Aasimov was both a scientist and a prolific writer. Some of his short stories are essentially in the form of 'reports'. I think that they might make a good example to start with. The fact that "I Robot" came out as a movie last year would help pique at least some people's interest.
It just hit me, rereading the article, that what happened is that he picked up some spyware that is messing with his 'results' and making it look like his google searches are resulting in porn ads. If so, then the company that is running the spyware could find themselves joindered in as a third-party defendant...
The charges that I can see being rendered against the spyware company would include: interference with contractual relationships (with the people paying for the real targeted google ads), libel (making it look like google is serving porn), unfair business practices... and I'm sure that an actual lawyer (of which google probably has many) could come up with some even more interesting causes of action.
It's yet another sign that I'm right when I say that MS is run by the Marketing group. It's someone like that who would think that announcing a $1B marketing^w research budget would suddenly (and magically) increase the output from the development group.
It might do some good in marketing, where you can just go out an buy more time slots with the money, but R&D requires that you have people with ideas to develop and the skill to capitalize on the ideas. Many of those have walked away from MS long time ago, or just never bothered to walk in the door.
I mean: we're talking about a company that told the European Court that they don't understand Windows well enough to document it properly. -- Even if they were lying, it still bodes ill for their R&D plans.
My first experience with admining SUN boxes was around 1986. I had a grand scheme of networking large portions of the Biochemistry department starting with a small group of machines consisting of an old SUN-1 (serial #300 etched in the back by hand), an SGI box, and a VAX 750.
The documentation I found included the original documentation (from the early '80s) where SUN announced the 'The Network is the Computer' logo and declared that they would never again sell a machine without a network.
By the time I arrived on the scene, this attitude was so endemic to the company that I got into something of an email bitch-fight with a SUN sales rep who sold me an ethernet card without a MAC address EPROM. When I complained about this he responded that I should simply clone the address from the first ethernet card in the box.
When I told him that this was the first ethernet card for the box, he tried to explain to me how all SUN boxes were sold with Ethernet cards in them. I replied by telling him the serial number of the machine, and suggesting that he ask someone who was with the company when that machine was built "I suggest Bill Joy".
He quietly shipped the eeprom.
_____
The impressive thing about SUN's "the Network is the Computer" idea was not with thin clients -- but rather with the smart clients and fast central file-server model that allowed dataless and diskless desktop machines with lots of computing power on each desktop,
Among other things, this allowed the "one login -- any machine" environment where you could use your login on any UNIX box in the system that allowed you -- and could even move across CPU/Manufacturer lines with a minimal ammount of kludging. They manaaged to extend this to the point where you could fly across the country and log into a SUN box in New York and have minimal impact from the fact that your files were on a server in San Francisco (presuming that your cross-country network was reasonably fast.
The ubiquitous UNIX networking is also part of the reason why SUN became the a central part of the.com internet backbone... they had so much experience with providing rock-solid IP-based networking that you knew that you didn't have to worry about that part of your system.
(( Even though the original heart of UNIX was DEC, DEC insisted on their own big-iron server-based networking system and only moved to TCP/IP when it was clear that DecNet had lost the battle.)).
I think that SUN's eventual downfall was that they got stuck on the same seductive path as DEC and IBM -- that of trying to hold on to the high-cost high-profit margin world of big-iron. This was despite the fact that their entry into the market was at a relatively low end (albeit a $20K-60K low end). They climbed the ladder into the stratosphere of big iron, while effectively abandoning their original base of 'cheap' workstations and so left themselves vulnerable to the creeping featurism at the low end that they had originally mastered.
By the time they returned to tending the low end, it was too late -- Microsoft's termites had eaten their original base into an undependable sponge. Now, they have to re-establish that base, but against MS's millions of termites, and with their high-end being eaten into. It's not a fun position to be in.
We really owe the "millions of jobs" to Ken Thompson and Dennis Ritchie and Bell Labs for developing UNIX and making it available to academia virtually for free.
No. The people you owe that to is AT&T's lawyers. The reason why Unix became (pseudo) Open Source is that AT&T was a legislated monopoly. Part of their consent decree was that they couldn't go into markets outside of Telephones, and they couldn't supress technology.
This got them into a rather tight bind. When someone asked AT&T to send them a copy of Unix (so that they could use the chess program that was written on it), the lawyers tried to nix the sale complaining that they'd be sued for going outside the Unix market -- so they got sued for not releasing UNIX.
Having lost the lawsuit, they started selling UNIX systems and were promptly sued a second time -- this time for releasing UNIX. They also lost this second suit.
The lawyers looked at the seemingly conflicting decisions and found that, while they couldn't restrict the UNIX technology, neither could they market or support it.
Their solution was that -- for the appropriate price (depending on whether you were a university or company) and signing an interesting non-disclosure agreement, you got a tape dump of a running UNIX box (including source) and a hearty "good luck!". You could share source/fixes with other institutions who had a similar license (( which soon turned out to include just about every major university )), but not with that uninteresting portion of the universe known as "anybody else".
This managed to satisfy both lawsuits because they were now neither marketing/supporting UNIX nor keeping it closeted.
Thus it is that the pseudo-open-source nature of UNIX was a legal kludge, not a conscious plan on the part of Thompson, Ritchie or anybody else. I expect that, if AT&T had had their original way, they would have never released the source to UNIX -- and it would have thus remained a little-known, ill-supported niche system.
Hotmail is now (presumably) a heavily Windows based system, but (if you read between the lines of the link in the parent article) they did it by walking away from the normal windows eye-candy system and going to a UNIXesque nearly purely scripted environment -- I wouldn't be at all surprised to find out that it's mostly a cygwin based environment (what better way to migrate off of *BSD/Solaris when Mr. Bill mandates the migration?).
Do you ever wonder why Microsoft came out with all of thos Unix/Windows integration tools at the turn of the millenium? This might be the answer.
Actually, SGI's problems stem mainly from their decision to drop MIPS and pin their future to the Itanium. . . . ..
Another way of describing SGI's mistake is that they expected Intel to be technological leaders, as opposed to market leaders.
The last CPU that Intel produced that was decent in it's first (or even second) revision was the 8080 -- and even that was mostly because there was very little to compare it to at the time.
Then Zillog came up with the Z80 which Intel cloned with the 8085. There was nothing really good about the 8086 -- in fact my pet theory was then (and is still now) that IBM chose it because it was so badly designed that it would never be real competition for the cash cow that their/360 mainfraim line was (something that couldn't be said about the Motorola and National Semiconductor chips).
Intel then tried to produce a real 32 bit chip -- a marketing driven bastard child that died in infancy. The 80186 and 80286 were attempts to clean up the worst aspects of the 8086 without throwing out the baby with the bathwater, but turned out to be little more than a foul tasting soup.
The '386 solved the problem by emulating the 8086 16 bit mode and providing an entirely new (well, kinda) 32 bit engine, but it wasn't until the pentium that they finally got even that right.
As I remember it nobody came to be a respected mover in the workstation market using an Intel-made chip. SGI and Apple went with Motorola. SGI eventually bought MIPS, and Apple rode the 68000 family for a decade before moving to another Motorola chip. DEC came out with the much-respected Alpha, and IBM/Motorola came out with the RS/6000 -- all of which allowed them to ride out (more or less) the MS/Intell steamroller.
By the end of the '90s I think that it was becomming clear that AMD was better at producing 'Intel' chips than Intel was. The outcome of the 64 bit 'intel' wars was no big surprise to me.
Given that history, I would have been very wary of betting the future of my company on Intel producing an industry-leading chip. History shows that the main thing that the Intel monopoly had going for it was that it was the 'standard' chip for the "Wintel" platform.
This is, I think, one of the few areas where Dems and Reps both gather and feel very free to take pot-shots at each other -- and the pot shots tend to be far better aimed on average than what you tend to find in the outside world.
As a result both sides are going to feel like Slashdot is full of members of "the other side".
I get a sense that the normal course of events is that you usually have a high concentration of one side or the other. Those in the majority commiserate among themselves and only a few braver members of 'the minority' pipe up from time to time. Thus the normal political experience is "we are the natural majority and their side doesn't make much sense" but there are pockets of 'the other side' where you can't really speak your mind.
Slashdot is that oddity where both sides get a good raking over the coals (in part, I think, because of a reasonably strong foreign contingent who often think that they're both off the wall.
It will make life soooo much easier for counterfieting rings... Once you get the knack of how to make a good-looking counterfit, you can pretend that you're from anywhere in the country.
And you'll have a false sense of security, too -- most people aren't going to have the tools to reliably recognize most half-decent forgeries, so all you'll need is a half-decent fake, but -- because most people will know them as 'secure' IDs, they'll just be accepted
at face value.
Most importantly, however: Being able to positively identify someone after they blow themselves up doesn't do much to stop terrorism.
Even after he was arrested, Mousaui is still trying to get himself killed.
Unfortunately, part of the purpose of patents today is defensive -- people now use patents to generate the equivalent of a mexican standoff--- I agree not to sue you over my pool of patents, if you agree not to sue me with your pool of patents. At that point it barely even matters if many of the patents are questionable -- What matters is that you have some to trade for the truce.
Generating a patent database might make it easier for the owner of a patent to justify it's existence (look! It's not in the database!) while not providing the community a defensive pool.
The DHS, put out a Request For Information (RFI) looking for someone who had the technology to read ID tags from
25 feet away at 55MPH... Through the skin of a bus... All the passengers at once.
They seem to suggest that they only want it so that they can identify people stopped at border checkpoints.
The silly thing is that his big-endian/little-endian program would break on a 64 bit bigendian system which would return '0', and not 0x12 or 0x78. You can use -2 and check for *(unsigned char *)i == 0xff or 0xfe.. but, in that case.
Going off on a tangent:
I have no idea what a 36 bit signed-magnitude integer mainfraim (( Yeah, they really existed -- CDC made them )) would return for *(unsigned char *) (int)-2. It would probably be 0x80 or 0x40 -- but it might be 0x800 (CDC used 6 bit characters, and case shifting was done by using a second 'escape' character -- rather like unicode, so a 'char' might be either a 6 bit or a 12 bit unit -- or an 8 bit unit just to keep from choking every C program under the Sun)
It's possible that part of the reason why they're not releasing their code is that they've snarfed some GPL code to improve their proprietary drivers, and they don't want the Open Source community to find out.
One way around that problem would be to just release the specs. That would allow the FLOSS community to come out with a good quality Open Source driver without the company having to admit to their theft.
If the FLOSS code is good enough, they might even decide to modify it and release an Open Source driver for Windows (hehe !-).
Reverse engineering a complex piece of hardware is not easy. It takes time and energy that would be far better spent just writing code. Many years ago I tried to reverse engineer driving of a serial chip. I spend days working on it before I realized that I could go down to the store (this was pre-web) and get the spec sheet for the chip. (Turns out it was a
6551)
Even though I'd already done most of the reverse engineering, the extra notes in the spec allowed me to improve my program greatly -- and with much assurance.
Graphics chips, on the other hand, are far more complex creatures, so it's going to be even easier to figure out that if you turn on this bit here after you turn off that bit over there, you enable the high-speed texturing using that, otherwise seemingly useless block of memory.
Once you have the docs, it's easy. Until then, you can spend months trying to figure out how to enable that new capability, and a day or two implementing what you just figured out.
On the other hand, chip manufacturers can gain from releasing their code to the public because the OS community can often take over the job of supporting the device, and sometimes figure out how to do things that would shock the original designers.
If nothing else, just release the specs, and see what the open source geeks can do. It may seem like there's no ability, but all it really takes is an army of mediocre geeks to do the bulk of the coding and one or two uber-geeks to tweak that last 5-10% that makes the difference between an ok driver and a really good one.
No one has died of a radiation-related accident in the history of the U.S. civilian nuclear reactor program.
Does infant mortality and miscarriages count? Child deaths and miscarriages apparently spiked around TMI after the accident there. -- Granted, it's not as dramatic as Homer's kid choking on a radioactive nugget, but the numbers were very statistically significant.
In other words, you don't have to go as far as wood river.
It's nice to see some greens finally start accepting what has been obvious to some of us for 30 or 40 years.
"Some greens" have always seen Nuclear power as a good idea (check my domain name). There's been a low-level dispute about whether or not the upsides of nuclear power exceed it's downsides.
As the disasterous implications of global warming have loomed ever larger, the downsides of nuclear power have started to loose their bite.
The 'badness' of Nuclear power has always been one of preference (or lack thereof), that has gotten, for some, to the point of dogma.
It started as a 'mom and apple pie' hysteric fear -- in the times of fallout and bomb shelters and relatively fresh pictures of the horrors of Hiroshima and Nagaski, where "nuclear" was most often coupled with "weapons". It was all to easy to mount a convincing attack on the problems of nuclear power, since the problems were really there, and the hysteric fear of atomic weapons was enough to flip what should have been a practical argument into "oooh, nukes! keep them away from me".
There was also the unspoken background that many US power plants were actually breeder nuclear reactors that were used as much to help manufacture weaponry as to generate power (( and weren't necessarily all to good at the latter, from some conversations that I've had with people who lived near them )).
The downsides of nuclear power have not gone away... it's just starting to look more and more like the lesser of two evils.
summary: anti-nuke was always a policy decision, not a scientific theory.
Global warming, on the other hand is a scientific theory that has been slowly worked it's way from 'interesting flake idea' to 'pretty much a proven fact' over the space of about 30+ years. There is roughly zero probability of a "Perry Mason Moment(tm)" where 10,000 scientists bread down on the stand and admit that it was all just an elaborate hoax.
If you've got a half dozen (or dozens of) people who have root login, and something goes on, all you have with a quick look at the log is that someone logged in at 4:15 pm and the system choked at 4:18.
If you force people to login as themselves and then SU (or sudo), then you know who was on the system with root when the system snarked ( And, if they use sudo instead of su, you can even have logs of the commands they used) -- it cuts down on the number of people you have to interview in the rush to figure out what broke the system.
Then, there's also the fact that if someone tricks you into doing something 'bad', it's less likely to catch you flatfooted as root if the only commands you exeute as 'root' are the ones that really need root.
As a last point: If you disable root logins (especially remote root logins), then a hacker needs to hunt down two passwords to get root access -- one for remote access and the other to get root.
Security isn't about making it impossible for an intruder to get in -- It's about making it hard enough that an attacker gives up and goes away -- even if they just go find an easier target (hi bill!).
So guys.... What was with the almost-daylong outage? Are we gonna get a description? I mean, we're a bunch of geeks here, so whatever it is that froze out slashdot for half a day is gonna be of some interest to us.
If you set it up so that LEDs are on in 10ms bursts, then there should be no real problem with sucking data out of them. If you leave them directly connected to the data flow, then yeah --- you're asking for trouble..
As for the comment about "who considers unencrypted traffic public", it's one thing to whisper 'cute' things to your girlfriend at a public phone. It's another to have it broadcast over the PA system. Although both are 'public', there's a difference in the nature of the beast.
It's silly to ask for trouble.
On the other hand, if the entire network is unencrypted (or simple WEP), then you've got easier ways to get at the data.
Slashdot Mirror
Yeah.. You didn't necessarily get a whole lot done doing an all-nighter with a computer, but it was often a whole lot more enjoyable!
Aasimov was both a scientist and a prolific writer. Some of his short stories are essentially in the form of 'reports'. I think that they might make a good example to start with. The fact that "I Robot" came out as a movie last year would help pique at least some people's interest.
You might be able to claim responsibility for the post by the method noted above, but you {would,should} never get the karma points.
The charges that I can see being rendered against the spyware company would include: interference with contractual relationships (with the people paying for the real targeted google ads), libel (making it look like google is serving porn), unfair business practices ... and I'm sure that an actual lawyer (of which google probably has many) could come up with some even more interesting causes of action.
His wife caught him "doing some research", and now he has to cover his butt by spending thousands of dollars on a frivolous lawsuit.
It might do some good in marketing, where you can just go out an buy more time slots with the money, but R&D requires that you have people with ideas to develop and the skill to capitalize on the ideas. Many of those have walked away from MS long time ago, or just never bothered to walk in the door.
I mean: we're talking about a company that told the European Court that they don't understand Windows well enough to document it properly. -- Even if they were lying, it still bodes ill for their R&D plans.
The documentation I found included the original documentation (from the early '80s) where SUN announced the 'The Network is the Computer' logo and declared that they would never again sell a machine without a network.
By the time I arrived on the scene, this attitude was so endemic to the company that I got into something of an email bitch-fight with a SUN sales rep who sold me an ethernet card without a MAC address EPROM. When I complained about this he responded that I should simply clone the address from the first ethernet card in the box.
When I told him that this was the first ethernet card for the box, he tried to explain to me how all SUN boxes were sold with Ethernet cards in them. I replied by telling him the serial number of the machine, and suggesting that he ask someone who was with the company when that machine was built "I suggest Bill Joy".
He quietly shipped the eeprom.
_____
The impressive thing about SUN's "the Network is the Computer" idea was not with thin clients -- but rather with the smart clients and fast central file-server model that allowed dataless and diskless desktop machines with lots of computing power on each desktop,
Among other things, this allowed the "one login -- any machine" environment where you could use your login on any UNIX box in the system that allowed you -- and could even move across CPU/Manufacturer lines with a minimal ammount of kludging. They manaaged to extend this to the point where you could fly across the country and log into a SUN box in New York and have minimal impact from the fact that your files were on a server in San Francisco (presuming that your cross-country network was reasonably fast.
The ubiquitous UNIX networking is also part of the reason why SUN became the a central part of the .com internet backbone... they had so much experience with providing rock-solid IP-based networking that you knew that you didn't have to worry about that part of your system.
(( Even though the original heart of UNIX was DEC, DEC insisted on their own big-iron server-based networking system and only moved to TCP/IP when it was clear that DecNet had lost the battle.)).
I think that SUN's eventual downfall was that they got stuck on the same seductive path as DEC and IBM -- that of trying to hold on to the high-cost high-profit margin world of big-iron. This was despite the fact that their entry into the market was at a relatively low end (albeit a $20K-60K low end). They climbed the ladder into the stratosphere of big iron, while effectively abandoning their original base of 'cheap' workstations and so left themselves vulnerable to the creeping featurism at the low end that they had originally mastered.
By the time they returned to tending the low end, it was too late -- Microsoft's termites had eaten their original base into an undependable sponge. Now, they have to re-establish that base, but against MS's millions of termites, and with their high-end being eaten into. It's not a fun position to be in.
No. The people you owe that to is AT&T's lawyers. The reason why Unix became (pseudo) Open Source is that AT&T was a legislated monopoly. Part of their consent decree was that they couldn't go into markets outside of Telephones, and they couldn't supress technology.
This got them into a rather tight bind. When someone asked AT&T to send them a copy of Unix (so that they could use the chess program that was written on it), the lawyers tried to nix the sale complaining that they'd be sued for going outside the Unix market -- so they got sued for not releasing UNIX.
Having lost the lawsuit, they started selling UNIX systems and were promptly sued a second time -- this time for releasing UNIX. They also lost this second suit.
The lawyers looked at the seemingly conflicting decisions and found that, while they couldn't restrict the UNIX technology, neither could they market or support it.
Their solution was that -- for the appropriate price (depending on whether you were a university or company) and signing an interesting non-disclosure agreement, you got a tape dump of a running UNIX box (including source) and a hearty "good luck!". You could share source/fixes with other institutions who had a similar license (( which soon turned out to include just about every major university )), but not with that uninteresting portion of the universe known as "anybody else".
This managed to satisfy both lawsuits because they were now neither marketing/supporting UNIX nor keeping it closeted.
Thus it is that the pseudo-open-source nature of UNIX was a legal kludge, not a conscious plan on the part of Thompson, Ritchie or anybody else. I expect that, if AT&T had had their original way, they would have never released the source to UNIX -- and it would have thus remained a little-known, ill-supported niche system.
How about a new one???
Do you ever wonder why Microsoft came out with all of thos Unix/Windows integration tools at the turn of the millenium? This might be the answer.
Another way of describing SGI's mistake is that they expected Intel to be technological leaders, as opposed to market leaders.
The last CPU that Intel produced that was decent in it's first (or even second) revision was the 8080 -- and even that was mostly because there was very little to compare it to at the time.
Then Zillog came up with the Z80 which Intel cloned with the 8085. There was nothing really good about the 8086 -- in fact my pet theory was then (and is still now) that IBM chose it because it was so badly designed that it would never be real competition for the cash cow that their /360 mainfraim line was (something that couldn't be said about the Motorola and National Semiconductor chips).
Intel then tried to produce a real 32 bit chip -- a marketing driven bastard child that died in infancy. The 80186 and 80286 were attempts to clean up the worst aspects of the 8086 without throwing out the baby with the bathwater, but turned out to be little more than a foul tasting soup.
The '386 solved the problem by emulating the 8086 16 bit mode and providing an entirely new (well, kinda) 32 bit engine, but it wasn't until the pentium that they finally got even that right.
As I remember it nobody came to be a respected mover in the workstation market using an Intel-made chip. SGI and Apple went with Motorola. SGI eventually bought MIPS, and Apple rode the 68000 family for a decade before moving to another Motorola chip. DEC came out with the much-respected Alpha, and IBM/Motorola came out with the RS/6000 -- all of which allowed them to ride out (more or less) the MS/Intell steamroller.
By the end of the '90s I think that it was becomming clear that AMD was better at producing 'Intel' chips than Intel was. The outcome of the 64 bit 'intel' wars was no big surprise to me.
Given that history, I would have been very wary of betting the future of my company on Intel producing an industry-leading chip. History shows that the main thing that the Intel monopoly had going for it was that it was the 'standard' chip for the "Wintel" platform.
As a result both sides are going to feel like Slashdot is full of members of "the other side".
I get a sense that the normal course of events is that you usually have a high concentration of one side or the other. Those in the majority commiserate among themselves and only a few braver members of 'the minority' pipe up from time to time. Thus the normal political experience is "we are the natural majority and their side doesn't make much sense" but there are pockets of 'the other side' where you can't really speak your mind.
Slashdot is that oddity where both sides get a good raking over the coals (in part, I think, because of a reasonably strong foreign contingent who often think that they're both off the wall.
And you'll have a false sense of security, too -- most people aren't going to have the tools to reliably recognize most half-decent forgeries, so all you'll need is a half-decent fake, but -- because most people will know them as 'secure' IDs, they'll just be accepted at face value.
Most importantly, however: Being able to positively identify someone after they blow themselves up doesn't do much to stop terrorism.
Even after he was arrested, Mousaui is still trying to get himself killed.
Generating a patent database might make it easier for the owner of a patent to justify it's existence (look! It's not in the database!) while not providing the community a defensive pool.
Probably 12/36/36, since characters are a multiple of 6 bits (either 6 or 12) and a 24 bit short just doesn't seem to make much sense.
They seem to suggest that they only want it so that they can identify people stopped at border checkpoints.
Going off on a tangent:
I have no idea what a 36 bit signed-magnitude integer mainfraim (( Yeah, they really existed -- CDC made them )) would return for *(unsigned char *) (int)-2. It would probably be 0x80 or 0x40 -- but it might be 0x800 (CDC used 6 bit characters, and case shifting was done by using a second 'escape' character -- rather like unicode, so a 'char' might be either a 6 bit or a 12 bit unit -- or an 8 bit unit just to keep from choking every C program under the Sun)
I bought the OS, too. I may not own the copyright, but I did buy the copy.
One way around that problem would be to just release the specs. That would allow the FLOSS community to come out with a good quality Open Source driver without the company having to admit to their theft.
If the FLOSS code is good enough, they might even decide to modify it and release an Open Source driver for Windows (hehe !-).
Even though I'd already done most of the reverse engineering, the extra notes in the spec allowed me to improve my program greatly -- and with much assurance.
Graphics chips, on the other hand, are far more complex creatures, so it's going to be even easier to figure out that if you turn on this bit here after you turn off that bit over there, you enable the high-speed texturing using that, otherwise seemingly useless block of memory.
Once you have the docs, it's easy. Until then, you can spend months trying to figure out how to enable that new capability, and a day or two implementing what you just figured out.
On the other hand, chip manufacturers can gain from releasing their code to the public because the OS community can often take over the job of supporting the device, and sometimes figure out how to do things that would shock the original designers.
If nothing else, just release the specs, and see what the open source geeks can do. It may seem like there's no ability, but all it really takes is an army of mediocre geeks to do the bulk of the coding and one or two uber-geeks to tweak that last 5-10% that makes the difference between an ok driver and a really good one.
That's how Open Source works.
Does infant mortality and miscarriages count? Child deaths and miscarriages apparently spiked around TMI after the accident there. -- Granted, it's not as dramatic as Homer's kid choking on a radioactive nugget, but the numbers were very statistically significant.
In other words, you don't have to go as far as wood river.
"Some greens" have always seen Nuclear power as a good idea (check my domain name). There's been a low-level dispute about whether or not the upsides of nuclear power exceed it's downsides.
As the disasterous implications of global warming have loomed ever larger, the downsides of nuclear power have started to loose their bite.
The 'badness' of Nuclear power has always been one of preference (or lack thereof), that has gotten, for some, to the point of dogma. It started as a 'mom and apple pie' hysteric fear -- in the times of fallout and bomb shelters and relatively fresh pictures of the horrors of Hiroshima and Nagaski, where "nuclear" was most often coupled with "weapons". It was all to easy to mount a convincing attack on the problems of nuclear power, since the problems were really there, and the hysteric fear of atomic weapons was enough to flip what should have been a practical argument into "oooh, nukes! keep them away from me".
There was also the unspoken background that many US power plants were actually breeder nuclear reactors that were used as much to help manufacture weaponry as to generate power (( and weren't necessarily all to good at the latter, from some conversations that I've had with people who lived near them )).
The downsides of nuclear power have not gone away... it's just starting to look more and more like the lesser of two evils.
summary: anti-nuke was always a policy decision, not a scientific theory.
Global warming, on the other hand is a scientific theory that has been slowly worked it's way from 'interesting flake idea' to 'pretty much a proven fact' over the space of about 30+ years. There is roughly zero probability of a "Perry Mason Moment(tm)" where 10,000 scientists bread down on the stand and admit that it was all just an elaborate hoax.
If you force people to login as themselves and then SU (or sudo), then you know who was on the system with root when the system snarked ( And, if they use sudo instead of su, you can even have logs of the commands they used) -- it cuts down on the number of people you have to interview in the rush to figure out what broke the system.
Then, there's also the fact that if someone tricks you into doing something 'bad', it's less likely to catch you flatfooted as root if the only commands you exeute as 'root' are the ones that really need root.
As a last point: If you disable root logins (especially remote root logins), then a hacker needs to hunt down two passwords to get root access -- one for remote access and the other to get root.
Security isn't about making it impossible for an intruder to get in -- It's about making it hard enough that an attacker gives up and goes away -- even if they just go find an easier target (hi bill!).
So guys.... What was with the almost-daylong outage? Are we gonna get a description? I mean, we're a bunch of geeks here, so whatever it is that froze out slashdot for half a day is gonna be of some interest to us.
As for the comment about "who considers unencrypted traffic public", it's one thing to whisper 'cute' things to your girlfriend at a public phone. It's another to have it broadcast over the PA system. Although both are 'public', there's a difference in the nature of the beast.
It's silly to ask for trouble.
On the other hand, if the entire network is unencrypted (or simple WEP), then you've got easier ways to get at the data.