Truly, I think the only place this may make sense is in a clustered environment where you don't want to be changing a monitor cable around between systems. However, even then it doesn't make sense since you should get a KVM controller.
As with 95% of all machine hacks, cute, but ultimately pointless.
However, you can also use systems of equatiosn where c=1 (unitless), leading to many entertaining conversions. There is also one where G=1, which leads to mass in units of meters. That's what I meant by not looking too closely.
The sun has been causing global warming? Now who would have ever expected a giant ball of uncontrolled nuclear explosions to change at all and have any effect on the warming of our planet.
So it is directly proportional. However, I didn't look closely at the units that they are using here, but thta shouldn't matter to the solution at hand.
When a boss needs to cut things to show he's ding things to keep costs under control, he invariably heads for 'the little things' first. Like that espresso machine. Or the supply of bottled water. Or Mountain Dew. In many cases these are what employees will consider made the place 'livable', and when the perceived quality of life drops, morale soon goes out the door as well. Especially when all the old guys tell all the new guys "Back in the day we had blah-di-blah blah"
Now, IANAEMTG (I Am Not An English Major, Thank God), but we seem to be prety lax on this whole concept of Haiku.
From the little I understand, part of the art of Haiku is to have the first two lines be completely unrelated, and the third tie them together.
Bane of Chernobyl, First End of Life for our subs, Curse you, O Xenon!
Also, I've often wondered how translatable 'syllable' to the original Japanese word are. I have absolutely zero knowledge of the language, but I'm certain an original Haiku would not translate with the same rhyme scheme we seem to imply.
Except that TV and Radio stations would want this-- they have to pay a bundle to 'licence' their spectrum.
Since the ultimate size of their audience is one of power of broadcasting and not of spectrum usage, I'm sure they would absolutely adore changes in the FCC that made the frequencies cheap and plentiful.
In fact, if the restructuring happened, and lots of smaller stations appeared, they may even like it more since they could then say, "We have to compete with these smaller stations, and they are allowed to say ShitPissFuckCuntCocksuckerMotherFuckerTits." So laws regarding content restriction could be lifted, and we all know that this would enhance overall radio as we know it. (A little sarcasm at the end thrown in for good measure).
And just to point out, while many of the 'hands-on' variety of positions on the boat/ship are enlisted, there's always an officer standing EOOW (Engineering Officer of the Watch), which is the Engineer's representative for the plant much like the OOD is the Captain's representative. So, this is the position that college students are actively recruited for.
And yes, the Sea-shore rotation does absolutely suck for these types. Even worse, from what I only hear, thankfully, is that when the boat comes in from being out at sea for 3 months, everyone get to go home right away except for the nukes-- it's a multi-hour shutdown procedure, and someone has to stay and watch it when it's off as well.
Actually from what I hear. That paper you signed going in has a clause that even if you're out. You can be reactivated, and be brought back. So unless one wants to be effectively owned for the rest of their lives, no! (For those who don't see this, it's an AC that I'm answering).
I'm actually in this group as well, and honestly, if you're in the country, and you're male, you're already owned. Remeber that Selective Service card you were supposed to fill out when you turned 18? The liklihood of that being used is actually higher than the liklihood of you getting called back after you resign your commission.
Now, I'm actually in the Active Reserves, which gives me around $400/drill weekend and I get to go on all sorts of nifty trips as well. The only thing I have to fear is if we go to war, but what, I ask you, are the odds of that happening these days?
Honestly, they're good-- but for the real reason that people should get PhD's. Some try to get a good SysAdmin job with a PhD, and find that experience wins out over the title-- as it should in that case.
A PhD is for learning how to do research. You may end up at a smaller university. You may end up at a research lab. You may end up with a post doc. The jobs are there-- but they are to do endless research.
Now, as a grad student, this is my impression at the moment, and I'm quite fine with it-- I want to do research. Saldy, I know many students in similar positions to mine stating, "But I don't want to do this my whole life..." These people possibly should not have sought a PhD.
Also Off the Main topic, but to fill in some stuff, I thought I'd mention that I took up the Navy on it's offer.
It's pretty good while you're in college-- right now they pay right around $40K/year for your last two years in school--not bad, since you have absolutely nothing ROTCish or Navyish to do for those two years (I even interned at a national lab while I was getting paid by the Navy-- hooray for double dipping.:)
Anyway, the job I signed on for was instructor, which meant that I taught onshore, never seeing a sub but as a tourist, for 4 years. And then I was out. If you want a military career- this is not the way to go. If you hate paperwork, this is not the way to go. If you despise bueracracies, 'the man', uniforms, power trips from idiots, or senseless rules, this is not for you.
However, it is a job, it gives you in-state tuition for whatever school you're in, it delays having to choose a real career for 4 years, and they do give a reasonable paycheck. (The instructor option is only open for technical majors, however-- otherwise you can go sub, not see the sky for 3 months at a time, go crazy, but get about a $12K signing bonus.)
However, you have somewhat proven my point. I was stating that high level languages had not reached the level of "put this data in the most efficient manner possible" problem. Sure, Perl is great. But notice that you, the programmer, still had to know which type you should be using.
Don't get me wrong, my only statement is that high level languages now, and quite possibly will always, require a deep understanding of the associations of the data they are maintaining.
The main thing that C offers over Perl is speed(always), and a more efficient use of memory, (generally).
And how do you know that's the most efficient way to store the data? Is it a B-Tree? A Fib. Heap? A stack? A sorted list? There are very large and detailed books on algorithm and data structure selection.
Now, one or two might do for most cases, but if you really care about efficiency of your sort and storing(and I mean _really_ care), then you'd have to look more closely at the algorithm you're choosing, and not just say, "Oh, Perl, do this."
The primes.java test used in the article is a bit tricky because they used "long" for all the calculation instead of "int", and GCC doesn't produce particularly efficient code for the 64-bit mod operation. After changing "long" to "int", gcj beats both IBM and Sun Linux JDKs. With "long" we are slightly faster than Sun, but a lot slower than IBM - so obviously their JIT must have better optimized "long" operations. From here.
So, it seems that the JIT compiler knew the underlying architecture and was able to use a better represenation of integers in it.
Specifically, my problem is the following:
Every Java program is really two (or more) programs running at the same time: something to compile it, and the program itself. Once it's run long enough, you may be able to assume that the program is fully compiled and runs just like native code-- except for all the bounds checking and possible garbage collection.
Now, it could be that this added amount of security is well worth it. However, it's going to take some really creative run time optimizations to gain back these losses.
You are greatly underestimating what people dwell on. Sure, to anyone in this field, they see little differnece between 95 and XP. For that matter, show someone in this field KDE 3.1 and they'll instantly know how to go about using it-- after they complain about the one vs. two click differnece (unless you set it up to act like MS as well).
However, when someone who was using 95 switches to XP, watch 'em freeze up. It's amazing to see the panic in a new users eys as things appear and disappear without asking. Or to watch him search for the somewhat more hidden windows explorer program that so many were used to.
And, I will will also state, the reason that Microsoft is more 'intuitive' is because of advertising. There is absolutely no reason to expect people to know to double click. Or to drag. Or what individual icons mean. Or any number of other things that get spread more by 'word of mouth' than by any real true 'intuition'.
And I will also state that you could switch most people over, just tell them it was a cheaper version of Windows, and they would retrain themselves.
(Thank God for typeahead find in Mozilla, I was just about to post this...)
Anyway, I think the somewhat big news here is that non miltary agencies will be moving to the SIPRNET. And switching over more "routine" communications to this systems has to be a good thing for a variety of reasons.
And for those to lazy to google, here's a link . SIPRNET is designed to encrypt and send traffic, and they use their own wres and relays. (Although I can't swear that they don't use some of the commericial wires as well.)
However, interpretted code (Java or Perl) will always run slower than natively compiled code. Even assuming JIT compiling.
And as far as algorithms go, I was unaware that high level languages had gotten to the point of "sort and store this information in the best possible way for this application" being an acceptable form of instruction.
Now, this being said, I can certainly see the day where the norm becomes Most code in Java/Perl + some of the more numerical or the body of tight loops being in C + some instructions hand optimized. (If this isn't he common practice already).
However, you'll always be sacrificing speed for readability. Even if you think it's 'good enough', you should still hope that noone wants that extra order of magnitude in performance you can get with a truly compiled language.
Just look at what happens everytime someone goes about writing an OS in Java. Someone invariably points out that the same thing would run 300% faster if C had been used instead.
And, I'm sorry to say, there will always be a need for performance.
We're currently running some experiments using IDE drives, and so we packed a standard tower case with 9 60GB hard drives. It wasn't easy, and took 4 IDE cards, and we had to search for a system that had enough PCI slots, but we did it.
The problem we've run into is a matter of power-- the power supply really doesn't like being chained out using splitters to all those drives.
With a 533MHz front side bus, and assuming a 128 bit wide structure (unrealistic), you're just at 8GB/sec. And theoretically you'd want to have something else beside disks on this system-- maybe memory and video? Maybe a network?
Slashdot Mirror
Truly, I think the only place this may make sense is in a clustered environment where you don't want to be changing a monitor cable around between systems. However, even then it doesn't make sense since you should get a KVM controller.
As with 95% of all machine hacks, cute, but ultimately pointless.
Notice how the Pythagorean Theorem...
Like this?.
Took about 30 seconds with google, and that's because I misspelled Pythagorean. Good thread, however.
However, you can also use systems of equatiosn where c=1 (unitless), leading to many entertaining conversions. There is also one where G=1, which leads to mass in units of meters. That's what I meant by not looking too closely.
The sun has been causing global warming? Now who would have ever expected a giant ball of uncontrolled nuclear explosions to change at all and have any effect on the warming of our planet.
I'm still believing it's the cow farts.
Actually, a quick googling found this:
r0=2GM/c^2 (Eqn 10.1.5)
So it is directly proportional. However, I didn't look closely at the units that they are using here, but thta shouldn't matter to the solution at hand.
When a boss needs to cut things to show he's ding things to keep costs under control, he invariably heads for 'the little things' first. Like that espresso machine. Or the supply of bottled water. Or Mountain Dew. In many cases these are what employees will consider made the place 'livable', and when the perceived quality of life drops, morale soon goes out the door as well. Especially when all the old guys tell all the new guys "Back in the day we had blah-di-blah blah"
*cough*SteveForbesRossPerot*cough*
Now, IANAEMTG (I Am Not An English Major, Thank God), but we seem to be prety lax on this whole concept of Haiku.
From the little I understand, part of the art of Haiku is to have the first two lines be completely unrelated, and the third tie them together.
Bane of Chernobyl,
First End of Life for our subs,
Curse you, O Xenon!
Also, I've often wondered how translatable 'syllable' to the original Japanese word are. I have absolutely zero knowledge of the language, but I'm certain an original Haiku would not translate with the same rhyme scheme we seem to imply.
Except that TV and Radio stations would want this-- they have to pay a bundle to 'licence' their spectrum.
Since the ultimate size of their audience is one of power of broadcasting and not of spectrum usage, I'm sure they would absolutely adore changes in the FCC that made the frequencies cheap and plentiful.
In fact, if the restructuring happened, and lots of smaller stations appeared, they may even like it more since they could then say, "We have to compete with these smaller stations, and they are allowed to say ShitPissFuckCuntCocksuckerMotherFuckerTits." So laws regarding content restriction could be lifted, and we all know that this would enhance overall radio as we know it. (A little sarcasm at the end thrown in for good measure).
And just to point out, while many of the 'hands-on' variety of positions on the boat/ship are enlisted, there's always an officer standing EOOW (Engineering Officer of the Watch), which is the Engineer's representative for the plant much like the OOD is the Captain's representative. So, this is the position that college students are actively recruited for.
And yes, the Sea-shore rotation does absolutely suck for these types. Even worse, from what I only hear, thankfully, is that when the boat comes in from being out at sea for 3 months, everyone get to go home right away except for the nukes-- it's a multi-hour shutdown procedure, and someone has to stay and watch it when it's off as well.
Actually from what I hear. That paper you signed going in has a clause that even if you're out. You can be reactivated, and be brought back. So unless one wants to be effectively owned for the rest of their lives, no! (For those who don't see this, it's an AC that I'm answering).
I'm actually in this group as well, and honestly, if you're in the country, and you're male, you're already owned. Remeber that Selective Service card you were supposed to fill out when you turned 18? The liklihood of that being used is actually higher than the liklihood of you getting called back after you resign your commission.
Now, I'm actually in the Active Reserves, which gives me around $400/drill weekend and I get to go on all sorts of nifty trips as well. The only thing I have to fear is if we go to war, but what, I ask you, are the odds of that happening these days?
Honestly, they're good-- but for the real reason that people should get PhD's. Some try to get a good SysAdmin job with a PhD, and find that experience wins out over the title-- as it should in that case.
A PhD is for learning how to do research. You may end up at a smaller university. You may end up at a research lab. You may end up with a post doc. The jobs are there-- but they are to do endless research.
Now, as a grad student, this is my impression at the moment, and I'm quite fine with it-- I want to do research. Saldy, I know many students in similar positions to mine stating, "But I don't want to do this my whole life..." These people possibly should not have sought a PhD.
Just to share, I'm getting paid to get a PhD in CS. Every school I know of does this for every technical major-- nobody pays for a PhD in CS.
Sure, you have to live like a dog for multiple years, making between 13-17K a year, but hey, if you can live that cheap, there are no problems.
Also Off the Main topic, but to fill in some stuff, I thought I'd mention that I took up the Navy on it's offer.
:)
It's pretty good while you're in college-- right now they pay right around $40K/year for your last two years in school--not bad, since you have absolutely nothing ROTCish or Navyish to do for those two years (I even interned at a national lab while I was getting paid by the Navy-- hooray for double dipping.
Anyway, the job I signed on for was instructor, which meant that I taught onshore, never seeing a sub but as a tourist, for 4 years. And then I was out. If you want a military career- this is not the way to go. If you hate paperwork, this is not the way to go. If you despise bueracracies, 'the man', uniforms, power trips from idiots, or senseless rules, this is not for you.
However, it is a job, it gives you in-state tuition for whatever school you're in, it delays having to choose a real career for 4 years, and they do give a reasonable paycheck. (The instructor option is only open for technical majors, however-- otherwise you can go sub, not see the sky for 3 months at a time, go crazy, but get about a $12K signing bonus.)
When you go fishing, why do you have to take 2 mormons?
Because if you only take one, he'll drink all your beer.
However, you have somewhat proven my point. I was stating that high level languages had not reached the level of "put this data in the most efficient manner possible" problem. Sure, Perl is great. But notice that you, the programmer, still had to know which type you should be using.
Don't get me wrong, my only statement is that high level languages now, and quite possibly will always, require a deep understanding of the associations of the data they are maintaining.
The main thing that C offers over Perl is speed(always), and a more efficient use of memory, (generally).
And how do you know that's the most efficient way to store the data? Is it a B-Tree? A Fib. Heap? A stack? A sorted list? There are very large and detailed books on algorithm and data structure selection.
Now, one or two might do for most cases, but if you really care about efficiency of your sort and storing(and I mean _really_ care), then you'd have to look more closely at the algorithm you're choosing, and not just say, "Oh, Perl, do this."
So, it seems that the JIT compiler knew the underlying architecture and was able to use a better represenation of integers in it.
Specifically, my problem is the following:
Every Java program is really two (or more) programs running at the same time: something to compile it, and the program itself. Once it's run long enough, you may be able to assume that the program is fully compiled and runs just like native code-- except for all the bounds checking and possible garbage collection.
Now, it could be that this added amount of security is well worth it. However, it's going to take some really creative run time optimizations to gain back these losses.
You are greatly underestimating what people dwell on. Sure, to anyone in this field, they see little differnece between 95 and XP. For that matter, show someone in this field KDE 3.1 and they'll instantly know how to go about using it-- after they complain about the one vs. two click differnece (unless you set it up to act like MS as well).
However, when someone who was using 95 switches to XP, watch 'em freeze up. It's amazing to see the panic in a new users eys as things appear and disappear without asking. Or to watch him search for the somewhat more hidden windows explorer program that so many were used to.
And, I will will also state, the reason that Microsoft is more 'intuitive' is because of advertising. There is absolutely no reason to expect people to know to double click. Or to drag. Or what individual icons mean. Or any number of other things that get spread more by 'word of mouth' than by any real true 'intuition'.
And I will also state that you could switch most people over, just tell them it was a cheaper version of Windows, and they would retrain themselves.
(Thank God for typeahead find in Mozilla, I was just about to post this...)
Anyway, I think the somewhat big news here is that non miltary agencies will be moving to the SIPRNET. And switching over more "routine" communications to this systems has to be a good thing for a variety of reasons.
And for those to lazy to google, here's a link . SIPRNET is designed to encrypt and send traffic, and they use their own wres and relays. (Although I can't swear that they don't use some of the commericial wires as well.)
However, interpretted code (Java or Perl) will always run slower than natively compiled code. Even assuming JIT compiling.
And as far as algorithms go, I was unaware that high level languages had gotten to the point of "sort and store this information in the best possible way for this application" being an acceptable form of instruction.
Now, this being said, I can certainly see the day where the norm becomes Most code in Java/Perl + some of the more numerical or the body of tight loops being in C + some instructions hand optimized. (If this isn't he common practice already).
However, you'll always be sacrificing speed for readability. Even if you think it's 'good enough', you should still hope that noone wants that extra order of magnitude in performance you can get with a truly compiled language.
Just look at what happens everytime someone goes about writing an OS in Java. Someone invariably points out that the same thing would run 300% faster if C had been used instead.
And, I'm sorry to say, there will always be a need for performance.
Always.
Well, then, I sit corrected.
Just to share.
We're currently running some experiments using IDE drives, and so we packed a standard tower case with 9 60GB hard drives. It wasn't easy, and took 4 IDE cards, and we had to search for a system that had enough PCI slots, but we did it.
The problem we've run into is a matter of power-- the power supply really doesn't like being chained out using splitters to all those drives.
Just an FYI.
With a 533MHz front side bus, and assuming a 128 bit wide structure (unrealistic), you're just at 8GB/sec. And theoretically you'd want to have something else beside disks on this system-- maybe memory and video? Maybe a network?