you can really spite commercial interests from 'exploiting' your code to their own ends. Congrats. You're now in an interesting situation. If the commercial interest is well funded, they will develop their own proprietary version of your gadget, and depending on the strength of their marketing and sales teams, they could potentially make their proprietary solution the standard for the market and actually end up cutting out your free solution purely because you forbade them from using it. Or, if they're not well funded, they could just rip you off anyway, committing a crime, and justifying you staying up late at night unpacking firmware binaries and running 'strings' just to ensure that someone is not cheating you out of your hard spent time and effort (like the effort you're then spending to ensure that they're not ripping you off for the effort you made earlier). In the first case, you potentially hurt the community by forcing the introduction of competing proprietary standards. In the latter, you just hurt yourself and small commercial interests who may simply not have any other easy way to become profitable.
I'll admit this is all kind of dramatic. In the end, I too am an avid open source advocate, and have released more software under the GPL than I have under the MIT. But the real deal is that the majority of open source projects are so small and are abandoned so early out of the gate that they are like grains of sand on a beach somewhere. The only thing that their perceived significance might dwarf is their author's litigation budget which matters more than the actual COPYING file included with the package. Unless copyright shifts such that the government takes up the baton in terms of enforcing licenses directly through federal administration (as the record, and tv, and film companies want) that situation wont change, and this discussion will remain largely a moot point.
to be honest, that's pretty sweet =)
This might be a stupid question, but does it really actually make sense to run big number crunching applications on top of WINE? I know WINE performance is really quite amazing considering what it is, but at the same time, if you have a job that you expect to run for a long time doing a lot of work (especially a threaded one) repeatedly you should probably re-implement it if you can natively. OpenMP is kind of fun
For your standard persistent terminals, SCREEN is really your best bet
you're way overstating this. Nobody will deny that true end-to-end QOS is the best solution. However, saying that you can't get any 'true success' with traffic shaping is just bunk! As a network engineer you should appreciate that so many problems are caused by misconfiguration and not any fault with (or the pricetag on) the hardware.
exactly! it's really necessary to do some outgoing traffic shaping just to ensure that you dont ever fill that buffer cause the results really really suck! so much of ensuring a good connection ends up coming down to ensuring that you dont step on any of the grumpy stuff in the middle's toes.
This isnt really true though. TCP/IP as it's designed sends packets at a faster and faster rate until it ends up with loss, and then scales back. You can't control how fast data comes in, but you can control how fast you forward it. Therefore, QOS/Scheduling on the internal interface of any router can be used to scale the incoming traffic on the other side by keeping the connections from becoming too hot in the first place. Can you truly limit the incoming bandwidth of your router? not really. But, you can truly limit the incoming bandwidth of all the hosts on the internal interface of the router, which amounts to the same thing in the end. I would suggest reading through the advanced linux router and traffic control howto. it's a very good document.
A man who chooses to represent himself has a fool for a client.
Actually, as I stop to think about it, it occurs to me that there are far larger problems here. Benjamin (as in Walter) makes clear in his critique of violence, law is used to replace violence by the state. That is, the state moves to obtaining a monopoly on the use of violence.
Now that we're that far, let's think about how this applies to internet governance. Large scale DOS attacks have been sort of illegal since the 90s (if my memory/understanding serves me correctly). However, it begs the question, can the new Cyber Command legally execute DOS attacks against hosts, domestic or foreign, for purposes that serve to defend the people (state)? Further, CAN-SPAM makes it illegal to forge mail headers, but can the federal government legally forge mail headers in the execution of an investigation (such as sting operations)? How much authority have we already given to the government without our consent?
Finally, if legislation passes that massively increases the scope of the government's monitoring abilities, then we are suggesting that the government has a right to monitor, and that we do not have a right to monitor (as invasion of privacy is a form of violence) potentially making the use of things like packet sniffers illegal under all but the most unusual of circumstances. Finally, we have to ask the question, if ISPs push back against government monitoring efforts, what recourse does the government then have? If the government needed to put a police station in a 'dangerous neighborhood,' could they use eminent domain to acquire property to do so? Could the government use eminent domain to... lets say... buy a peering facility to facilitate traffic monitoring?
What annoys me about this article is the glorification of the system that they have created. P2P networks use open protocols which are exceedingly well documented. Almost every bittorrent client that I've ever seen has a specific window with which to view the peers you are connected to. Literally, someone could copy and paste code from an OSS project, couple it with a dictionary of keywords, and then end up with this system. It's infuriating that someone could be taking so much credit for it. Finally, and most egregiously, they fail to point out that this entire system is ineffective in monitoring people with even the least bit of intelligence (you would imagine if you were transferring one of the only types of illegal content on the internet, that you would use an encrypted session), further indicating that this system is probably just a totally behind the scenes funded starting point for the enforcement of copyright through active federal monitoring of open P2P systems.
Sometimes it's just really not a good idea to push a piece of software out to hundreds of millions of people on its first release just because they use/update your other products. This is the real way that it could come back and bite them, and it certainly seems to have.