This is _not_ a troll, but nor am I a karma whore either. In all seriousness, we should look to MS-Windows, specifically 95/NT for an easy simple, intuitive user interface. My son learnt it at age 5 without much instruction and he's now better than me.
Microsoft deserves credit where it is due. No-one doubts their marketing prowess, or ruthlessness either. I believe they have put considerable effort into UI design via expensive user panels and focus groups. They certainly haven't put effective effort into OS design, which has become recursive layers of kludge frequently going random. But there is no reason that we should not incorporate User Interface findings to the extent possible. Furthermore, it would ease transition.
To the specific question about OS incorporating user feedback, as a modest OS developer (cpuburn) I feel particularly qualified to comment. I get some user feedback, and if it seems like a good idea to me, I will incorporate it.
But if I don't like it, forget it, and I may be capricious. One MS-Windows user wanted start & stop buttons added. I firmly declined, even though I conceed it will improve ease-of-use. I prefer the user learn about TaskMan. If he wants buttons, let him add them. The code _is_ GPL.
Slackware is BSDish vs SysVish UN*X
on
Slackware Updates
·
· Score: 2
I have used Linux since kernel 1.3.15 days, and at that time, Slackware was the most polished and complete distribution. For quite some time, it languished, and only recently has seen the maintenance/updating found on more recent distro's, notably RedHat.
I've used and continue to use RedHat extensively. It is certainly the best Linux distro for Alpha architecture. It is exceptionally easy for beginners to install, yet remains very powerful via the `rpm` mechanism of package and especially dependancy updating.
Yet Slackware still has powerful attractions. It can certainly be installed on less hard-disk and RAM than RedHat. It leads to better UNIX understanding [albeit less easily] by using an editor for sysadmin rather than `linuxconf` tools. Furthermore, I find the file layout easier to understand, especially the BSD-style/etc/rc.d init scripts. I detest the byzantine SysV init scripts found in RedHat.
Yes, using anonymity _may_ reduce the impact of what is said. Particularly if there is no apparent need for anonymity. Paranoia doesn't look good.
But where the reader agrees with the need for anonymity [controversial/retributive material] then I don't think the impact is reduced much. Sometimes, publishing nominatively may appear foolhardy or risk/publicity seeking behaviour.
Hasn't the US Supreme Court defended anonymity as an important element of free speech? Without anonymity, there is always fear of retribution.
Not that Presidents or administrations have ever had much interest in supporting freedom. Even if they are sworn to defend the US Constitution, their main interest is in governing the populace. Freedoms make that job harder.
Do two wrongs make a right? Bezos' defense merely another form of the question. The trite answer "of course not" is naive, because in many situtations, two wrongs do seem to make a right. For example "The Prisoners Dilemma" in modern game theory.
More important is the fundamental question of intellectual property rights. IP has piggybacked onto physical property rights. But IP & RP are fundamentally different things, and the most important commercial use of IP, publishing, has no analog in the real-property world.
When you sell wheat, you aren't worried about someone copying it. You havent sold the farm. When you publish, you have sold the farm because copying is trivially easy. But you won't necessarily notice copying the way you would if someone stole your bushel of wheat.
Thanks for your reply. ctcm measures `movsd` which is a load and a store for each word. If I should have 4112 MB/s @ 539 MHz, that's 7.6 bytes read and 7.6 bytes written each clock. Not very likely unless I've got a 128bit path or dual ported SRAM. But you are right, ctcm is more a measure of bandwidth than latency.
I dug out my pseudrorandom access asm timer. I measure 10.7 Mreads/s from DRAM (9.1 busclocks), 20.0 Mreads/s from L2 (27 Celeron CPU clks) and 525 Mreads/s from L1 (1.03 CPU clks). So L1 seems single cycle, but L2 looks oddly slow, perhaps due to unintended thrashing.
As for the power/die budget, I'm afraid I don't know enough about chip feature design. But from all the micrographs I've seen, L1 is a fairly small portion of the die, so doubling it wouldn't be too painful. It also appears disproportionately large compared to L2, so something like this has probably been done.
I think the latencies you gave are for L2 cache. As for power budgets, I hardly think the P6 core is that great. But it _is_ alot better than an Alpha 21264 633 MHz with 107W (47Amps @ 2.35V).
As for SDRAM latency, I've measured ~9 Mreads/sec for pseudorandom P6 addresses. Now that _is_ 11 busclocks, but remember the P6 always accesses DRAM by full cache lines, so latency is (11-3)=7 busclocks (more for later bytes). So 42-66 CPU clocks if the multiplier is 6x.
As for L1 latency, I don't recall my read rates. But they'd have to be _very_ fast to allow ctcm to report 2700 MB/s @ 539 MHz in `movsd` to L1. This sounds like 1 clock/transfer to me, and the L1 might even be double (read & write ) ported.
First, high compliments on an outstanding and insightful article on one high-end of the computing business. It is easy to forget there are other aspects than the max-CPU performance sought by hobbyists.
The discussion of architectural performance benefits was very clear and insightful. There are obvious limits to multi-issue architectures.
A few corrections, if indeed I am correct: Main memory fetch is _not_ the oft-quoted "hundreds of CPU cycles". Typical SDRAM timing is 6-1-1-1, or 9 bus cycles per 32byte cache line. For a 600 MHz CPU with a 6x multiplier, this is 54 cycles, plus perhaps a few for page misses, etc.
Also, AFAIK at least Intel's P6 x86 core has a 1 CPU cycle latency L1 cache. Such a fast cache is necessary to make up for the risible shortage of x86 registers and helps considerably with stack-based operations such as often generated by `c` code. I do know that I can realize three RISC-type uops per clock cycle when 33-50% of the uops are loads from L1.
AFAIK, Jon Johansen is going to be tried in Norway. There is another DeCSS trial going on in the US against some websites.
But note how this MPAA tough-guy is afraid of admitting he even made a polite request for a teenager's arrest. He implies they did nothing at all, when the Norwegian prosecutor says otherwise.
I think it would be quite an unwelcome precedent if a Norwegian citizen were to be extradited to the US for breaking a US law while physically located in Norway. IANAL, but extradition treaties are all about returning fugitives, not extraterritorial law enforcement. Even wire fraud is prosecuted in the home country unless the fraudster willingly travels to the victim country.
Put it this way: There are some neoNazi groups in the US whose writings are illegal under German law. Do you think the US police could turn them over to the German police even if they wanted to? But it might not be too smart for them to travel to Germany. (But no-one has ever accused them of being smart).
If this man flew 51 missions in WW2, he's no baby boomer (usually considered the cohorts born 1946-1965). More like one of their parents, most likely born before 1927.
Regardless, he seems scared of being implicated with the arrest of the Norwegian teenager who authord DeCSS (Johanssen?). Most of his replies were evasive, as is customary among PHBs. But that one wasn't. I infer he's feeling some heat.
Cat 5 is std for 1000baseTX, not just NatSemi
on
Cheap Gigabit Ether
·
· Score: 2
According to the Gigabit Ethernet group draft standard IEEE 802.3ab, Gigabit copper 1000base TXshould run on all decent Cat 5 installations.
It does this by running single duplex over all four pairs at 125 MHz. The coding is changed to increase the number bits per symbol from 0.8 to 1.25. Simple wiring screw-up like mixing-up tip & ring are already handled by most 100baseTX ethernet transceivers. But crossover-vs-not isn't, and split pairs are unfixable.
Your Cat5 working 100baseTX is supposed to run 1000baseTX just fine. But it won't if you've left pairs unconnected, or stole them for a second run or phone. Poor crimping might also hurt.
That said, the real question is what you can do with all that bandwidth. Most hard-disks cannot sustain even 10 MB/s that 100baseTX provides. And it's hardly a high spped internet solution. It only runs 100m from the hub. The real problem with internet has always been interbuilding: the last mile between cable heads and user buildings.
Well yes, there probably is less than 2^24 (16 M) entries. But that's not the problem. The real problem is how do you search 1-16M entries in the miniscule time you have available?
Figure that each of the pipes into a MAE is at least 1000 Mbit/s. The average packet might be 200 bytes = ~1600 bits. 1.6 ms per packet, or search 1M entries at 2500 GByte/s. Of course, you do parallel processing to reduce this, but you still need alot of fast cache SRAM.
I agree with you. I never said, and didn't mean to imply that IPv6 was started as any sort of conspiracy. It was plainly started with the loftiest of technical goals: improving routing and multicasting.
Routing is a _terrible_ job. Just think of what happens to a packet that arrives in MAE East. It came in one one pipe. Presumably it doesn't go back out on that one. But on which one of the other pipes does it go? Surely those super routers cannot keep and search 2^32 entries.
I like your point about unintended consequences. Unfortunately I think IPv6 has large ones in terms of losing anonymity.
Sure, all of this _could_ work. But what if your ISP won't cooperate, or cannot, because Courts view it as "irresponsible" not to embed userid in IP?
You can use the active anonymity measures you describe. But you or tha forwarder will stand out like a sore thumb, and attract more sophisticated tracing. Right now anonymity isn't too bad because everyone has it to some extent. The herd protects those who really need anonymity.
Yes, you can be identified under IPv4. But it requires the cooperation of your ISP who has to grep through the logs. If they can find him before the logs rotate to/dev/null. Not easy to do, and certainly not on a large scale.
With IPv6, none of this rigamarole would be required. You would always be identified. So gone would be the casual anonymity has has helped the net grow. As you say, for "hard" [bulletproof?] anonymity, you have to go through logless anonymizers.
I'm sorry you find my comments nonsense, perhaps I can be more clear. I agree IPv6 _does_ have significant advantages for routing. But this thread isn't about future advantages, it's about disadvantages.
I'm not really talking about perfect anonymity either, but rather the ease of tracing a user that say surfs to a website. Currently, all the httpd knows is the 32 bit IP address. This is nearly useless because many other people use the same IP, and the user is unlikely to use that one again soon. Unless some part of the IPv6 is randomized, one (non-portable) machine will map to one IPv6 address. Maybe ISPs will still do some dynamic assignments on dial-ups. Quite possibly not if we don't do anything about it.
As for the 32 bit IPv4 "running out", I disagree. Yes, Class A's are gone, and B's and C's harder to get. So what? There are 4.16 _billion_ addresses available for ~200 million computers/routers. And it's not like there aren't any solutions. NAT/masq/proxying greatly expand the effective address space.
IPv6 certainly does help routing. But you _cannot_ just make up your MAC and have your router still recognize you. Cable systems use MAC for access control. I don't know about *DSL.
IIRC, IPv6 is 64 routing bits (mostly topo/geographical) , 32 bits for the ISP and 32 bits for your ISP to use as it sees fit. MAC/userid most likely.
I am hardly a consiracy theorist, but I can see the authorities in virtually all countries wanting the ability to track users without the cooperation of the ISPs. It is in the nature of police to want to know everything.
Well, yes of course *DSL and cable modems have effectively static IPs. But currently they are hardly worth tracking since the vast majority of users (not traffic) is dial-up. Frankly I see this continuing for quite some time. Many people's 'net usage is pretty light, just some email and information surfing that isn't worth an extra $40/mo to speed up.
I realize that IPv6 is _mostly_ about improving routing. And there's some reason for it. But on most cable systems, you cannot change your MAC and have the router still recognize your machine. Most are single user, or a most a few family members.
As for the current dynamic IP/dialup logging, of course my ISP's netadmin can trace me. But if someone else wants to, they have to ask by email/phone. That's not trivially easy, and the logs may have rotated;). My ISP would probably need some convincing.
The thing is, I know how to write software. I don't know how to manipulate the masses into calling for guarentees of freedom from their governments. I don't know how to make judges understand the issues the way I do.
I'm very sorry to say I don't think there is a way. The people in government, police & courts are mostly there because they have chosen it as a career. Why would they with the low pay? Because it satisfies their desire/need to control others. Freedom becomes as hard to defend as it is for individualists to unite!
I like your choice of words. 128 bits makes the hair on the back of my neck stand up.
The ISPs certainly could do proxying, and effectively not implement the userid bits, or dynamically (randomly) assign them.
But will they? Can you imagine AOL doing this? No! The current argument is that dynamic IP's are technologically necessary because the ISPs have 10-15 times more users than IP addresses.
But when IPv6 makes it feasible to identify users, ISP's will lose this defense. The police/Courts would likely view it as "irresponsible" not to identify users. Look for some small buried regulation that requires it.
IPv6 requires 128(!) bit IP addresses in place of the current 32 bit addresses. It is the end of anonymity on the 'net. Some of the extra bits will identify you as a particular user, others will be for routing to your location. HTTP cookies would become redundant.
I see user-selected anonymity as a cornerstone of the phenomenal growth of the 'net. When you want to read or write something that you fear would subject you meddling or critisism from busybodies, you can currently do so anonymously easily. Dial-up accounts and dynamic IP addresses make user tracing difficult.
Of course, there are always those control-addicts who will mention that anonymity assists criminals, and others who do things they don't like. That is the price of freedom. No-one ever said freedom was free, or even cheap. Unfortunately, control-addicts do not value freedom for others. By piercing anonymity, you stifle many activies.
The three days is _very_ odd. There are only two reasons for crashes that long: broken customized hardware, and lack of skilled technical people. Customized hardware should have spares on the shelf.
A shortage of skilled techs is very likely. The private sector offers much higher pay, and govt job security is only worth so much. So how is the US govt going to fix this?
Surely not by paying more. They might re-introduce the DRAFT (aka Selective Service) for a "War on Drugs" or a "War on Poverty" or "War on Pollution" or whatever. But they'd take all the technically proficient draftees and use them to service govt departments. Not sensitive areas like the NSA which would be filled by transfering long-term govt employees, but backfilling for them.
Perhaps a bit paranoid, but the price of freedom is vigilance.
This is very odd--Why would the NSA let us know their machines are down, and they can't get them up? Are they trying to lull us into complacency? Disinformation is an important tactic of spyies.
Now, I recognize the importance of National Security, and the role the NSA plays in it. But frankly, with ECHELON etc, they haven't convinced me that their actions are solely devoted to defending against foreign threats. That is the law, and I think they bend or break it.
This bothers me a bit too. The GPL covers "derivative work", although just how much code needs to be included to make something "derivitive" is not clear. And copyright doesn't protect ideas, just their expression. So someone could read your code, see a nifty idea/algorithm, and use it with impunity.
Detection is another problem. How are you going to know they used your code? They don't publish theirs! You could always try to figure out what compiler & options they used (but watch that EULA!], compile your code that way, and compare. Very tough.
Fortunately, once you really think they did steal GPL code and you're willing to start a lawsuit, things paradoxically get easier. IANAL, but you have subpoena power for depositions and discovery. They will have to produce their source code, and employees will have to answer were it came from.
Slashdot Mirror
This is _not_ a troll, but nor am I a karma whore either. In all seriousness, we should look to MS-Windows, specifically 95/NT for an easy simple, intuitive user interface. My son learnt it at age 5 without much instruction and he's now better than me.
Microsoft deserves credit where it is due. No-one doubts their marketing prowess, or ruthlessness either. I believe they have put considerable effort into UI design via expensive user panels and focus groups. They certainly haven't put effective effort into OS design, which has become recursive layers of kludge frequently going random. But there is no reason that we should not incorporate User Interface findings to the extent possible. Furthermore, it would ease transition.
To the specific question about OS incorporating user feedback, as a modest OS developer (cpuburn) I feel particularly qualified to comment. I get some user feedback, and if it seems like a good idea to me, I will incorporate it.
But if I don't like it, forget it, and I may be capricious. One MS-Windows user wanted start & stop buttons added. I firmly declined, even though I conceed it will improve ease-of-use. I prefer the user learn about TaskMan. If he wants buttons, let him add them. The code _is_ GPL.
I have used Linux since kernel 1.3.15 days, and at that time, Slackware was the most polished and complete distribution. For quite some time, it languished, and only recently has seen the maintenance/updating found on more recent distro's, notably RedHat.
/etc/rc.d init scripts. I detest the byzantine SysV init scripts found in RedHat.
I've used and continue to use RedHat extensively. It is certainly the best Linux distro for Alpha architecture. It is exceptionally easy for beginners to install, yet remains very powerful via the `rpm` mechanism of package and especially dependancy updating.
Yet Slackware still has powerful attractions. It can certainly be installed on less hard-disk and RAM than RedHat. It leads to better UNIX understanding [albeit less easily] by using an editor for sysadmin rather than `linuxconf` tools. Furthermore, I find the file layout easier to understand, especially the BSD-style
Yes, using anonymity _may_ reduce the impact of what is said. Particularly if there is no apparent need for anonymity. Paranoia doesn't look good.
But where the reader agrees with the need for anonymity [controversial/retributive material] then I don't think the impact is reduced much. Sometimes, publishing nominatively may appear foolhardy or risk/publicity seeking behaviour.
Hasn't the US Supreme Court defended anonymity as an important element of free speech? Without anonymity, there is always fear of retribution.
Not that Presidents or administrations have ever had much interest in supporting freedom. Even if they are sworn to defend the US Constitution, their main interest is in governing the populace. Freedoms make that job harder.
Do two wrongs make a right? Bezos' defense merely another form of the question. The trite answer "of course not" is naive, because in many situtations, two wrongs do seem to make a right. For example "The Prisoners Dilemma" in modern game theory.
More important is the fundamental question of intellectual property rights. IP has piggybacked onto physical property rights. But IP & RP are fundamentally different things, and the most important commercial use of IP, publishing, has no analog in the real-property world.
When you sell wheat, you aren't worried about someone copying it. You havent sold the farm. When you publish, you have sold the farm because copying is trivially easy. But you won't necessarily notice copying the way you would if someone stole your bushel of wheat.
-- Robert
Thanks for your reply. ctcm measures `movsd` which is a load and a store for each word. If I should have 4112 MB/s @ 539 MHz, that's 7.6 bytes read and 7.6 bytes written each clock. Not very likely unless I've got a 128bit path or dual ported SRAM. But you are right, ctcm is more a measure of bandwidth than latency.
I dug out my pseudrorandom access asm timer. I measure 10.7 Mreads/s from DRAM (9.1 busclocks), 20.0 Mreads/s from L2 (27 Celeron CPU clks) and 525 Mreads/s from L1 (1.03 CPU clks). So L1 seems single cycle, but L2 looks oddly slow, perhaps due to unintended thrashing.
As for the power/die budget, I'm afraid I don't know enough about chip feature design. But from all the micrographs I've seen, L1 is a fairly small portion of the die, so doubling it wouldn't be too painful. It also appears disproportionately large compared to L2, so something like this has probably been done.
I think the latencies you gave are for L2 cache. As for power budgets, I hardly think the P6 core is that great. But it _is_ alot better than an Alpha 21264 633 MHz with 107W (47Amps @ 2.35V).
As for SDRAM latency, I've measured ~9 Mreads/sec for pseudorandom P6 addresses. Now that _is_ 11 busclocks, but remember the P6 always accesses DRAM by full cache lines, so latency is (11-3)=7 busclocks (more for later bytes). So 42-66 CPU clocks if the multiplier is 6x.
As for L1 latency, I don't recall my read rates. But they'd have to be _very_ fast to allow ctcm to report 2700 MB/s @ 539 MHz in `movsd` to L1. This sounds like 1 clock/transfer to me, and the L1 might even be double (read & write ) ported.
-- Robert
First, high compliments on an outstanding and insightful article on one high-end of the computing business. It is easy to forget there are other aspects than the max-CPU performance sought by hobbyists.
The discussion of architectural performance benefits was very clear and insightful. There are obvious limits to multi-issue architectures.
A few corrections, if indeed I am correct: Main memory fetch is _not_ the oft-quoted "hundreds of CPU cycles". Typical SDRAM timing is 6-1-1-1, or 9 bus cycles per 32byte cache line. For a 600 MHz CPU with a 6x multiplier, this is 54 cycles, plus perhaps a few for page misses, etc.
Also, AFAIK at least Intel's P6 x86 core has a 1 CPU cycle latency L1 cache. Such a fast cache is necessary to make up for the risible shortage of x86 registers and helps considerably with stack-based operations such as often generated by `c` code. I do know that I can realize three RISC-type uops per clock cycle when 33-50% of the uops are loads from L1.
AFAIK, Jon Johansen is going to be tried in Norway. There is another DeCSS trial going on in the US against some websites.
But note how this MPAA tough-guy is afraid of admitting he even made a polite request for a teenager's arrest. He implies they did nothing at all, when the Norwegian prosecutor says otherwise.
I think it would be quite an unwelcome precedent if a Norwegian citizen were to be extradited to the US for breaking a US law while physically located in Norway. IANAL, but extradition treaties are all about returning fugitives, not extraterritorial law enforcement. Even wire fraud is prosecuted in the home country unless the fraudster willingly travels to the victim country.
Put it this way: There are some neoNazi groups in the US whose writings are illegal under German law. Do you think the US police could turn them over to the German police even if they wanted to? But it might not be too smart for them to travel to Germany. (But no-one has ever accused them of being smart).
If this man flew 51 missions in WW2, he's no baby boomer (usually considered the cohorts born 1946-1965). More like one of their parents, most likely born before 1927.
Regardless, he seems scared of being implicated with the arrest of the Norwegian teenager who authord DeCSS (Johanssen?). Most of his replies were evasive, as is customary among PHBs. But that one wasn't. I infer he's feeling some heat.
According to the Gigabit Ethernet group draft standard IEEE 802.3ab, Gigabit copper 1000base TXshould run on all decent Cat 5 installations.
It does this by running single duplex over all four pairs at 125 MHz. The coding is changed to increase the number bits per symbol from 0.8 to 1.25. Simple wiring screw-up like mixing-up tip & ring are already handled by most 100baseTX ethernet transceivers. But crossover-vs-not isn't, and split pairs are unfixable.
Your Cat5 working 100baseTX is supposed to run 1000baseTX just fine. But it won't if you've left pairs unconnected, or stole them for a second run or phone. Poor crimping might also hurt.
That said, the real question is what you can do with all that bandwidth. Most hard-disks cannot sustain even 10 MB/s that 100baseTX provides. And it's hardly a high spped internet solution. It only runs 100m from the hub. The real problem with internet has always been interbuilding: the last mile between cable heads and user buildings.
Well yes, there probably is less than 2^24 (16 M) entries. But that's not the problem. The real problem is how do you search 1-16M entries in the miniscule time you have available?
Figure that each of the pipes into a MAE is at least 1000 Mbit/s. The average packet might be 200 bytes = ~1600 bits. 1.6 ms per packet, or search 1M entries at 2500 GByte/s. Of course, you do parallel processing to reduce this, but you still need alot of fast cache SRAM.
I'm surprised this didn't happen earlier. Probably took awhile to draft the complaint fully.
IANAL, but AFAIK, It is almost routine to launch a countersuit. It helps settlement. And in this case it's more than justified.
I agree with you. I never said, and didn't mean to imply that IPv6 was started as any sort of conspiracy. It was plainly started with the loftiest of technical goals: improving routing and multicasting.
Routing is a _terrible_ job. Just think of what happens to a packet that arrives in MAE East. It came in one one pipe. Presumably it doesn't go back out on that one. But on which one of the other pipes does it go? Surely those super routers cannot keep and search 2^32 entries.
I like your point about unintended consequences. Unfortunately I think IPv6 has large ones in terms of losing anonymity.
Sure, all of this _could_ work. But what if your ISP won't cooperate, or cannot, because Courts view it as "irresponsible" not to embed userid in IP?
You can use the active anonymity measures you describe. But you or tha forwarder will stand out like a sore thumb, and attract more sophisticated tracing. Right now anonymity isn't too bad because everyone has it to some extent. The herd protects those who really need anonymity.
Yes, you can be identified under IPv4. But it requires the cooperation of your ISP who has to grep through the logs. If they can find him before the logs rotate to /dev/null. Not easy to do, and certainly not on a large scale.
With IPv6, none of this rigamarole would be required. You would always be identified. So gone would be the casual anonymity has has helped the net grow. As you say, for "hard" [bulletproof?] anonymity, you have to go through logless anonymizers.
I'm sorry you find my comments nonsense, perhaps I can be more clear. I agree IPv6 _does_ have significant advantages for routing. But this thread isn't about future advantages, it's about disadvantages.
I'm not really talking about perfect anonymity either, but rather the ease of tracing a user that say surfs to a website. Currently, all the httpd knows is the 32 bit IP address. This is nearly useless because many other people use the same IP, and the user is unlikely to use that one again soon. Unless some part of the IPv6 is randomized, one (non-portable) machine will map to one IPv6 address. Maybe ISPs will still do some dynamic assignments on dial-ups. Quite possibly not if we don't do anything about it.
As for the 32 bit IPv4 "running out", I disagree. Yes, Class A's are gone, and B's and C's harder to get. So what? There are 4.16 _billion_ addresses available for ~200 million computers/routers. And it's not like there aren't any solutions. NAT/masq/proxying greatly expand the effective address space.
IPv6 certainly does help routing. But you _cannot_ just make up your MAC and have your router still recognize you. Cable systems use MAC for access control. I don't know about *DSL.
IIRC, IPv6 is 64 routing bits (mostly topo/geographical) , 32 bits for the ISP and 32 bits for your ISP to use as it sees fit. MAC/userid most likely.
I am hardly a consiracy theorist, but I can see the authorities in virtually all countries wanting the ability to track users without the cooperation of the ISPs. It is in the nature of police to want to know everything.
Well, yes of course *DSL and cable modems have effectively static IPs. But currently they are hardly worth tracking since the vast majority of users (not traffic) is dial-up. Frankly I see this continuing for quite some time. Many people's 'net usage is pretty light, just some email and information surfing that isn't worth an extra $40/mo to speed up.
;). My ISP would probably need some convincing.
I realize that IPv6 is _mostly_ about improving routing. And there's some reason for it. But on most cable systems, you cannot change your MAC and have the router still recognize your machine. Most are single user, or a most a few family members.
As for the current dynamic IP/dialup logging, of course my ISP's netadmin can trace me. But if someone else wants to, they have to ask by email/phone. That's not trivially easy, and the logs may have rotated
I'm very sorry to say I don't think there is a way. The people in government, police & courts are mostly there because they have chosen it as a career. Why would they with the low pay? Because it satisfies their desire/need to control others. Freedom becomes as hard to defend as it is for individualists to unite!
I like your choice of words. 128 bits makes the hair on the back of my neck stand up.
The ISPs certainly could do proxying, and effectively not implement the userid bits, or dynamically (randomly) assign them.
But will they? Can you imagine AOL doing this? No! The current argument is that dynamic IP's are technologically necessary because the ISPs have 10-15 times more users than IP addresses.
But when IPv6 makes it feasible to identify users, ISP's will lose this defense. The police/Courts would likely view it as "irresponsible" not to identify users. Look for some small buried regulation that requires it.
IPv6 requires 128(!) bit IP addresses in place of the current 32 bit addresses. It is the end of anonymity on the 'net. Some of the extra bits will identify you as a particular user, others will be for routing to your location. HTTP cookies would become redundant.
I see user-selected anonymity as a cornerstone of the phenomenal growth of the 'net. When you want to read or write something that you fear would subject you meddling or critisism from busybodies, you can currently do so anonymously easily. Dial-up accounts and dynamic IP addresses make user tracing difficult.
Of course, there are always those control-addicts who will mention that anonymity assists criminals, and others who do things they don't like. That is the price of freedom. No-one ever said freedom was free, or even cheap. Unfortunately, control-addicts do not value freedom for others. By piercing anonymity, you stifle many activies.
The three days is _very_ odd. There are only two reasons for crashes that long: broken customized hardware, and lack of skilled technical people. Customized hardware should have spares on the shelf.
A shortage of skilled techs is very likely. The private sector offers much higher pay, and govt job security is only worth so much. So how is the US govt going to fix this?
Surely not by paying more. They might re-introduce the DRAFT (aka Selective Service) for a "War on Drugs" or a "War on Poverty" or "War on Pollution" or whatever. But they'd take all the technically proficient draftees and use them to service govt departments. Not sensitive areas like the NSA which would be filled by transfering long-term govt employees, but backfilling for them.
Perhaps a bit paranoid, but the price of freedom is vigilance.
-- Robert
This is very odd--Why would the NSA let us know their machines are down, and they can't get them up? Are they trying to lull us into complacency? Disinformation is an important tactic of spyies.
Now, I recognize the importance of National Security, and the role the NSA plays in it. But frankly, with ECHELON etc, they haven't convinced me that their actions are solely devoted to defending against foreign threats. That is the law, and I think they bend or break it.
-- Robert
This bothers me a bit too. The GPL covers "derivative work", although just how much code needs to be included to make something "derivitive" is not clear. And copyright doesn't protect ideas, just their expression. So someone could read your code, see a nifty idea/algorithm, and use it with impunity.
Detection is another problem. How are you going to know they used your code? They don't publish theirs! You could always try to figure out what compiler & options they used (but watch that EULA!], compile your code that way, and compare.
Very tough.
Fortunately, once you really think they did steal GPL code and you're willing to start a lawsuit, things paradoxically get easier. IANAL, but you have subpoena power for depositions and discovery. They will have to produce their source code, and employees will have to answer were it came from.
-- Robert