With decent airflow (no windows left, building is hundreds of feet tall with no horizontal obstructions at that level), and enough temperature, virtually anything will burn. The airflow and temperature are competing factors to some extent, but I'm sure that fire got quite hot.
It isn't like they mail the prefix to you and have you hand-type it. They're handed out by NDP, which effectively works the same way as DHCP - the ISP will give you a prefix.
Even if they don't need to change prefixes, they still have incentive to do so. If you want a static prefix you will have to pay them more. If you don't care, they can give you a new one frequently.
That is what most ISPs do today with IPv4. Why wouldn't they do the same with IPv6? There is no technical reason why ISPs couldn't give all their customers static IPs (via DHCP), and just renumber them when there is a need to redo their network. They just prefer not to do it this way.
Reclaiming all those class As would really only delay IPv4 exhaustion by a few months. That was true 10 years ago. That was true 20 years ago. That will be true 30 years from now, when we're all on IPv8.
It wasn't reclaiming class A space that got us 10 years, but rather NAT and the existing pool lasting as long as it did.
Of course, the only purpose of the APAP in this medication is to kill the patient if they dare to abuse it.
There is no reason that the two couldn't be unbundled so that the doctor and patient could work together to ensure that just the right amount of each is administered (likely zero on the APAP front, but the doctor can prescribe it if appropriate). This is just prohibition-style thinking, just like sticking strychnine in industrial alcohol.
Liability is handled as well, since hospitals (and even their in-house pharmacies in particular) often have to insure against potential drug-related liabilities anyway (misdiagnosis, mislabeling, adverse drug interactions, improper storage procedures, etc).
There are a lot of things that can go wrong when manufacturing a pill beyond what goes wrong when you administer a pill. When you make an allergy pill after having made a super-strong heart medication, will the recipient of the allergy pill get their heart slowed down as an added bonus due to cross-contamination?
Also, there are also a lot more storage restrictions on the ingredients that go into pills than on the pills themselves. The pills contain antioxidants, preservatives, antimicrobials, and so on, and are tested as formulated to withstand variations in storage conditions similar to what a typical consumer would subject them to. The raw materials are often expected to be stored in very controlled and ideal conditions (not exposed to air, protected from light, kept at a specific temperature (which might be -70C), kept at a specific humidity (not too dry, not too wet), and so on. Since those raw materials cost millions to make at manufacturing scale anyway, the cost to store or even ship them at optimal conditions isn't a huge expense.
Then when manufacturing equipment is done being used and is cleaned out there is often analytical testing done to demonstrate that no residue remains on the equipment. This is done EVERY time the equipment is used. The tablets that are manufactured are also sampled and tested to ensure proper quality. Some of those tests can take a month to perform (like sterility testing - filter the IV fluid through a filter, then put the filter in growth media, and give anything that was caught by the filter plenty of time to grow).
I'm not saying these are insurmountable obstacles. However, printing tablets is unlikely to be just like loading up your inject printer with a few cartridges and hitting print. Compounding pharmacies today are being scrutinized very closely because there have been some high-profile lapses in quality.
This article was remarkably sparse on details, but I suspect that this is part of some huge manufacturing process analogous to 3D printing, done on an industrial scale. So, the tablets might be built in layers 10,000 at a time over a timespan of a minute or two, or something like that.
It seems like this short article is really two shorter articles in one. There is the specific product approval, which I suspect is for large scale manufacture (but I'm not certain of that), and then some general speculation about some in-the-hospital technology.
Also, the summary speaks of making tablets identical so that patients aren't confused. That seems a bit odd to me since the whole point of making pills look different is so that patients aren't confused. I know somebody who is on numerous medications and when they spill their pill organizer it is quite a task to try to sort things out and get them right again, since pills are hard enough to tell apart as it is.
You could argue that this technology would let you make 1 super-pill tailored to the individual that they could take once a day and not have to take 14 different medications. That makes sense on paper, but in reality from talking to people who take lots of pills many of them end up tweaking their schedules. They'll take some pills before eating, others after, some at different times, they might split some and take them at multiple times instead of once. While I'm sure there isn't clinical data to back this up, when it comes down to individuals sometimes there is a need for flexibility. True, many patients abuse this flexibility to their detriment, but if these patients were given the choice of a take-it-or-leave-it super pill I suspect most would take it, and the first time they get a stomach ache or a leg cramp or whatever they'd just stop taking it.
Just make the specs vague, self-contradictory, and long. Very, very, long.
When you get to software on this scale just writing the specs is an insanely difficult problem. Most big systems fail due to poor requirements, and I think that is because good requirements are a LOT harder to get right than people appreciate, especially for waterfall-style projects that are always the favorites of big RFP projects like these.
You really shouldn't be using NAT with IPv6. The idea is one machine, one address. Given DNS I'm not sure why you would want fixed fully external IPs.
Ok, so how do I set up my internal DNS server so that everything works fine when my ISP gives me a new prefix every 24 hours, or every time the router is rebooted, or every time they feel like changing my prefix?
The advantage of private addresses is that they're handed out by a DHCP server that you can control, not by your ISP.
Again, I think IPv6 users today are spoiled by tunnel brokers and the like who are giving out static prefixes. I don't think that the likes of Comcast are going to be going that route when they finally embrace IPv6 fully. If nothing else a static prefix is something they can sell, and you could also argue that not having static prefixes makes life easier for them if they want to change their internal network topology.
For residential users who like to know where their devices are... that one is a bit tougher. Really DNS is your friend. How often do you change ISPs anyway?
How exactly do you set up internal DNS when all your IPs are subject to change?
You don't have to change ISPs to have your IPv6 prefix change. Your ISP need only assign you a new one. It is in their interest to do so often so that you have an incentive to pay more for a non-residential connection, and also because it gives them more flexibility in renumbering.
By the time you finished with all the lawsuits/etc the addresses wouldn't be useful. We need to get past IPv4. Putting it off for another few months won't help that.
That said, there are a lot of things about IPv6 that are rather annoying. There aren't really a lot of good DHCP options if you want to use NAT, and if you don't want to use NAT then anytime your router prefix changes the external IPs of all the hosts on the network change. That is a fairly big change from how things work today, and I think most early adopters don't notice because they tend to have static IPs, but that is unlikely to be the case once it is mass-adopted.
So, my sense is that Windows 10 is the "odd-numbered" version that I'll eventually end up upgrading to. Right now I do most of my serious stuff on linux, and maintain a decent windows 7 PC mainly for gaming purposes (since I still have many games that are not linux-compatible). I tend to view my windows PC more like a gaming console as a result.
So, when is the right time to upgrade? I suspect that it will be once DirectX 12 is available, stable, and in some kind of use. At the very least I'll want to wait until the rumor mill indicates that my graphics drivers (Radeon) are reasonably stable on Win10.
I'm also torn on whether to upgrade in place or just wipe and reinstall (especially since I'm still running my OS from the non-SSD drive in the system). Is installing Win10 from scratch supported, and free?
My guess is that it will be at least a few months before I'll be upgrading...
Well, ultimately you can do whatever you can get away with. Build some oil platforms, and what is anybody going to do about it? Of course, others could also start building platforms as well in competition, and then the Chinese are left with the same dilemma. The only way to truly enforce exclusivity is to either go to war or start imposing economic sanctions/etc, but if the US were to start putting sanctions on China that could get messy fast with all the trade.
Nobody really wants to get into a shooting war over this stuff.
A compromise on the length argument is a registration fee, that is exponential. So, having copyright for a year might be a buck or two. Having it for 5 years might be a few hundred. Having it for 10 might be a few million. Having it for 15 might start to go into the billions range. That can obviously be tuned and tweaked. The point is that copyright is a public sacrifice and should be done for public benefit. Companies can decide how valuable that renewal is and let it go or buy in.
Many SaaS vendors are moving towards new generation of logins. I see many vendors removing OpenID in general and we're seeing an equally high number of companies embracing SSO.
Ugh, unfortunately the SaaS vendor I'm working with right now isn't one of them.
OAuth2? Nope. Another password to remember/reset/etc.
Webservices? Nope. Drop a file onto an FTP site which is polled, and poll the site yourself.
XML? Sort-of - it is their least-preferred file format which they try to avoid at all costs. Oh yeah, they have failures to parse xml files that W3 validates (for syntax, not semantics). I'm sure that there are issues with their non-xml-file parsers for the majority of the files we're dealing with since they avoid xml, but since you can't just use validators/etc to check the files we just run into them from time to time in production.
But, hey, our internal IT group is just as brain dead as they seem to also be writing their own XML parsers judging by some of the failures I've seen.
Password managers are essentially making a bet that the risk of your hard drive being compromised is far less likely than a website being compromised.
If your hard drive is compromised then your keystrokes are being logged and your cookies are being extracted, and any website you log into will be compromised. The password manager isn't really adding that much more risk here.
I somewhat agree, but there is a tragedy of the commons element to this. I do think that it makes sense to encourage reciprocation, via things like cross-licensing. By all means make it free to countries that have low incomes and so on.
I feel similarly about drug patents. I think that governments should start doing end-to-end R&D and keep the patents. They should be licensed for free to any domestic manufacturer, to manufacturers in countries that make similar investments and reciprocate, and for local domestic consumption in the third world. That would make drugs dirt cheap (due to the tax-based funding and competition for manufacture), but also encourage other first-world countries to make similar investments which increases the value of the public domain.
When you just force countries that invest to make their content freely, then you encourage freeloaders.
And I'm fine with "cross-licensing" being cross-domain. Maybe one country makes lots of free movies, and another country has really amazing free museums with subsidized hotel rates, so the two countries make their services available to the citizens of the other.
IMHO, one of the remaining hurdles to us getting past the Great Filter is the proliferation of technology and doomsday weaponry to all corners of the globe.
Honestly, I really only see the solution to this problem being the proliferation of humans to such an extent that warfare using the most powerful weapons available is not a threat to a substantial portion of the human race. If people colonized half the galaxy a nuclear war would take centuries to reach everybody simply due to the speed of light, to say nothing of effective countermeasures.
Obviously that isn't going to happen anytime soon.
The problem is that it is very difficult to put the genie back in the bottle. The design of a nuclear weapon is basically just information, and everybody can see just how hard it is to keep information under wraps. Over time we've seen increasingly more dangerous weapons coming into the hands of the general public. Maybe if the entirety of the human population could be kept under surveillance we could completely prevent the proliferation of such technology, but simply having an organization capable of such a feat is in some sense an existential threat of a different sort.
Furthermore, refueling an airplane does not take less than two minutes. That's the time you need just to plug in the fuel line.
He said, "But it would make loading and unloading the plane a matter of two minutes or something which is less time than it takes to refuel the plane."
Really, Hacking Team was just doing things the way a very small segment of society which currently holds most financial capital thinks everybody should be operating.
FTFY - SOPA, TPP, etc.. are not products of the Software industry. I am pretty sure I agree with your point under the surface, but the generalization is plain wrong.
They are certainly the way the software industry thinks everybody should be operating, which is all I claimed. I did not claim that all of those laws/treaties/etc were products of the software industry. I'm not sure how you can claim that the Uniform Computer Information Transactions Act wasn't though.
Simple access doesn't count as distribution in this case. The GPLv2 applies to code distributed under the GPLv2. It isn't a property of the code itself, and the fact that you have GPLv2ed code doesn't mean you have to give it to me, nor do I get the right to the code by simply having access. If you deliberately give me the code, you have to do so under the GPLv2, and I have all the rights that grants.
You give your employee the code when you give them access to it. Before they couldn't see the code. Now they can. They gave it to you.
The FSF holds that having employees work on company code isn't distribution.
That's nice, but they aren't the authors of the code in question, even if they're the authors of the license. If the kernel authors intended the code to be copyable by the employee and the license says that it is, then it is.
Consider that I've got a lot of company-owned proprietary code on my work computer. If that counted as distributing it to me, I'd own one copy of the code.
Well, you do have one copy of the code in your possession - the one on the server. That doesn't mean that you can make another copy of the code without permission from the copyright holder.
Similarly, putting copy of GPLv2ed code on a work computer is not distribution, and the programmer doesn't have a copy under GPLv2.
The wording of the GPLv2 isn't terribly helpful here. It uses phrases like "copy or distribute" in some places and "distribute" in others and doesn't define either term. If nothing else it really seems open to litigation. It would be sensible to make stuff like this explicit in a license.
It would be an interesting case - no doubt an expensive one if there were deep enough pockets involved. Ultimately the law is whatever the courts say it is, and there is only one way to find out...
But there's another aspect of this. Say my company downloads the Linux kernel and we internally make some changes to it and use it on our servers in its modified form. Jim is one of the coders. Linux is released under GPLv2. Does that mean that Jim can take our changes home with him?
No.
The GPLv2 kicks in only when the company redistributes the code along with the modifications, and those modifications are available to the recipients that we've specified.
This is a common argument but I'm not convinced that it is airtight. How about this:
But there's another aspect of this. Say my company buys a Windows DVD and we install it on 47 of our servers. Jim is one of the coders. Windows DVDs are not licensed for multiple installations. Does that mean that Jim can call up MS and pocket a reward?
No.
The Windows License kicks in only when the company redistributes copies of Windows, and those copies are available to the recipients that we've specified.
The problem with this argument is that copyright applies anytime you make a copy of anything. Copyright says you can't install Linux anywhere. What lets you install it is the license. The license for Windows says you aren't allowed to copy it at all except to install it once. The license for Linux says you're only allowed to copy it if the copy is GPLv2 along with any modifications you've made, and you make the source available to any recipients. So, your modifications MUST be GPLv2 the moment they're installed with the kernel. Since they are GPLv2, you've given permission to anybody with access to them to redistribute them already, and that includes Jim.
Now, you can still tell Jim not to redistribute it and fire him if he does and blacklist him so that his kids don't have any kind of future as is fashionable in the US, and I'm not sure that a court would take issue with that. However, I'm not convinced that a court would hold up an employer's right to sue an employee for redistributing Linux kernel modifications if you could argue that a kernel was copied anywhere at all. Keep in mind that the software industry has been pushing for super-conservative interpretations of copyright such that simply loading a copy of a portion of the program into RAM to execute it is covered. Under that kind of interpretation (which is the basis of any kind of end-use licensing) you're copying the kernel the moment you open the source in an editor.
But, you do expose the general weakness with GPLv2 in that it was really designed with a world of software sold in boxes (ie the 90s). The wording really needed cleanup and licenses like GPLv3 and AGPLv3 do much of that.
And the original point stands... have you audited the source code? Have you compiled / built the product from the source code? If not, then have the alleged source code isn't any better than trusting the Windows source code.
I have done both. So, I guess you'd say I'm fine then?
Most people who use FOSS audit it in the same way that any major corporation does due diligence for stuff they buy. They make sure that people they trust are looking into it.
Sure, you can sneak stuff into reputably-distributed FOSS just like you can sneak stuff into Windows CDs. The fact is that there probably still are a lot more people checking the integrity of something like RHEL than Windows, simply by virtue of it being FOSS.
Slashdot Mirror
With decent airflow (no windows left, building is hundreds of feet tall with no horizontal obstructions at that level), and enough temperature, virtually anything will burn. The airflow and temperature are competing factors to some extent, but I'm sure that fire got quite hot.
It isn't like they mail the prefix to you and have you hand-type it. They're handed out by NDP, which effectively works the same way as DHCP - the ISP will give you a prefix.
Even if they don't need to change prefixes, they still have incentive to do so. If you want a static prefix you will have to pay them more. If you don't care, they can give you a new one frequently.
That is what most ISPs do today with IPv4. Why wouldn't they do the same with IPv6? There is no technical reason why ISPs couldn't give all their customers static IPs (via DHCP), and just renumber them when there is a need to redo their network. They just prefer not to do it this way.
Reclaiming all those class As would really only delay IPv4 exhaustion by a few months. That was true 10 years ago. That was true 20 years ago. That will be true 30 years from now, when we're all on IPv8.
It wasn't reclaiming class A space that got us 10 years, but rather NAT and the existing pool lasting as long as it did.
Of course, the only purpose of the APAP in this medication is to kill the patient if they dare to abuse it.
There is no reason that the two couldn't be unbundled so that the doctor and patient could work together to ensure that just the right amount of each is administered (likely zero on the APAP front, but the doctor can prescribe it if appropriate). This is just prohibition-style thinking, just like sticking strychnine in industrial alcohol.
Liability is handled as well, since hospitals (and even their in-house pharmacies in particular) often have to insure against potential drug-related liabilities anyway (misdiagnosis, mislabeling, adverse drug interactions, improper storage procedures, etc).
There are a lot of things that can go wrong when manufacturing a pill beyond what goes wrong when you administer a pill. When you make an allergy pill after having made a super-strong heart medication, will the recipient of the allergy pill get their heart slowed down as an added bonus due to cross-contamination?
Also, there are also a lot more storage restrictions on the ingredients that go into pills than on the pills themselves. The pills contain antioxidants, preservatives, antimicrobials, and so on, and are tested as formulated to withstand variations in storage conditions similar to what a typical consumer would subject them to. The raw materials are often expected to be stored in very controlled and ideal conditions (not exposed to air, protected from light, kept at a specific temperature (which might be -70C), kept at a specific humidity (not too dry, not too wet), and so on. Since those raw materials cost millions to make at manufacturing scale anyway, the cost to store or even ship them at optimal conditions isn't a huge expense.
Then when manufacturing equipment is done being used and is cleaned out there is often analytical testing done to demonstrate that no residue remains on the equipment. This is done EVERY time the equipment is used. The tablets that are manufactured are also sampled and tested to ensure proper quality. Some of those tests can take a month to perform (like sterility testing - filter the IV fluid through a filter, then put the filter in growth media, and give anything that was caught by the filter plenty of time to grow).
I'm not saying these are insurmountable obstacles. However, printing tablets is unlikely to be just like loading up your inject printer with a few cartridges and hitting print. Compounding pharmacies today are being scrutinized very closely because there have been some high-profile lapses in quality.
This article was remarkably sparse on details, but I suspect that this is part of some huge manufacturing process analogous to 3D printing, done on an industrial scale. So, the tablets might be built in layers 10,000 at a time over a timespan of a minute or two, or something like that.
It seems like this short article is really two shorter articles in one. There is the specific product approval, which I suspect is for large scale manufacture (but I'm not certain of that), and then some general speculation about some in-the-hospital technology.
Also, the summary speaks of making tablets identical so that patients aren't confused. That seems a bit odd to me since the whole point of making pills look different is so that patients aren't confused. I know somebody who is on numerous medications and when they spill their pill organizer it is quite a task to try to sort things out and get them right again, since pills are hard enough to tell apart as it is.
You could argue that this technology would let you make 1 super-pill tailored to the individual that they could take once a day and not have to take 14 different medications. That makes sense on paper, but in reality from talking to people who take lots of pills many of them end up tweaking their schedules. They'll take some pills before eating, others after, some at different times, they might split some and take them at multiple times instead of once. While I'm sure there isn't clinical data to back this up, when it comes down to individuals sometimes there is a need for flexibility. True, many patients abuse this flexibility to their detriment, but if these patients were given the choice of a take-it-or-leave-it super pill I suspect most would take it, and the first time they get a stomach ache or a leg cramp or whatever they'd just stop taking it.
Just make the specs vague, self-contradictory, and long. Very, very, long.
When you get to software on this scale just writing the specs is an insanely difficult problem. Most big systems fail due to poor requirements, and I think that is because good requirements are a LOT harder to get right than people appreciate, especially for waterfall-style projects that are always the favorites of big RFP projects like these.
You really shouldn't be using NAT with IPv6. The idea is one machine, one address. Given DNS I'm not sure why you would want fixed fully external IPs.
Ok, so how do I set up my internal DNS server so that everything works fine when my ISP gives me a new prefix every 24 hours, or every time the router is rebooted, or every time they feel like changing my prefix?
The advantage of private addresses is that they're handed out by a DHCP server that you can control, not by your ISP.
Again, I think IPv6 users today are spoiled by tunnel brokers and the like who are giving out static prefixes. I don't think that the likes of Comcast are going to be going that route when they finally embrace IPv6 fully. If nothing else a static prefix is something they can sell, and you could also argue that not having static prefixes makes life easier for them if they want to change their internal network topology.
For residential users who like to know where their devices are... that one is a bit tougher. Really DNS is your friend. How often do you change ISPs anyway?
How exactly do you set up internal DNS when all your IPs are subject to change?
You don't have to change ISPs to have your IPv6 prefix change. Your ISP need only assign you a new one. It is in their interest to do so often so that you have an incentive to pay more for a non-residential connection, and also because it gives them more flexibility in renumbering.
Heh, sounds like Windows 3.1 or VMS. :)
By the time you finished with all the lawsuits/etc the addresses wouldn't be useful. We need to get past IPv4. Putting it off for another few months won't help that.
That said, there are a lot of things about IPv6 that are rather annoying. There aren't really a lot of good DHCP options if you want to use NAT, and if you don't want to use NAT then anytime your router prefix changes the external IPs of all the hosts on the network change. That is a fairly big change from how things work today, and I think most early adopters don't notice because they tend to have static IPs, but that is unlikely to be the case once it is mass-adopted.
So, my sense is that Windows 10 is the "odd-numbered" version that I'll eventually end up upgrading to. Right now I do most of my serious stuff on linux, and maintain a decent windows 7 PC mainly for gaming purposes (since I still have many games that are not linux-compatible). I tend to view my windows PC more like a gaming console as a result.
So, when is the right time to upgrade? I suspect that it will be once DirectX 12 is available, stable, and in some kind of use. At the very least I'll want to wait until the rumor mill indicates that my graphics drivers (Radeon) are reasonably stable on Win10.
I'm also torn on whether to upgrade in place or just wipe and reinstall (especially since I'm still running my OS from the non-SSD drive in the system). Is installing Win10 from scratch supported, and free?
My guess is that it will be at least a few months before I'll be upgrading...
Well, ultimately you can do whatever you can get away with. Build some oil platforms, and what is anybody going to do about it? Of course, others could also start building platforms as well in competition, and then the Chinese are left with the same dilemma. The only way to truly enforce exclusivity is to either go to war or start imposing economic sanctions/etc, but if the US were to start putting sanctions on China that could get messy fast with all the trade.
Nobody really wants to get into a shooting war over this stuff.
A compromise on the length argument is a registration fee, that is exponential. So, having copyright for a year might be a buck or two. Having it for 5 years might be a few hundred. Having it for 10 might be a few million. Having it for 15 might start to go into the billions range. That can obviously be tuned and tweaked. The point is that copyright is a public sacrifice and should be done for public benefit. Companies can decide how valuable that renewal is and let it go or buy in.
Many SaaS vendors are moving towards new generation of logins. I see many vendors removing OpenID in general and we're seeing an equally high number of companies embracing SSO.
Ugh, unfortunately the SaaS vendor I'm working with right now isn't one of them.
OAuth2? Nope. Another password to remember/reset/etc.
Webservices? Nope. Drop a file onto an FTP site which is polled, and poll the site yourself.
XML? Sort-of - it is their least-preferred file format which they try to avoid at all costs. Oh yeah, they have failures to parse xml files that W3 validates (for syntax, not semantics). I'm sure that there are issues with their non-xml-file parsers for the majority of the files we're dealing with since they avoid xml, but since you can't just use validators/etc to check the files we just run into them from time to time in production.
But, hey, our internal IT group is just as brain dead as they seem to also be writing their own XML parsers judging by some of the failures I've seen.
Password managers are essentially making a bet that the risk of your hard drive being compromised is far less likely than a website being compromised.
If your hard drive is compromised then your keystrokes are being logged and your cookies are being extracted, and any website you log into will be compromised. The password manager isn't really adding that much more risk here.
I somewhat agree, but there is a tragedy of the commons element to this. I do think that it makes sense to encourage reciprocation, via things like cross-licensing. By all means make it free to countries that have low incomes and so on.
I feel similarly about drug patents. I think that governments should start doing end-to-end R&D and keep the patents. They should be licensed for free to any domestic manufacturer, to manufacturers in countries that make similar investments and reciprocate, and for local domestic consumption in the third world. That would make drugs dirt cheap (due to the tax-based funding and competition for manufacture), but also encourage other first-world countries to make similar investments which increases the value of the public domain.
When you just force countries that invest to make their content freely, then you encourage freeloaders.
And I'm fine with "cross-licensing" being cross-domain. Maybe one country makes lots of free movies, and another country has really amazing free museums with subsidized hotel rates, so the two countries make their services available to the citizens of the other.
Because, you know, others have rights too.
This is America. If you feel endangered, remember the second amendment.
It was self-defense your honor. I saw him rub his nose and sneeze.
IMHO, one of the remaining hurdles to us getting past the Great Filter is the proliferation of technology and doomsday weaponry to all corners of the globe.
Honestly, I really only see the solution to this problem being the proliferation of humans to such an extent that warfare using the most powerful weapons available is not a threat to a substantial portion of the human race. If people colonized half the galaxy a nuclear war would take centuries to reach everybody simply due to the speed of light, to say nothing of effective countermeasures.
Obviously that isn't going to happen anytime soon.
The problem is that it is very difficult to put the genie back in the bottle. The design of a nuclear weapon is basically just information, and everybody can see just how hard it is to keep information under wraps. Over time we've seen increasingly more dangerous weapons coming into the hands of the general public. Maybe if the entirety of the human population could be kept under surveillance we could completely prevent the proliferation of such technology, but simply having an organization capable of such a feat is in some sense an existential threat of a different sort.
Furthermore, refueling an airplane does not take less than two minutes. That's the time you need just to plug in the fuel line.
He said, "But it would make loading and unloading the plane a matter of two minutes or something which is less time than it takes to refuel the plane."
Ie, refueling takes MORE than two minutes.
Really, Hacking Team was just doing things the way a very small segment of society which currently holds most financial capital thinks everybody should be operating.
FTFY - SOPA, TPP, etc.. are not products of the Software industry. I am pretty sure I agree with your point under the surface, but the generalization is plain wrong.
They are certainly the way the software industry thinks everybody should be operating, which is all I claimed. I did not claim that all of those laws/treaties/etc were products of the software industry. I'm not sure how you can claim that the Uniform Computer Information Transactions Act wasn't though.
Simple access doesn't count as distribution in this case. The GPLv2 applies to code distributed under the GPLv2. It isn't a property of the code itself, and the fact that you have GPLv2ed code doesn't mean you have to give it to me, nor do I get the right to the code by simply having access. If you deliberately give me the code, you have to do so under the GPLv2, and I have all the rights that grants.
You give your employee the code when you give them access to it. Before they couldn't see the code. Now they can. They gave it to you.
The FSF holds that having employees work on company code isn't distribution.
That's nice, but they aren't the authors of the code in question, even if they're the authors of the license. If the kernel authors intended the code to be copyable by the employee and the license says that it is, then it is.
Consider that I've got a lot of company-owned proprietary code on my work computer. If that counted as distributing it to me, I'd own one copy of the code.
Well, you do have one copy of the code in your possession - the one on the server. That doesn't mean that you can make another copy of the code without permission from the copyright holder.
Similarly, putting copy of GPLv2ed code on a work computer is not distribution, and the programmer doesn't have a copy under GPLv2.
The wording of the GPLv2 isn't terribly helpful here. It uses phrases like "copy or distribute" in some places and "distribute" in others and doesn't define either term. If nothing else it really seems open to litigation. It would be sensible to make stuff like this explicit in a license.
It would be an interesting case - no doubt an expensive one if there were deep enough pockets involved. Ultimately the law is whatever the courts say it is, and there is only one way to find out...
But there's another aspect of this. Say my company downloads the Linux kernel and we internally make some changes to it and use it on our servers in its modified form. Jim is one of the coders. Linux is released under GPLv2. Does that mean that Jim can take our changes home with him?
No.
The GPLv2 kicks in only when the company redistributes the code along with the modifications, and those modifications are available to the recipients that we've specified.
This is a common argument but I'm not convinced that it is airtight. How about this:
But there's another aspect of this. Say my company buys a Windows DVD and we install it on 47 of our servers. Jim is one of the coders. Windows DVDs are not licensed for multiple installations. Does that mean that Jim can call up MS and pocket a reward?
No.
The Windows License kicks in only when the company redistributes copies of Windows, and those copies are available to the recipients that we've specified.
The problem with this argument is that copyright applies anytime you make a copy of anything. Copyright says you can't install Linux anywhere. What lets you install it is the license. The license for Windows says you aren't allowed to copy it at all except to install it once. The license for Linux says you're only allowed to copy it if the copy is GPLv2 along with any modifications you've made, and you make the source available to any recipients. So, your modifications MUST be GPLv2 the moment they're installed with the kernel. Since they are GPLv2, you've given permission to anybody with access to them to redistribute them already, and that includes Jim.
Now, you can still tell Jim not to redistribute it and fire him if he does and blacklist him so that his kids don't have any kind of future as is fashionable in the US, and I'm not sure that a court would take issue with that. However, I'm not convinced that a court would hold up an employer's right to sue an employee for redistributing Linux kernel modifications if you could argue that a kernel was copied anywhere at all. Keep in mind that the software industry has been pushing for super-conservative interpretations of copyright such that simply loading a copy of a portion of the program into RAM to execute it is covered. Under that kind of interpretation (which is the basis of any kind of end-use licensing) you're copying the kernel the moment you open the source in an editor.
But, you do expose the general weakness with GPLv2 in that it was really designed with a world of software sold in boxes (ie the 90s). The wording really needed cleanup and licenses like GPLv3 and AGPLv3 do much of that.
Sure, but this is all stuff that is par for the course with laws like SOPA, TPP, UCITA, and so on.
Really, Hacking Team was just doing things the way the software industry thinks everybody should be operating.
And the original point stands... have you audited the source code? Have you compiled / built the product from the source code? If not, then have the alleged source code isn't any better than trusting the Windows source code.
I have done both. So, I guess you'd say I'm fine then?
Most people who use FOSS audit it in the same way that any major corporation does due diligence for stuff they buy. They make sure that people they trust are looking into it.
Sure, you can sneak stuff into reputably-distributed FOSS just like you can sneak stuff into Windows CDs. The fact is that there probably still are a lot more people checking the integrity of something like RHEL than Windows, simply by virtue of it being FOSS.