At work, I'm building a 200 TB storage system for a particle accelerator. It is a small part of a much larger grid that will eventually need about 3 - 5 petabytes of online disk space, and about twice that much on tape for backups.
I see myself buying a lot of these when the enterprise version comes out. Heck, with current systems taken into account, I'd only need about 3 SATAbeasts (48-disk enclosure) with these suckers in it.
Rule of thumb in computing: There is *always* someone out there needing more capacity, whether it is CPU, memory, disk, or bandwidth.
Most of the Java-related speed problems fit in one of three categories:
1) VM start-up time. 2) Over-architecting objects. OO languages tend to encourage you to make everything into an poorly-designed object. If you are doing numerical work and you end up creating millions of objects, your doing something wrong. 3) Crappy GUI libraries. Makes the UI miserable for many people (it is getting better, but slowly). If this is your interaction with Java, I understand your dislike for it.
If you apply the language to well-architected server programs, it's great.
Explorer automatically suggests downloading the substantially more capable FREE Word Viewer
Yes, but what if I want to edit a simple word document? What if I am working on a group project, and someone sends me the final report ridden with typos?
I guess I would just have to open it up with Mac OS X's built in "TextEditor" - which reads HTML,.doc,.rtf, and plain-old-text.
What is the world when all Apple computers ship with a basic.doc editor but new Windows computers don't?
You say that, but then I - as a member of that other 5% of the population - absolutely crave the ability to watch movies on my screen in high def. I *can* see the difference, and I think it's great.
However, no matter how great I think it is, I'm not dumb enough to buy a HDDVD/BluRay right now. Everytime I watch a DVD on my TV, I curse the consumer electronics industry for not being able to come up with a sane solution to fit my needs.
As much as it is ignored by 95% percent of the population, the HD-DVD/BluRay war is a deep disappointment for the other 5%.
No offense, but Java would be a particularly bad choice for this application. The real work done here is "gluing" two things together at a system level - the Roomba program and the Wii program. Java is uniquely miserable at interacting on the system level. Where Java excels is object-oriented architectural design and huge enterprise-level programs.
So, if you want a scalable application server, use Java. If you need to tie two programs together, use a scripting language like Python or Perl. If you know you can limit yourself to a Unix shell environment, Perl may be best (and this is from a Python junkie!)
According to this Vista is going to bring 100,000 new jobs to the US.
The question here is, of course, which one of these will be true:
1) Vista will unlock new potential markets for companies, allowing them to hire new programmers to add features to existing products or create new ones.
2) Vista will increase the barrier of entry for programs, meaning 100,000 new jobs will be created just to be able to support it, even in the absence of new features.
As someone who every so often has to see Windows when he visits home to fix things, I hope for the former. That would imply that Vista was done nicely, and I don't have to bend-over-backwards fixing my family's computers.
Meanwhile, I'll continue on with my Mac laptop, Linux servers, and continue to live without Windows. I don't think I'm missing much.
If something has an interior, that to me at least would logically mean there is an exterior. What's on the exterior?
False, in general (I'm not sure about this specific case, however). Go look up a Klein bottle. It is a mathematical object with no defineable interior or exterior (which makes it very hard to integrate over, in the traditional sense. Damned non-orientable objects.)
The first PotC was such an awesome movie because Disney thought it would suck. They half-made it, and tossed it aside. Because they thought it would fail, they let the director and Johnny Depp do their thing - no market droids wanted to touch it.
Then, when it was all wrapped up and finished, they watched it and said "Damn, this is pretty good." But, before they could market the hell out of it and reinsert more special effects, the thing came out.
You could tell Disney's initial plan was to dump it because it was a "builder" - the thing made increasing amounts of money over the summer.
Goes to show how good movies can be if the market droids leave it alone.
I still think the EU (and Russia) has a long way before they "get over it" with respect to WWII, like the GP suggested.
I've visited mass graves in St. Petersburg. I've talked to people who lost family in the war. There's still a lot of emotion tied up into that war.
To too many (but not all, hopefully) US citizens, WWII was the "great adventure" that their grandfather or great-grandfather went on, and came back with a lot of stories about.
To many people in Europe/Russia, WWII was something that may have closely affected the town that they grew up in. Emotional wounds don't just get up and heal because we are in the Computer Age.
That's a disappointing article from Joel, he's usually more observant than that. I'd expect him to be able to make the distinction between "simplicity of an application" and "simplicity of an interface". You can have a feature-rich application which has an extremely simple interface.
My Mac laptop has a simple interface that both me and my wife enjoy. However, it is perfectly as functional as my linux desktop, who is much more complex.
An iPod's interface is simple; the device itself is complex. Same with gmail.
Both authors make the mistake of equating "ease of use" with "lack of features".
A good college shouldn't expect you to know HOW to program in C++. A Good College should teach how to program first and foremost, where the example language is C++.
I had friends in Georgia Tech who were decent Java programmers who did miserable in their introductory programming classes because the professor chose an extremely obscure language that no one knew beforehand. This way, he knew that no one came in who knew programming, but didn't know the concepts. By choosing a weird language, he could force concepts first, specific languages later. They hated it, got a poor grade, but came out better programmers.
On the same note, a mathematician does not differentiate between solutions of ax^2 + bx + c = 0 and x^2 + 5x + 1 = 0; knowing how to solve the quadratic equation is the important part, the second is just an example to make the theory easier.
They spent all their time learning about useless crap like advanced multivariable calculus, matrix theory, and other math crap instead of learning how to program.
You are making the mistake of mixing up a trade school programming degree with a CS degree. If you want to learn skills which will allow you to get an entry-level programming job today, go to the trade school.
If you want to learn how to approach the profession with a scientific outlook, which will enable you to get an entry-level programming job today and get a new one tomorrow when industry makes a major shift in languages or paradigms, go to the university. The math will make your programming better. The multivariate calc is possibly the least useful math class a CS major could take. More important is set theory, matrix theory, probability, and statistics.
Another way to evaluate degrees is this: were you taught how to think and program, or how to write Java? If Java becomes dead in 5 years, the person who learned how to program will shrug his shoulders and move on. The person who learned how to write Java will become unemployed. (Note: I'm just picking on Java; replace it with language X, and the statement holds true).
The Internet2 is the research network that universities use for high-bandwidth needs. Universities formed a consortium so there would be a fancy organization that NSF would fund network connections and that they could leverage their size to get cheaper bandwidth.
Buying a 10Gbps line from I2 that travels over Level3's network is about 10x cheaper than buying that bandwidth through L3 directly. Of course, it can be damned expensive to buy up the fiber from your university to one of the I2 hubs. Further, I2 doesn't like you to connect directly to their hubs - they prefer to connect through a "regional network". So, the Great Plains region, for example, set up a regional network called the GPN, and then the GPN buys the connection to Abilene (the name of the current I2 network).
There's some money available for research into networking technologies such as the connection-oriented software mentioned above, but most of the traffic is simple IPV4 (and some IPV6).
So, for some people, the I2 means that they can turn over 200 TB of disk cache every 2 weeks or so (which are the requirements for some sites associated with the Large Hadron Collider, like us). For others, it means that they can stream classes in HDTV across the United States.
For another good chunk of users, it means that they can trade movies/songs at gigabit speeds, if the p2p application can support that. Most can't. It still takes a bit of networking black-magic to get line-speed for gigabit ports, but not nearly as much as it used to.
Our office bought the LVM-42W2 for video conferencing over the summer. Since then, 4 of us have bought the exact same model. It's got tons of inputs (all the various analog ones, 1 VGA, 2 DVI, 1 HDMI). It can do 1080p. It is cheap - finding it for $1500 is not hard, I think. I haven't run into any quirks.
The difference between the Westinghouse and the $3000 Samsung is that the Samsung has lots of nice filters on it, whereas the Westinghouse only has the standard brightness/contrast/etc. Three points: 1) Your 1080p/1080i source doesn't need any expensive upconverting filter technologies. 2) You'll want a nicer up-converting image for DVD sources. This can be remedied by buying a nice $100 DVD player which does the up-conversion, instead of having the TV do it. 3) Unless you have lots of nice TVs at home already, you won't be able to tell the difference between the Westinghouse and a $3000 set once you get it in your living room. The only way to see that the $3000 set has a marginally better picture is to put them next to each other. So, the extra $1500 in cost goes away once you take the set home, and in the worst case can be remedied by buying a nice DVD player (cost: $100).
I friggin' love my TV and, at $1500, my wife even let my buy it.
You can start a case with a bad patent, but you won't get very far.
SCO started a case with, what, nothing but a good press release? After several years of litigation, they still have nothing to show, but the trial date is "sometime next year".
You can drag patent cases on indefinitely. The last statistic I heard is that it cost $1 million to fight a patent case; I bet you can multiply that by many times if you're facing MS or IBM.
Still, in the end, if I was OO.o, I'd be flattered that some other company wanted to dedicate programmers to my project. However, I'd still be wary about including it.
Finally, the.doc format is not all that old. It really became firm in, what, the 1990s? That plus 20 years is 2010. This is assuming that MS has not patented any part of the format in the 90's or early 2000's, when the "patent everything" explosion started. I think that's a poor assumption.
No, if a major patent holder really wanted to, they could keep the Open Source community up to its ears in lawsuits indefinitely.
First of all, note that this is not a fork of the code. Novell is developing a plugin to read the OpenXML format, a Microsoft format.
Let me repeat, They are not forking OpenOffice.
Hell, the sourceforge project is called "odf-converter", not "Novell's evil plan for OO.o".
Further, the only way that I could read the press release from Novell in order to interpret it as "Novell is forking OpenOffice.org" is by the sentence which refers to the current OpenOffice.org product as "Novell's OpenOffice.org". That sounds more like a marketing intern not understanding how OOo and open source works out, not a secret decision on Novell's part.
Finally, I really hate the attitude that many of those contributing to Slashdot has taken toward Novell's current projects. It's fairly one-sided. They are not violating the law. They are not violating the GPL. They are not violating the spirit of the GPL.
The point of the GPL is that anyone can take your code, change it, and redistribute it, as long as they follow the rules. You can't make a distinction between people redistributing your software who you like and those who you don't like.
There's a lot of you who are sounding like Bush-style Republicans who want free speech for themselves, but not for those saying things they don't agree with. I bet a lot of you beating up on Novell today for taking advantage of the GPL are the same who beat up on Newt Gingrich the other day when he wanted to restrict free speech on the Internet. Hypocrites.
If you don't like Novell's contributions, don't accept them; if you think Novell is trying to get OpenXML into OO.o so MS can sue RedHat for patent infringement, think again. I doubt OpenXML is any more patent-ridden than the.doc format, or that there aren't any patent violations in the Linux kernel or OO.o already.
In other words, Novell can't paint any bigger target on Linux's back than there already is. MS and IBM have so many ambiguous patents that they can sue any Linux user for the indefinite future.
Believe it or not, Novell may just be trying to differentiate its product so people would buy it over their competitor's product. You know, effectively compete in the business world. That sort of thing.
Groklaw used to be a place where I could get a detailed analysis of legal issues I didn't understand. Now, it seems to have disintegrated into blind zealotry. Maybe they were trying to be funny in the article, and I just didn't get the joke...
But then, you have to gather a statistically significant number of these to rule out luck. So, you'd need to get those right several times to validate your model. Accounting for varing CO2 emissions, of course, complicates it. I doubt there's enough evidence time-history (following a previous prediction) to falsify anyone's theory. That's the problem.
It may be that I'm just the crazy mathematician, but if you get published in a well-respected journal, your results should be statistically significant. If your results are statistically significant, by the way the mathematics works, you had to have gathered enough evidence to prove yourself false, but the data actually proved your hypothesis true.
So, unless the journals in the field are complete crap, there's enough evidence to falsify the results about the historical climate. However, where things get a bit fuzzy is when you try and apply historical data to predict the future. The real problem is that no one knows how exactly to compare the past with the future; does the earth do the same thing, or have humans altered the climate in some way that the earth is going to do something crazy over the next climate cycle?
That's the real tough nut to crack. You get some very loud people saying "It's part of the natural climate cycle, human CO2 is just noise in a larger trend" and some others saying "We're fundamentally altering the climate dynamics of the planet". Without a either a time machine or a second planet Earth to compare the trends to, there is no way we can falsify the theories which apply the results to the future.
specifically seeking this set of conditions, and ignoring unknown data that simply doesnt fit, I think that's somewhat biased...
No, not quite -
It's like you want to study the eating habits of 500 lb people. However, there's just too many skinny people to record. So, you ignore the eating habits of people under 400 lbs. Sure, it's possible that all the scales in the world are wrong and all people who are weighed at 100 lbs are actually 500 lbs, but that risk is considered acceptable.
The hardware level triggers and software reconstructors of CMS and ATLAS will cull out low energy events. There's an extremely small probability that something interesting in low energy events (1 in a trillion?); there's a better probability (1/10000 perhaps) that something interesting happened in high energy collisions. If you had to throw one out due to the amount of storage space you had available, which would you record?
The way that I heard it is that every single war of aggression since the dawn of the industrial age has been a loss for the aggressor.
I, for one, can't think of any major conflicts which break this rule. Iraq I, Iraq II, WWI, WWII, Vietnam, Afghanistan (Britain and Russia), Korea. I'm sure there are other conflicts I am forgetting; maybe some other country breaks this rule.
The third Afghani war (US in Afghanistan) looked like a sure victory for the aggressor until this year. Now it's looking a little shaky.
Simply put, we don't have any way to effectively fight a determined insurgency.
An increase of only a couple of degrees of temperature can cause massive, and I mean massive, problems. They expect to loose a couple of magnets here and there, but the loss of the detector would be devastating. Not only does the magnet get destroyed, but there's enough energy in the beam (which would no longer be focused) to punch big holes in walls. This is why it is fairly deep underground.
To guard against this (at least in the sister project, CMS), they have 4 huge tanks (about the capacity of two semi-trucks, or more) of liquid helium sitting right above ground. If there's ever an emergency loss of power, these will flood the magnets with liquid helium to protect the magnets.
There's a reason that this project cost $8 billion and is billed as the most complex scientific experiment in the history of man.
Slashdot Mirror
At work, I'm building a 200 TB storage system for a particle accelerator. It is a small part of a much larger grid that will eventually need about 3 - 5 petabytes of online disk space, and about twice that much on tape for backups.
I see myself buying a lot of these when the enterprise version comes out. Heck, with current systems taken into account, I'd only need about 3 SATAbeasts (48-disk enclosure) with these suckers in it.
Rule of thumb in computing: There is *always* someone out there needing more capacity, whether it is CPU, memory, disk, or bandwidth.
Most of the Java-related speed problems fit in one of three categories:
1) VM start-up time.
2) Over-architecting objects. OO languages tend to encourage you to make everything into an poorly-designed object. If you are doing numerical work and you end up creating millions of objects, your doing something wrong.
3) Crappy GUI libraries. Makes the UI miserable for many people (it is getting better, but slowly). If this is your interaction with Java, I understand your dislike for it.
If you apply the language to well-architected server programs, it's great.
Yes, but what if I want to edit a simple word document? What if I am working on a group project, and someone sends me the final report ridden with typos?
I guess I would just have to open it up with Mac OS X's built in "TextEditor" - which reads HTML,
What is the world when all Apple computers ship with a basic
You say that, but then I - as a member of that other 5% of the population - absolutely crave the ability to watch movies on my screen in high def. I *can* see the difference, and I think it's great.
However, no matter how great I think it is, I'm not dumb enough to buy a HDDVD/BluRay right now. Everytime I watch a DVD on my TV, I curse the consumer electronics industry for not being able to come up with a sane solution to fit my needs.
As much as it is ignored by 95% percent of the population, the HD-DVD/BluRay war is a deep disappointment for the other 5%.
No offense, but Java would be a particularly bad choice for this application. The real work done here is "gluing" two things together at a system level - the Roomba program and the Wii program. Java is uniquely miserable at interacting on the system level. Where Java excels is object-oriented architectural design and huge enterprise-level programs.
So, if you want a scalable application server, use Java. If you need to tie two programs together, use a scripting language like Python or Perl. If you know you can limit yourself to a Unix shell environment, Perl may be best (and this is from a Python junkie!)
The question here is, of course, which one of these will be true:
1) Vista will unlock new potential markets for companies, allowing them to hire new programmers to add features to existing products or create new ones.
2) Vista will increase the barrier of entry for programs, meaning 100,000 new jobs will be created just to be able to support it, even in the absence of new features.
As someone who every so often has to see Windows when he visits home to fix things, I hope for the former. That would imply that Vista was done nicely, and I don't have to bend-over-backwards fixing my family's computers.
Meanwhile, I'll continue on with my Mac laptop, Linux servers, and continue to live without Windows. I don't think I'm missing much.
No, you got it all wrong:
In capitalist West, rich telcos own government.
In Soviet Russia, government owns telcos!
(no really, no joke)
Because the Wii has such a sexy name?
Beats the hell outta me.
You forgot a step: Drink 10 beers, *then* start playing the Wii. See if you can break something then.
Most (but not all, of course) of the incidents involving breaking stuff also involve plenty of alcohol
Futurama depiction
I'd have to agree with you, buddy.
The first PotC was such an awesome movie because Disney thought it would suck. They half-made it, and tossed it aside. Because they thought it would fail, they let the director and Johnny Depp do their thing - no market droids wanted to touch it.
Then, when it was all wrapped up and finished, they watched it and said "Damn, this is pretty good." But, before they could market the hell out of it and reinsert more special effects, the thing came out.
You could tell Disney's initial plan was to dump it because it was a "builder" - the thing made increasing amounts of money over the summer.
Goes to show how good movies can be if the market droids leave it alone.
I still think the EU (and Russia) has a long way before they "get over it" with respect to WWII, like the GP suggested.
I've visited mass graves in St. Petersburg. I've talked to people who lost family in the war. There's still a lot of emotion tied up into that war.
To too many (but not all, hopefully) US citizens, WWII was the "great adventure" that their grandfather or great-grandfather went on, and came back with a lot of stories about.
To many people in Europe/Russia, WWII was something that may have closely affected the town that they grew up in. Emotional wounds don't just get up and heal because we are in the Computer Age.
That's a disappointing article from Joel, he's usually more observant than that. I'd expect him to be able to make the distinction between "simplicity of an application" and "simplicity of an interface". You can have a feature-rich application which has an extremely simple interface.
My Mac laptop has a simple interface that both me and my wife enjoy. However, it is perfectly as functional as my linux desktop, who is much more complex.
An iPod's interface is simple; the device itself is complex. Same with gmail.
Both authors make the mistake of equating "ease of use" with "lack of features".
My god, if an admin can do that on a school's budget, why the hell are they working for the school?
It's still a fairly good trick to get Windows and Linux to coexist gracefully based on Linux servers. That's worth some money.
A good college shouldn't expect you to know HOW to program in C++. A Good College should teach how to program first and foremost, where the example language is C++.
I had friends in Georgia Tech who were decent Java programmers who did miserable in their introductory programming classes because the professor chose an extremely obscure language that no one knew beforehand. This way, he knew that no one came in who knew programming, but didn't know the concepts. By choosing a weird language, he could force concepts first, specific languages later. They hated it, got a poor grade, but came out better programmers.
On the same note, a mathematician does not differentiate between solutions of ax^2 + bx + c = 0 and x^2 + 5x + 1 = 0; knowing how to solve the quadratic equation is the important part, the second is just an example to make the theory easier.
If you want to learn how to approach the profession with a scientific outlook, which will enable you to get an entry-level programming job today and get a new one tomorrow when industry makes a major shift in languages or paradigms, go to the university. The math will make your programming better. The multivariate calc is possibly the least useful math class a CS major could take. More important is set theory, matrix theory, probability, and statistics.
Another way to evaluate degrees is this: were you taught how to think and program, or how to write Java? If Java becomes dead in 5 years, the person who learned how to program will shrug his shoulders and move on. The person who learned how to write Java will become unemployed. (Note: I'm just picking on Java; replace it with language X, and the statement holds true).
The Internet2 is the research network that universities use for high-bandwidth needs. Universities formed a consortium so there would be a fancy organization that NSF would fund network connections and that they could leverage their size to get cheaper bandwidth.
Buying a 10Gbps line from I2 that travels over Level3's network is about 10x cheaper than buying that bandwidth through L3 directly. Of course, it can be damned expensive to buy up the fiber from your university to one of the I2 hubs. Further, I2 doesn't like you to connect directly to their hubs - they prefer to connect through a "regional network". So, the Great Plains region, for example, set up a regional network called the GPN, and then the GPN buys the connection to Abilene (the name of the current I2 network).
There's some money available for research into networking technologies such as the connection-oriented software mentioned above, but most of the traffic is simple IPV4 (and some IPV6).
So, for some people, the I2 means that they can turn over 200 TB of disk cache every 2 weeks or so (which are the requirements for some sites associated with the Large Hadron Collider, like us). For others, it means that they can stream classes in HDTV across the United States.
For another good chunk of users, it means that they can trade movies/songs at gigabit speeds, if the p2p application can support that. Most can't. It still takes a bit of networking black-magic to get line-speed for gigabit ports, but not nearly as much as it used to.
Can I hear an "Amen" to that.
Our office bought the LVM-42W2 for video conferencing over the summer. Since then, 4 of us have bought the exact same model. It's got tons of inputs (all the various analog ones, 1 VGA, 2 DVI, 1 HDMI). It can do 1080p. It is cheap - finding it for $1500 is not hard, I think. I haven't run into any quirks.
The difference between the Westinghouse and the $3000 Samsung is that the Samsung has lots of nice filters on it, whereas the Westinghouse only has the standard brightness/contrast/etc. Three points:
1) Your 1080p/1080i source doesn't need any expensive upconverting filter technologies.
2) You'll want a nicer up-converting image for DVD sources. This can be remedied by buying a nice $100 DVD player which does the up-conversion, instead of having the TV do it.
3) Unless you have lots of nice TVs at home already, you won't be able to tell the difference between the Westinghouse and a $3000 set once you get it in your living room. The only way to see that the $3000 set has a marginally better picture is to put them next to each other.
So, the extra $1500 in cost goes away once you take the set home, and in the worst case can be remedied by buying a nice DVD player (cost: $100).
I friggin' love my TV and, at $1500, my wife even let my buy it.
You can drag patent cases on indefinitely. The last statistic I heard is that it cost $1 million to fight a patent case; I bet you can multiply that by many times if you're facing MS or IBM.
Still, in the end, if I was OO.o, I'd be flattered that some other company wanted to dedicate programmers to my project. However, I'd still be wary about including it.
Finally, the
No, if a major patent holder really wanted to, they could keep the Open Source community up to its ears in lawsuits indefinitely.
First of all, note that this is not a fork of the code. Novell is developing a plugin to read the OpenXML format, a Microsoft format.
.doc format, or that there aren't any patent violations in the Linux kernel or OO.o already.
Let me repeat, They are not forking OpenOffice.
Hell, the sourceforge project is called "odf-converter", not "Novell's evil plan for OO.o".
Further, the only way that I could read the press release from Novell in order to interpret it as "Novell is forking OpenOffice.org" is by the sentence which refers to the current OpenOffice.org product as "Novell's OpenOffice.org". That sounds more like a marketing intern not understanding how OOo and open source works out, not a secret decision on Novell's part.
Finally, I really hate the attitude that many of those contributing to Slashdot has taken toward Novell's current projects. It's fairly one-sided. They are not violating the law. They are not violating the GPL. They are not violating the spirit of the GPL.
The point of the GPL is that anyone can take your code, change it, and redistribute it, as long as they follow the rules. You can't make a distinction between people redistributing your software who you like and those who you don't like.
There's a lot of you who are sounding like Bush-style Republicans who want free speech for themselves, but not for those saying things they don't agree with. I bet a lot of you beating up on Novell today for taking advantage of the GPL are the same who beat up on Newt Gingrich the other day when he wanted to restrict free speech on the Internet. Hypocrites.
If you don't like Novell's contributions, don't accept them; if you think Novell is trying to get OpenXML into OO.o so MS can sue RedHat for patent infringement, think again. I doubt OpenXML is any more patent-ridden than the
In other words, Novell can't paint any bigger target on Linux's back than there already is. MS and IBM have so many ambiguous patents that they can sue any Linux user for the indefinite future.
Believe it or not, Novell may just be trying to differentiate its product so people would buy it over their competitor's product. You know, effectively compete in the business world. That sort of thing.
Groklaw used to be a place where I could get a detailed analysis of legal issues I didn't understand. Now, it seems to have disintegrated into blind zealotry. Maybe they were trying to be funny in the article, and I just didn't get the joke...
It may be that I'm just the crazy mathematician, but if you get published in a well-respected journal, your results should be statistically significant. If your results are statistically significant, by the way the mathematics works, you had to have gathered enough evidence to prove yourself false, but the data actually proved your hypothesis true.
So, unless the journals in the field are complete crap, there's enough evidence to falsify the results about the historical climate. However, where things get a bit fuzzy is when you try and apply historical data to predict the future. The real problem is that no one knows how exactly to compare the past with the future; does the earth do the same thing, or have humans altered the climate in some way that the earth is going to do something crazy over the next climate cycle?
That's the real tough nut to crack. You get some very loud people saying "It's part of the natural climate cycle, human CO2 is just noise in a larger trend" and some others saying "We're fundamentally altering the climate dynamics of the planet". Without a either a time machine or a second planet Earth to compare the trends to, there is no way we can falsify the theories which apply the results to the future.
This is the real big problem
No, not quite -
It's like you want to study the eating habits of 500 lb people. However, there's just too many skinny people to record. So, you ignore the eating habits of people under 400 lbs. Sure, it's possible that all the scales in the world are wrong and all people who are weighed at 100 lbs are actually 500 lbs, but that risk is considered acceptable.
The hardware level triggers and software reconstructors of CMS and ATLAS will cull out low energy events. There's an extremely small probability that something interesting in low energy events (1 in a trillion?); there's a better probability (1/10000 perhaps) that something interesting happened in high energy collisions. If you had to throw one out due to the amount of storage space you had available, which would you record?
The way that I heard it is that every single war of aggression since the dawn of the industrial age has been a loss for the aggressor.
I, for one, can't think of any major conflicts which break this rule. Iraq I, Iraq II, WWI, WWII, Vietnam, Afghanistan (Britain and Russia), Korea. I'm sure there are other conflicts I am forgetting; maybe some other country breaks this rule.
The third Afghani war (US in Afghanistan) looked like a sure victory for the aggressor until this year. Now it's looking a little shaky.
Simply put, we don't have any way to effectively fight a determined insurgency.
You haven't seen many papers in particle physics, have you?
Many times the author list is longer than the article itself.
An increase of only a couple of degrees of temperature can cause massive, and I mean massive, problems. They expect to loose a couple of magnets here and there, but the loss of the detector would be devastating. Not only does the magnet get destroyed, but there's enough energy in the beam (which would no longer be focused) to punch big holes in walls. This is why it is fairly deep underground.
To guard against this (at least in the sister project, CMS), they have 4 huge tanks (about the capacity of two semi-trucks, or more) of liquid helium sitting right above ground. If there's ever an emergency loss of power, these will flood the magnets with liquid helium to protect the magnets.
There's a reason that this project cost $8 billion and is billed as the most complex scientific experiment in the history of man.