I'm not arguing for or against. I try not to do much bit slinging where I can avoid it. In fact I think most people do... till they REALLY need to. Then, for everyone's sake, I hope they do it well.
However... effing HOW? I can't even get people to use PARAMETERIZED QUERIES. How in the HELLS am I going to get them to use a crypto system?
Sure, I can make it easy to some degree, but then how do we do queries on anything but the unencrypted bits? Sure, there are ways (oh there are ways. Each more devilishly complex and mathematically involved than the last)
Than we have to store the keys. So many keys... and the IVs and the data to know what information they point to >_<
Now where to put the keys... gonna need to be FAST and reliable. Gee, it sure would be nice if someone had an infrastructure for that sort of thing... Aww hell.
In apt systems at least... Create an overlay filesystem based on / chroot into the overlay run updates have a handy dandy "boot into overlay" option (I've done this, knoppix does this, it isn't hard and they CAN put it in there with no hassle) test test more give it the okay! reboot into a "collapse the changes into the mainline root" mode which will then reboot normally (into the freshly collapsed root). One can collapse the two using rsync. It will be very fast Profit.
Take you two hours to do or about three weeks if you wanted it to be useful for the unwashed masses. Nobody does this because.... ummm... Actually I'm not sure, but I'll bet it has a lot to do with POTENTIALLY breaking things and you can't put that into a distro.
However, feel free to whip it up and enjoy. I just take a nice sturdy backup before a major change. If you think that restoring a backup is too long / costly / unreliable then you have a problem with you backup system. Also: test environments are handy. I guess those are good arguments against having a built-in "try and see" mode. Breaks things and there are better solutions than mucking about with your boot config and filesystems for most.
I'm not going to argue that having well oiled, well supported piece of desktop software is a great way to do things. With the right arrangement of technologies it isn't even that hard to make a client-server pair. I'll leave the technology of choice to the implementer as well as the definition of hard.
Having said that, I develop a semi-internal website for a living at a multinational corporation that needs to be used by a few hundred different people ranging in age from 20 to 70 across 17 timezones on their hardware (they are distributors and sales people). Putting this application on the web was the only way to get anyone to use it and it works very well. It didn't used to, but then they hired me and now it does.
I'm no shaman. Heck, I use PHP up front and Java in the backend with no framework for the web side and tomcat / Apache Cayenne as the Java servlet container / DB interface on the back. This means, I hope, I'm not pushing an agenda or a philosophy, but rather advocating sane development and care no matter what the tools or job at hand.
Software, no matter where, is only as good as the thought put into it. Time needed varies depending on what and how you do it, of course.
I really find it odd to admit, but I've found that developing in IE9 is a pleasure in so far as development in IE can be. Seriously! They have dev tools built-in (not as good as firebug, but it is enough to work out kinks!) It supports SVG (I do some things as vectors when I know that they are going to scale for the situation. It is for generic JS tools that are meant to be used anywhere in our web suite. simply: var hasSvg = false; try{hasSvg = document.implementation.hasFeature("http://www.w3.org/TR/SVG11/feature#BasicStructure", "1.1");}catch{;}
and you get a boolean to decide to include SVG graphics or raster). Best of all it, so far, renders everything I've tested on it about pixel for pixel (user inputs aside) as FF4. Even CSS rounded corners!
Perhaps IE9 has some flaws, but it is everything I need it to be! Hooray web standards!
This is simplified, don't take this completely literally, but get this first. I'll use a car analogy.
You and several other clowns are in a clown and some of them are juggling. You are driving so you can't look at them. You can't look because you are doing a precision maneuver with several other clown cars. As part of the act they are also exchanging juggling objects with other cars. Even though you can't look at the jugglers you can sense what they are doing due to the fact that their motions and transfer of momentum are throwing you off course. It is important that you stay on course to make the jump. God help you if you hit the ramp like like the last guy did, but the kids like to see this act up close.
If the jugglers are throwing around tennis balls your course will be effected differently than if they were throwing juggling pins.
Now, back in the world of the article you've got the same thing. Atoms with electrons flying around and shared by chemical bonds. The shape of the electrons effects the shape of the molecule. More specifically the shape of the charge around the electron effects the shape.
Don't try to watch the objects being juggled, watch the clown cars try to stay in formation on their way to the jump over lion pen.
It took a long time because the measurements are so delicately precise and spurious data had to be discounted and filtered from the signal. The measurements weren't averaged but they were mercilessly filtered and subjected to analysis to take the "noise floor" down this low.
I am not a physicist. Someone correct me or clarify if I was dead wrong. Thanks!
I even said that string theory is a large unknown. It is most likely bunk. Might not be, but that is not my area of expertise.
You mentioned that the models are "unconsciously" steering their experiments. They aren't. They are 100% fully aware of how the models are steering their experiments. That is how it works. Gotta test the model, how else do you test a model other than devise a test for it?
More importantly that is how plenty of models die. The tests show them to be wrong. Creativity certainly has its place in science, that is how the big leaps are made. Then they go and check and re check and publish and others do all that all over again.
Science is repeatable. If I can do something, so can you! Not only that but science intermingles. If someone comes up with the wrong idea but don't manage to prove it wrong then other pieces won't fit. Not only are the habits of scientists self-verifying but the science itself is too.
Most importantly if a scientist is wrong they will admit it. There will be whining, some gnashing of teeth, and all the human drama to some degree, but in the end it is either accept the evidence or be shamed out of a career. Truth (of this variety) has a way of becoming pervasive.
Again I ask, because I think this is key: What things specifically do you disagree with outside of those I mentioned? Some examples? Anything other than undirected mistrust?
You are still missing the point. Much of what we consider to be "modern" science is still a house of cards which is not built on a known quantity.
Such as? I can think of string theory (a large unknown with no testable predictions and thus not used for any real science as of yet), dark matter (though we can observe the effects so we know it to be there, they just don't know what it is yet. Lots of very bad theories coming out of this one. I can say that because all but one of them is wrong, and we might not even have the one yet!), and dark energy (which recently got a big boost in observational evidence!)
In the classical scientific method, you have to have a known value that has been verified to be true.
No. Not always. The periodic table of elements was envisioned before there was enough evidence to prove that model correct. Ditto for gravity, "germs", electricity, relativity (general and special), and quantum physics. Those models predicted things that had yet to be observed based on mathematics and known quantities. Their predictions have been tested and proven right. Theories that were wrong are less remembered because they were wrong and abandoned. Steady state, for example.
Mathematical models are not an adequate replacement for known values derived from direct observation.
No, they compliment them. Models predict, observations confirm or contradict. Some models work well only sometimes, like gravity, but within their realm of usefulness they advance science to the point that they break down and science happens all over again. Constant testing and verification. This, just like every other useful model, will be tested mercilessly.
If you think that math and science are strange bedfellows then I encourage you to attempt science without mathematical models. I just don't see how a lack of modeling is an advantage. Clarify if you are willing, but I'm just not seeing it. Mathematics have become more important and much more the focus but it works and it has worked well. Saying that the models don't yield observable quantities nor come from observable quantities is just wrong. Some models take decades. DECADES (I can't stress this enough. Some of Einsteins predictions are still being tested today that have never been tested before such as frame dragging) before their predictions are testable let alone proven.
I'm curious, would CTR be less vulnerable if one XORed before encryption? Call the operation CXR. Where ^ is the XOR operator CXR(block(i)) = encrypt(IV ^ i ^ block(i))
I'm not sure if there is analysis that can be done on the block at that point that makes this undesirable. Methinks not because as far as I know having a well known IV in, say, CBC is not a vulnerability. That implies to me that the security still rests firmly in the key. At the very least it stops being vulnerable to bitwise changes and reinstates the Confusion and Diffusion principals.
There might also be some magic in reading the whole block (since we are talking about block level devices) and having, say, a CBC over the block with an IV calculated with encrypt(IV ^ i) but I think that goes out of scope of my question.
Route counts are climbing fast Moving IP blocks around from their nice chunky/8 homes will make it necessary to advertise subnets AS numbers will not be issued to the fragmented blocks once the routing tables are a nice fat size and some older routers can't handle it (again, this type of thing has happened before) OR when they decide to just not hand out AS numbers for these fragmented blocks (to force the issue) No AS number, no ability to route a subnet differently from the entire block. No resale value in an unrouteable block Upgrades needed to handle the growing routing table AND/OR blocks are too large and unwieldy to be moved to where the customers need them IPv6 wins by default because of the need to upgrade either way, even if it has to coexist for a long time
Kind of a high level view, and I don't know all the ins and outs of AS number assignment, but I think that strangling that resource would work nicely even if a few policies needed changing.
You are talking about carrier grade NAT64/44. For the bandwidth they handle this would be a significant investment in addition hardware to provide a translation mechanism for ipv6 -> ipv4. Money that will only be in addition to the cost of upgrading their hardware, firmware, and network layout to ultimately support ipv6 overall.
Your device does, in fact, have to communicate "directly" with the endpoint servers maintained by third parties in some fashion. Adding in an additional layer of translation solves nothing in the long term and only serves to absorb money for very short term gain. Such systems may be required due to what amounts to wanton incompetence or malicious intransigence of network operators who do not transition to ipv6 in a timely manner.
The morality of the situation, however, is an entirely different discussion from the technical one.
IPv6 operates above layer two. Switches of any kind whatsoever that *ahem* "support" ipv4 will also operate equally well with ipv6. Layer 3 "switches" not withstanding, of course. Those are more closely related to routers than standard layer two switches.
As far as routers go: no, we don't shit money. We know how to type. Specifically we know how to type into our router's configuration to turn on ipv6. IPv6 routers are magical beats carved out of unicorn ivory and powered by the souls of freshly deceased cobol programmers, they are commodity. As they should be a decade and more after ipv6 was ratified as a standard.
Granted, some of Cisco's muti gigabit scalable routers do, in fact, route in hardware to a degree and cannot support ipv6 in the fast path, but I don't actually know terribly much about that, I fear. Never had a need to run that sort of bandwidth.
Not that I doubt that management is intransigent for reasons that they hold dear BUT... what does the stateful DHCP service provide them in the IPv6 context? What excuse are they pulling out to "require" this. I'm interested in knowing.
Sweet! You mean to say that all websites and application specific servers for mobile phones have been migrated to ipv6! Awesome!
Oh wait... hold on a second... Almost the entirety of the English speaking Internet still isn't on ipv6?
Whats that you say? Not even friggin' slashdot?
I wonder if THAT is why.
Now having said that: Every computer I'm an admin for is 100% ipv6 compatible and all of my servers have AAAA dns records alongside their A records. I've even got a nice little OSPFv3 infrastructure running. It isn't friggin rocket surgery, but everyone is dragging their ass on the ground like the problem will become someone else's, when in reality it will shortly become everyone's. All of my efforts are in vain so long as there is a dearth of IPv6 accessible content.
By the by, are you running IPv6?
Lastly: For everyone who says that it is "hard" for large network to migrate, and they they have to re-learn everything yadda yadda:
IPv6 is easier to work with on a large scale thanks to the simplified routing tables that it affords as well as the shotgun approach to address assignment. Every single link is a/64 at minimum (and maximum, given the number addresses in a/64) and the blocks can be handed out ham-fistedly because of the mind boggling size of the space. If they have hardware that does not support ipv6 then they should blame themselves. Large network operators have NO EXCUSE. They knew this was coming and their profit margins are wide enough that they could have thrown money at it.
Actually there have been some studies on the feasibility of solar and wind to not only power the world as is, but to keep up with demand and the need to bring the entirety of the world to at LEAST the level of western europe in terms of energy per capita (a critical metric in the evaluation of quality of life).
I don't have the study handy, but this video quotes it a href=http://video.google.com/videoplay?docid=-1518007279479871760#>http://video.google.com/videoplay?docid=-1518007279479871760#
for those of you without the time to watch it the short of it is that we would have to increase the production amount of things like steel, glass and aluminum to levels never before seen (low integer multiples of current levels). It would have to be a concerted worldwide effort just to keep up with the growing demand, let alone get to a stable state where we can use those raw materials for other things, like building houses or cars.
Nuclear is, for now, the only option. Fusion after that I hope.
The upshot: use QKD to transmit the key, then rely on classical encryption schemes
Only if you want a point to point link between you and everyone you need to talk to (amazon, gmail, etc) or trust the intermediate nodes (AT&T, Verizon, Comcast, the government who can wiretap anything they damn well please by court order)
Then again, computing power increases so quickly that I doubt AES will be secure for long.
I've done a lot of programing but not a lot of UI work outside of the browser (where most of my user interaction has been). However, I've found that the JAVA SWT to be "okay" for desktop GUI apps. Sure, it isn't as flexible as opengl or javascript/html(xml)/css, but it gets the job done for your standard inputs and outputs. If you want to do something fancier than the most basic 2d graphics (or minecraft-like 3d) you won't like it and neither will your users.
Mostly what you should ask yourself is "Can a browser do this?" and if the answer is no (sounds like it is) then ask, "Do I need anything fancy?" Those two questions will drive your decision.
Good luck. Dealing with the GUI is an order of magnitude more fiddly than all the algorithms work you will ever encounter. I should know, I'm marrying a graphics designer;)
The fact that you're too much of a chicken shit to post your opinion with a name suggests that you do in fact give a damn, enough at least to not associate your own handle with your own oxygen wasting stupidity.
Look up e fuses, that'd be how they do it. Have 128 efuses on the CPU for the "code" and one efuse to disable the commands to "blow" the efuse. That way you can't blow all the fuses in the field and then input all ones (or zeros, depending on how the CPU treats a blown fuse) as the unlock code. The serial number on the CPU is then effectively a nonce for looking up the proper code in their central database (that or fed into a Message Authentication Code [MAC] since it's easier to calculate from a secret than to maintain a large reliable database, but I digress). They've got the technology, and they seem to think it's worth the effort. Then again, they thought the same thing about the PIII serial numbers...
Debian, making installing dependencies a reflex rather than a compulsory chore. That alone would have gotten my praise. Then they also bolted on an incredibly stable and useful kernel and software stack on top of that.
Good show! (I know I got the order wrong, but thats the order of importance to me)
Slashdot Mirror
I'm not arguing for or against. I try not to do much bit slinging where I can avoid it. In fact I think most people do... till they REALLY need to. Then, for everyone's sake, I hope they do it well.
Here is to the bit singers!
I found this that provides some info about the lack of unsigned. I'm reading it now. http://darksleep.com/player/JavaAndUnsignedTypes.html#why_no_unsigned_types
Why?
How is
Widget a = new Widget();
Widget b = getBaseWidget();
b = a + b;
Any better than
Widget a = new Widget();
Widget b = getBaseWidget();
b.addNewModule(a);
Are you playing keyboard golf? Do you like having to remember what each operator does in every different combination? Does your IDE not auto-complete?
Really, tell me. Overloading operators seems like a good way to help ruin the readability of your code.
Okay. I'll bite.
Yes encrypting the data is good enough http://everything2.com/title/Thermodynamics+limits+on+cryptanalysis and http://en.wikipedia.org/wiki/Grover's_algorithm (if quantum computers come up then grover's algorithm still isn't much use. See the max speedup of the algorithm)
However...
effing HOW? I can't even get people to use PARAMETERIZED QUERIES. How in the HELLS am I going to get them to use a crypto system?
Sure, I can make it easy to some degree, but then how do we do queries on anything but the unencrypted bits? Sure, there are ways (oh there are ways. Each more devilishly complex and mathematically involved than the last)
Than we have to store the keys. So many keys... and the IVs and the data to know what information they point to >_<
Now where to put the keys... gonna need to be FAST and reliable. Gee, it sure would be nice if someone had an infrastructure for that sort of thing... Aww hell.
In apt systems at least...
Create an overlay filesystem based on /
chroot into the overlay
run updates
have a handy dandy "boot into overlay" option (I've done this, knoppix does this, it isn't hard and they CAN put it in there with no hassle)
test
test more
give it the okay!
reboot into a "collapse the changes into the mainline root" mode which will then reboot normally (into the freshly collapsed root). One can collapse the two using rsync. It will be very fast
Profit.
Take you two hours to do or about three weeks if you wanted it to be useful for the unwashed masses. Nobody does this because.... ummm... Actually I'm not sure, but I'll bet it has a lot to do with POTENTIALLY breaking things and you can't put that into a distro.
However, feel free to whip it up and enjoy. I just take a nice sturdy backup before a major change. If you think that restoring a backup is too long / costly / unreliable then you have a problem with you backup system. Also: test environments are handy. I guess those are good arguments against having a built-in "try and see" mode. Breaks things and there are better solutions than mucking about with your boot config and filesystems for most.
I guess I don't have a point... Enjoy either way!
I'm not going to argue that having well oiled, well supported piece of desktop software is a great way to do things. With the right arrangement of technologies it isn't even that hard to make a client-server pair. I'll leave the technology of choice to the implementer as well as the definition of hard.
Having said that, I develop a semi-internal website for a living at a multinational corporation that needs to be used by a few hundred different people ranging in age from 20 to 70 across 17 timezones on their hardware (they are distributors and sales people). Putting this application on the web was the only way to get anyone to use it and it works very well. It didn't used to, but then they hired me and now it does.
I'm no shaman. Heck, I use PHP up front and Java in the backend with no framework for the web side and tomcat / Apache Cayenne as the Java servlet container / DB interface on the back. This means, I hope, I'm not pushing an agenda or a philosophy, but rather advocating sane development and care no matter what the tools or job at hand.
Software, no matter where, is only as good as the thought put into it. Time needed varies depending on what and how you do it, of course.
I really find it odd to admit, but I've found that developing in IE9 is a pleasure in so far as development in IE can be. Seriously! They have dev tools built-in (not as good as firebug, but it is enough to work out kinks!) It supports SVG (I do some things as vectors when I know that they are going to scale for the situation. It is for generic JS tools that are meant to be used anywhere in our web suite. simply:
var hasSvg = false; try{hasSvg = document.implementation.hasFeature("http://www.w3.org/TR/SVG11/feature#BasicStructure", "1.1");}catch{;}
and you get a boolean to decide to include SVG graphics or raster). Best of all it, so far, renders everything I've tested on it about pixel for pixel (user inputs aside) as FF4. Even CSS rounded corners!
Perhaps IE9 has some flaws, but it is everything I need it to be! Hooray web standards!
This is simplified, don't take this completely literally, but get this first. I'll use a car analogy.
You and several other clowns are in a clown and some of them are juggling. You are driving so you can't look at them. You can't look because you are doing a precision maneuver with several other clown cars. As part of the act they are also exchanging juggling objects with other cars. Even though you can't look at the jugglers you can sense what they are doing due to the fact that their motions and transfer of momentum are throwing you off course. It is important that you stay on course to make the jump. God help you if you hit the ramp like like the last guy did, but the kids like to see this act up close.
If the jugglers are throwing around tennis balls your course will be effected differently than if they were throwing juggling pins.
Now, back in the world of the article you've got the same thing. Atoms with electrons flying around and shared by chemical bonds. The shape of the electrons effects the shape of the molecule. More specifically the shape of the charge around the electron effects the shape.
Don't try to watch the objects being juggled, watch the clown cars try to stay in formation on their way to the jump over lion pen.
It took a long time because the measurements are so delicately precise and spurious data had to be discounted and filtered from the signal. The measurements weren't averaged but they were mercilessly filtered and subjected to analysis to take the "noise floor" down this low.
I am not a physicist. Someone correct me or clarify if I was dead wrong. Thanks!
I even said that string theory is a large unknown. It is most likely bunk. Might not be, but that is not my area of expertise.
You mentioned that the models are "unconsciously" steering their experiments. They aren't. They are 100% fully aware of how the models are steering their experiments. That is how it works. Gotta test the model, how else do you test a model other than devise a test for it?
More importantly that is how plenty of models die. The tests show them to be wrong. Creativity certainly has its place in science, that is how the big leaps are made. Then they go and check and re check and publish and others do all that all over again.
Science is repeatable. If I can do something, so can you! Not only that but science intermingles. If someone comes up with the wrong idea but don't manage to prove it wrong then other pieces won't fit. Not only are the habits of scientists self-verifying but the science itself is too.
Most importantly if a scientist is wrong they will admit it. There will be whining, some gnashing of teeth, and all the human drama to some degree, but in the end it is either accept the evidence or be shamed out of a career. Truth (of this variety) has a way of becoming pervasive.
Again I ask, because I think this is key: What things specifically do you disagree with outside of those I mentioned? Some examples? Anything other than undirected mistrust?
You are still missing the point. Much of what we consider to be "modern" science is still a house of cards which is not built on a known quantity.
Such as? I can think of string theory (a large unknown with no testable predictions and thus not used for any real science as of yet), dark matter (though we can observe the effects so we know it to be there, they just don't know what it is yet. Lots of very bad theories coming out of this one. I can say that because all but one of them is wrong, and we might not even have the one yet!), and dark energy (which recently got a big boost in observational evidence!)
In the classical scientific method, you have to have a known value that has been verified to be true.
No. Not always. The periodic table of elements was envisioned before there was enough evidence to prove that model correct. Ditto for gravity, "germs", electricity, relativity (general and special), and quantum physics. Those models predicted things that had yet to be observed based on mathematics and known quantities. Their predictions have been tested and proven right. Theories that were wrong are less remembered because they were wrong and abandoned. Steady state, for example.
Mathematical models are not an adequate replacement for known values derived from direct observation.
No, they compliment them. Models predict, observations confirm or contradict. Some models work well only sometimes, like gravity, but within their realm of usefulness they advance science to the point that they break down and science happens all over again. Constant testing and verification. This, just like every other useful model, will be tested mercilessly.
If you think that math and science are strange bedfellows then I encourage you to attempt science without mathematical models. I just don't see how a lack of modeling is an advantage. Clarify if you are willing, but I'm just not seeing it. Mathematics have become more important and much more the focus but it works and it has worked well. Saying that the models don't yield observable quantities nor come from observable quantities is just wrong. Some models take decades. DECADES (I can't stress this enough. Some of Einsteins predictions are still being tested today that have never been tested before such as frame dragging) before their predictions are testable let alone proven.
I'm curious, would CTR be less vulnerable if one XORed before encryption? Call the operation CXR.
Where ^ is the XOR operator
CXR(block(i)) = encrypt(IV ^ i ^ block(i))
I'm not sure if there is analysis that can be done on the block at that point that makes this undesirable. Methinks not because as far as I know having a well known IV in, say, CBC is not a vulnerability. That implies to me that the security still rests firmly in the key. At the very least it stops being vulnerable to bitwise changes and reinstates the Confusion and Diffusion principals.
There might also be some magic in reading the whole block (since we are talking about block level devices) and having, say, a CBC over the block with an IV calculated with encrypt(IV ^ i) but I think that goes out of scope of my question.
Oh ye gods yes they will.
Here is how I see this going down:
Route counts are climbing fast /8 homes will make it necessary to advertise subnets
Moving IP blocks around from their nice chunky
AS numbers will not be issued to the fragmented blocks once the routing tables are a nice fat size and some older routers can't handle it (again, this type of thing has happened before) OR when they decide to just not hand out AS numbers for these fragmented blocks (to force the issue)
No AS number, no ability to route a subnet differently from the entire block. No resale value in an unrouteable block
Upgrades needed to handle the growing routing table AND/OR blocks are too large and unwieldy to be moved to where the customers need them
IPv6 wins by default because of the need to upgrade either way, even if it has to coexist for a long time
Kind of a high level view, and I don't know all the ins and outs of AS number assignment, but I think that strangling that resource would work nicely even if a few policies needed changing.
I love deadlines. I like the whooshing sound they make as they fly by.
~Douglas Noel Adams
You are talking about carrier grade NAT64/44. For the bandwidth they handle this would be a significant investment in addition hardware to provide a translation mechanism for ipv6 -> ipv4. Money that will only be in addition to the cost of upgrading their hardware, firmware, and network layout to ultimately support ipv6 overall.
Your device does, in fact, have to communicate "directly" with the endpoint servers maintained by third parties in some fashion. Adding in an additional layer of translation solves nothing in the long term and only serves to absorb money for very short term gain. Such systems may be required due to what amounts to wanton incompetence or malicious intransigence of network operators who do not transition to ipv6 in a timely manner.
The morality of the situation, however, is an entirely different discussion from the technical one.
IPv6 operates above layer two. Switches of any kind whatsoever that *ahem* "support" ipv4 will also operate equally well with ipv6. Layer 3 "switches" not withstanding, of course. Those are more closely related to routers than standard layer two switches.
As far as routers go: no, we don't shit money. We know how to type. Specifically we know how to type into our router's configuration to turn on ipv6. IPv6 routers are magical beats carved out of unicorn ivory and powered by the souls of freshly deceased cobol programmers, they are commodity. As they should be a decade and more after ipv6 was ratified as a standard.
Granted, some of Cisco's muti gigabit scalable routers do, in fact, route in hardware to a degree and cannot support ipv6 in the fast path, but I don't actually know terribly much about that, I fear. Never had a need to run that sort of bandwidth.
Not that I doubt that management is intransigent for reasons that they hold dear BUT... what does the stateful DHCP service provide them in the IPv6 context? What excuse are they pulling out to "require" this. I'm interested in knowing.
Sweet! You mean to say that all websites and application specific servers for mobile phones have been migrated to ipv6! Awesome!
Oh wait... hold on a second... Almost the entirety of the English speaking Internet still isn't on ipv6?
Whats that you say? Not even friggin' slashdot?
I wonder if THAT is why.
Now having said that: Every computer I'm an admin for is 100% ipv6 compatible and all of my servers have AAAA dns records alongside their A records. I've even got a nice little OSPFv3 infrastructure running. It isn't friggin rocket surgery, but everyone is dragging their ass on the ground like the problem will become someone else's, when in reality it will shortly become everyone's. All of my efforts are in vain so long as there is a dearth of IPv6 accessible content.
By the by, are you running IPv6?
Lastly: For everyone who says that it is "hard" for large network to migrate, and they they have to re-learn everything yadda yadda:
IPv6 is easier to work with on a large scale thanks to the simplified routing tables that it affords as well as the shotgun approach to address assignment. Every single link is a /64 at minimum (and maximum, given the number addresses in a /64) and the blocks can be handed out ham-fistedly because of the mind boggling size of the space. If they have hardware that does not support ipv6 then they should blame themselves. Large network operators have NO EXCUSE. They knew this was coming and their profit margins are wide enough that they could have thrown money at it.
Actually there have been some studies on the feasibility of solar and wind to not only power the world as is, but to keep up with demand and the need to bring the entirety of the world to at LEAST the level of western europe in terms of energy per capita (a critical metric in the evaluation of quality of life).
I don't have the study handy, but this video quotes it a href=http://video.google.com/videoplay?docid=-1518007279479871760#>http://video.google.com/videoplay?docid=-1518007279479871760#
for those of you without the time to watch it the short of it is that we would have to increase the production amount of things like steel, glass and aluminum to levels never before seen (low integer multiples of current levels). It would have to be a concerted worldwide effort just to keep up with the growing demand, let alone get to a stable state where we can use those raw materials for other things, like building houses or cars.
Nuclear is, for now, the only option. Fusion after that I hope.
The upshot: use QKD to transmit the key, then rely on classical encryption schemes
Only if you want a point to point link between you and everyone you need to talk to (amazon, gmail, etc) or trust the intermediate nodes (AT&T, Verizon, Comcast, the government who can wiretap anything they damn well please by court order)
Then again, computing power increases so quickly that I doubt AES will be secure for long.
Yes it will.
I've done a lot of programing but not a lot of UI work outside of the browser (where most of my user interaction has been). However, I've found that the JAVA SWT to be "okay" for desktop GUI apps. Sure, it isn't as flexible as opengl or javascript/html(xml)/css, but it gets the job done for your standard inputs and outputs. If you want to do something fancier than the most basic 2d graphics (or minecraft-like 3d) you won't like it and neither will your users.
Mostly what you should ask yourself is "Can a browser do this?" and if the answer is no (sounds like it is) then ask, "Do I need anything fancy?" Those two questions will drive your decision.
Good luck. Dealing with the GUI is an order of magnitude more fiddly than all the algorithms work you will ever encounter. I should know, I'm marrying a graphics designer ;)
You deserve + mod. I lack mod points, so have this waffle instead (>'.')>#
The fact that you're too much of a chicken shit to post your opinion with a name suggests that you do in fact give a damn, enough at least to not associate your own handle with your own oxygen wasting stupidity.
*clap clap clap*
Even I caught a little heat off that one.
Nice reference
Look up e fuses, that'd be how they do it. Have 128 efuses on the CPU for the "code" and one efuse to disable the commands to "blow" the efuse. That way you can't blow all the fuses in the field and then input all ones (or zeros, depending on how the CPU treats a blown fuse) as the unlock code. The serial number on the CPU is then effectively a nonce for looking up the proper code in their central database (that or fed into a Message Authentication Code [MAC] since it's easier to calculate from a secret than to maintain a large reliable database, but I digress). They've got the technology, and they seem to think it's worth the effort. Then again, they thought the same thing about the PIII serial numbers...
Debian, making installing dependencies a reflex rather than a compulsory chore. That alone would have gotten my praise. Then they also bolted on an incredibly stable and useful kernel and software stack on top of that.
Good show! (I know I got the order wrong, but thats the order of importance to me)