I don't think we were discussing what books you should pull code from, were we? That's not the subject on the door, anyway.
You are inferring that because I said that there are versions of this book in C,C++, and Fortran that I meant that you should use the code from the book. I did not say this. What I said was that it's a good book to learn from, as you have agreed. There were a lot of times looking through the book that I remember thinking "why did they do it that way? It's certainly convoluted."
You're wrong about the waste of memory, though. Go look again. Whenever they do new/malloc operations, they assign the pointer to an address that is one less than the address given so that the allocated array runs from 1 to n (accessing 0 would be illegal). While this is not convention, it is reasonable. This part adds fairly negligable expense, assuming you can remember that this is what they're doing. Optimizing compilers don't have any problem with this kind of range stuff, either, since the result is still linear, so this will most likely result in no difference at all.
I've used the same trick myself (making bigger images that have legal coordinates such as (-3,-10) ) with image mosiacs so I didn't have to translate the coordinate system.
Still, a lot of the algorithms themselves (not the code, usually) are useful exactly the way they are without variation even in production environments (the ubiquitous one that springs to mind is the Gaussian method with pivoting). You just have to keep in mind that the book is thirty years old, and you're going to be doing things the way that they've always been done, which is not necessarily the best way. You can write something using the Numerical Recipe, profile, and then get a different method than the recipe if that's what's sucking your cycles.
I think we have different ideas about what "stomps the snot" means. Adium would stomp the snot out of all other messenger clients, except that it's OSX-only.
If you've got someone who has already read Adam's extremely small collection of works (including HHGTG and the Dirk Gently books, his "we picked up all the clippings of stuff he wrote and put it into a book," and "Starship Titanic" based on his video game), then they might also like Terry Pratchett's books.
He tells a ripping good yarn, almost all parodies, usually insightful, and always funny. His most famous series is the Discworld books, of which there are 30 or so.
Reading through these books my favorite author slowly switched from Douglas to Pratchett. The humor style is almost the same, but Pratchett's a little more optimistic about the nature of life, which I find refreshing.
I wasn't aware that there was anyone who actually felt that three dimensional discrete calculus and C programming are both less difficult to understand than a VCR.
Go you!
But if you're going there, why not go ahead and get the master work on algorithms, Numerical Recipes, which is now available in C,C++, and Fortran versions. This, just like your suggestion, is hardcore programmers brainfood, not fluff.
For softer stuff, may I suggest O'Reilly books about scripting languages? It doesn't really matter which one. Pick one. Ruby, perl, python, etc. They give a good introduction to modern programming libraries, while not getting the newbie bogged down in having to write algorithms that are more complicated than they can deal with.
So...how about OSS efforts to develop Trillian clones?
I've already got the names picked out: Zaphod: OSX ('cause it's the hippest OS) Arthur: Zaurus (small brain, but still very valuable) Slartibartfast: Solaris (older than the rest) Marvin: EarthSim (here I am, brain the size of a planet, and they make me run an IM client. I'm not getting you down, am I? Because I wouldn't want to get you down. Life! Don't talk to me about life). Last two I can't think of any really good reason for the names, though: Eddie: BSD Ford: Linux
Right. You could be a secretary for the actual translator. Then your primary function would be to get coffee, collate, and if your boss isn't good, to convert Engrish into English.
Ah yes...because there's NO WAY that a group of people could be affected by the same growth-stunting mutagen and become the way they were. That is clearly impossible, and it is clearly another species.
You know, there are documented cases of humans (even groups of them - mutants who share the same parents or have the same disease) whose skeletal structure is as different from the norm as the norm is from all the other species in the Homo genus. Is it possible that we've got it wrong, and that we've only really found one species of man ever, or is there something besides skeletal remains on which to determine if there have been others in our genus?
He's neglected to mention some things from the first figure.
First of all, it's a 1 in 455 chance of being wiped out by asteroids, volcanic activity, comets, vampires, dark elves, zombies, or McDonald's, but he seemed to convenienly leave off the end of the list.
Secondly, he forgot to mention that this takes into account the fact that all humans who have not broken the code of the Greblor (roughly 96.3%) will be delivered by the benevolent lizard Godzilla back unto our home planet - a place of safety and prosperity in another dimension. Only the evil, self-destructive humans will remain.
Further, it is predicted that 97.1% of those who stay will be delivered in the second coming of Godzilla after having repented of their evil ways.
So as you can see, most of us have nothing to worry about. They neglected to mention the other parts of the report, which actually explain why the numbers are obviously true.
Level 3: Master of the entire energy capacity of a single galaxy Level 4: masters of the universe Level 5: All power put in one place and given to one man for justice - He-man.
Level 2 is only permanent if you don't piss off any of the higher leveled species so much that they wipe you out.
Most of those changes weren't as rapid as the one caused by the internet. At least, I don't think they were. So if we're measuring profundity by speed, the internet is probably winning.
I haven't really lived through their development, have you?
Also, I've taken a plane four times. I drive twice a day for fifteen minutes, and I talk on the phone twice a week. I use the can...well lets not get too specific, and just say that it's less than the amount of time I spend online. So if we're measuring profundity by time spent, then the internet wins for me, and there are probably others, too (well...I actually use lights even more than that). At the very least, I think we can cross "jetliner" off the list if this is the metric.
The last measure of profundity that I'd consider is in what I'd do without it. Now, I have no idea what most people's lives are actually like, so I can't say what's most important to them. Think about it yourself: Without indoor plumbing: I'd go outside to do that business. Not really a major change. Without engines: I live in a place where I can walk everywhere. Very little would change. Without jets: Those four trips I took? I wouldn't have. Without telephones: I'd just go over to people's houses when I wanted to communicate. Most people I call are within thirty minutes driving from me, or two hours walking. Or I'd use the internet.:) If there was never an internet: I'd have a differnt line of work, a new form of relaxing, I'd know a lot less than I do, and I wouldn't know anything about computers. I'd probably be an electrician. Without electric light: I'd probably be dead from a fire to my house, or blind from trying to read in the dark. Also, without light, there'd be no monitors, which basically means no home computers as we know them.
So...I'd say that the internet has a much more profound effect on my life than any of those things you said, except the lightbulb by any of the obvious measures of profundity.
If they had outlets in planes, then the answer would be yes. I'd bring a MiniATX machine and an LCD monitor. They're less fragile than laptops, and easier to repair.
He pretty much gave them free reign. ANY OSS at all!
Have you seen CPAN? Half of that code is something someone hacked up in a day! And what about all those sourceforge projects that have one developer and less than 10000 lines?
Meanwhile, almost every piece of code that this class is looking at is stuff that's already had a once over - heck, probably even been looked over thousands of times. No wonder they couldn't find any bugs. They were looking in the houses, not the motels.
I agree. Just the other day, I heard about a company went under when they fired their star programmer for leaving a disk labeled "main project-2004" on his desk (which was used in a presentation to the venture capitalists), instead of marking it "Sesame street - hardcore pr0n." He just didn't have a marker.
To make a long story short, the company didn't get the investment they needed due to a bad presentation, and they went under.
I don't think Fujitsu is really known for it's marking...shouldn't they have merged with another company?
The problem that the original poster understood (and that you're not getting) is that robotics is hard compared to programming. There's simply too much knowledge required to make something that actually does something.
As he was saying, you get a lot of bang for your buck with programming. You can make something that'll impress people within a day's work.
There is an exception, though - I think that if you stick to lego mindstorms or other canned robotic solutions (so that in general what you're talking about is programming robots rather than building them), then it become more possible - as long as you stay away from things like "navigating a maze" or "picking up an object using vision" which are the kind of things you get in college and professional level robotic competitions.
Fighting robots are WAY too expensive for high schools to actually be able to do (and I'm speaking as a friend of someone who made one that competed in battlebots), while simultaneously teaching very little about robotics, and a lot about remote control cars.
I'm not sure you're right about the calculators...I was known as "the calculator magi" in my school. People would come to me to get games, programs, and other things for the calculator, and I think I impressed a few. Sure, I didn't become popular just because I was good with my calculator, but I did meet a lot of people that way that I probably wouldn't have met otherwise. Oh, and I even got a scholarship for one of the programs I wrote for it. There are bound to be a few like me.
Ah...what metric are you using to determine the fractal dimension?
If you're talking about box count, this is not as accurate a metric as you are claiming. If you're talking about exact Capacity Dimension, then you do have a point: it can be used. It is also very difficult to determine precisely in almost all cases. It is difficult to prove that something actually has such a dimension other than by empirically showing that the box count is consistent enough to imply a fractal nature.
Box count is a fairly useful metric, actually. However, a better one that seems to show the same information while having little to do with fractals is edge density. It's also faster to compute.
A PDA with this kind of drive in it could be used to store 120 or so movies (well...mine only gets 320x240 resolution, and I'm assuming good compression like DIVX).
They already have video units like this, but for some reason they think that if your PDA does this then it doesn't have to be able to do anything else.
It's been developing for over thirty years - over a hundred, depending on how you count it (mathematical description of Cantor dust? Mandlebrot's ideas? What?). There are lots of "applications." We can at least say that a lot of discoveries have been made, but that none of them have been very applicable to anything.
This development speed is a lot slower than almost any other field that's only thirty years old. Now I'm willing to admit that development efforts could have been slow because no one's been clever enough to make use of it - after all, number theory didn't really take off until cryptography did, and that depended on the power of computers.
However, unlike number theory, Fractal math has many, many admirers - people like myself, at least, as well as brilliant mathematicians. Its fun because it's an art, too, and so there are a lot who have made it a hobby. Considering the amount of man-hours put into fractals, I would expect more output if they were genuinely very useful.
I think, rather, that very few things actually conform to fractal patterns, which makes their uses much more limited.
They have these things called "math books" filled with problems. The advantage there is that you can work with pencil and paper and actually struggle through problems.
Need help? A better place - one with actual information about how things work, rather than just how to solve certain kinds of problems - is Mathworld. This one will stick with you, too.
I still use it, and I'm getting my Masters in Engineering.
"Hey, those fractals look pretty cool, I think I'll sign up for this advanced fractals course." Me, circa 1999.
"Holy flying fluke, Batman, where did these equations come from?!?" Same era, but it was still interesting. I didn't give up, and I learned a lot about them.
One thing I learned is that knowledge of fractals is almost useless. Fractal compression doesn't compete very well with Wavelet compression (used now), and there's virtually no other area to which they have been applied successfully.
The only exception to this is the plasma fractal, which is used in landscape and cloud generation. It makes a decent PRNG, but stream ciphers are faster.
So...I kind of wish I hadn't seen those shiny colors. They distracted me from learning stuff that I would be able to make use of later.
It is a decent Matlab clone now. The missing piece? Libraries. Matlab has lots and lots of libraries that do tons of stuff.
Well, Octave does too thanks to Octave-forge. There's even a symbolic engine (as this is probably the most popular thing used). There are now a few areas that Octave rises above Matlab in performance - albiet very few.
The biggest lacking area now is Simulink. Matlab could be used to run hardware through a simulation engine. That's not going to happen with Octave until someone needs to do it.
But for school apps, and for algorithms in general, Octave is more than good enough.
Actually, for the past few years, I've been a TA for the Freshman Engineering course at Purdue University. The majority of this is course is about learning Matlab.
I can attest that everything we have ever had the students learn will work in both Matlab and Octave+Octaveforge.
As to price, that is not what they charge educational institutions. It's more like $100 per license. Also, most specialized engineering applications are actually quite expensive; this is nothing new (Autocad, Synopsis, and Cadence are some of the others I've come in contact with at school).
Law #4: When you post late at night after studying for many hours at the very end of the semester when you've fried your brain, don't expect it to be perfect.
Okay, so I meant the former. I'm surpised that post ended up coherent at all. I don't know what I was thinking posting in that state of mind.
Slashdot Mirror
I don't think we were discussing what books you should pull code from, were we? That's not the subject on the door, anyway.
You are inferring that because I said that there are versions of this book in C,C++, and Fortran that I meant that you should use the code from the book. I did not say this. What I said was that it's a good book to learn from, as you have agreed.
There were a lot of times looking through the book that I remember thinking "why did they do it that way? It's certainly convoluted."
You're wrong about the waste of memory, though. Go look again. Whenever they do new/malloc operations, they assign the pointer to an address that is one less than the address given so that the allocated array runs from 1 to n (accessing 0 would be illegal). While this is not convention, it is reasonable. This part adds fairly negligable expense, assuming you can remember that this is what they're doing. Optimizing compilers don't have any problem with this kind of range stuff, either, since the result is still linear, so this will most likely result in no difference at all.
I've used the same trick myself (making bigger images that have legal coordinates such as (-3,-10) ) with image mosiacs so I didn't have to translate the coordinate system.
Still, a lot of the algorithms themselves (not the code, usually) are useful exactly the way they are without variation even in production environments (the ubiquitous one that springs to mind is the Gaussian method with pivoting). You just have to keep in mind that the book is thirty years old, and you're going to be doing things the way that they've always been done, which is not necessarily the best way. You can write something using the Numerical Recipe, profile, and then get a different method than the recipe if that's what's sucking your cycles.
I think we have different ideas about what "stomps the snot" means. Adium would stomp the snot out of all other messenger clients, except that it's OSX-only.
Same could be said of Trillian, actually.
If you've got someone who has already read Adam's extremely small collection of works (including HHGTG and the Dirk Gently books, his "we picked up all the clippings of stuff he wrote and put it into a book," and "Starship Titanic" based on his video game), then they might also like Terry Pratchett's books.
He tells a ripping good yarn, almost all parodies, usually insightful, and always funny. His most famous series is the Discworld books, of which there are 30 or so.
Reading through these books my favorite author slowly switched from Douglas to Pratchett. The humor style is almost the same, but Pratchett's a little more optimistic about the nature of life, which I find refreshing.
I wasn't aware that there was anyone who actually felt that three dimensional discrete calculus and C programming are both less difficult to understand than a VCR.
Go you!
But if you're going there, why not go ahead and get the master work on algorithms, Numerical Recipes, which is now available in C,C++, and Fortran versions. This, just like your suggestion, is hardcore programmers brainfood, not fluff.
For softer stuff, may I suggest O'Reilly books about scripting languages? It doesn't really matter which one. Pick one. Ruby, perl, python, etc. They give a good introduction to modern programming libraries, while not getting the newbie bogged down in having to write algorithms that are more complicated than they can deal with.
So...how about OSS efforts to develop Trillian clones?
I've already got the names picked out:
Zaphod: OSX ('cause it's the hippest OS)
Arthur: Zaurus (small brain, but still very valuable)
Slartibartfast: Solaris (older than the rest)
Marvin: EarthSim (here I am, brain the size of a planet, and they make me run an IM client. I'm not getting you down, am I? Because I wouldn't want to get you down. Life! Don't talk to me about life).
Last two I can't think of any really good reason for the names, though:
Eddie: BSD
Ford: Linux
Right. You could be a secretary for the actual translator. Then your primary function would be to get coffee, collate, and if your boss isn't good, to convert Engrish into English.
Ah yes...because there's NO WAY that a group of people could be affected by the same growth-stunting mutagen and become the way they were. That is clearly impossible, and it is clearly another species.
You know, there are documented cases of humans (even groups of them - mutants who share the same parents or have the same disease) whose skeletal structure is as different from the norm as the norm is from all the other species in the Homo genus. Is it possible that we've got it wrong, and that we've only really found one species of man ever, or is there something besides skeletal remains on which to determine if there have been others in our genus?
He's neglected to mention some things from the first figure.
First of all, it's a 1 in 455 chance of being wiped out by asteroids, volcanic activity, comets, vampires, dark elves, zombies, or McDonald's, but he seemed to convenienly leave off the end of the list.
Secondly, he forgot to mention that this takes into account the fact that all humans who have not broken the code of the Greblor (roughly 96.3%) will be delivered by the benevolent lizard Godzilla back unto our home planet - a place of safety and prosperity in another dimension. Only the evil, self-destructive humans will remain.
Further, it is predicted that 97.1% of those who stay will be delivered in the second coming of Godzilla after having repented of their evil ways.
So as you can see, most of us have nothing to worry about. They neglected to mention the other parts of the report, which actually explain why the numbers are obviously true.
Level 3: Master of the entire energy capacity of a single galaxy
Level 4: masters of the universe
Level 5: All power put in one place and given to one man for justice - He-man.
Level 2 is only permanent if you don't piss off any of the higher leveled species so much that they wipe you out.
Most of those changes weren't as rapid as the one caused by the internet. At least, I don't think they were. So if we're measuring profundity by speed, the internet is probably winning.
:)
I haven't really lived through their development, have you?
Also, I've taken a plane four times. I drive twice a day for fifteen minutes, and I talk on the phone twice a week. I use the can...well lets not get too specific, and just say that it's less than the amount of time I spend online. So if we're measuring profundity by time spent, then the internet wins for me, and there are probably others, too (well...I actually use lights even more than that). At the very least, I think we can cross "jetliner" off the list if this is the metric.
The last measure of profundity that I'd consider is in what I'd do without it. Now, I have no idea what most people's lives are actually like, so I can't say what's most important to them. Think about it yourself:
Without indoor plumbing: I'd go outside to do that business. Not really a major change.
Without engines: I live in a place where I can walk everywhere. Very little would change.
Without jets: Those four trips I took? I wouldn't have.
Without telephones: I'd just go over to people's houses when I wanted to communicate. Most people I call are within thirty minutes driving from me, or two hours walking. Or I'd use the internet.
If there was never an internet: I'd have a differnt line of work, a new form of relaxing, I'd know a lot less than I do, and I wouldn't know anything about computers. I'd probably be an electrician.
Without electric light: I'd probably be dead from a fire to my house, or blind from trying to read in the dark. Also, without light, there'd be no monitors, which basically means no home computers as we know them.
So...I'd say that the internet has a much more profound effect on my life than any of those things you said, except the lightbulb by any of the obvious measures of profundity.
One of the big problems there is that WEP is incredibly insecure, even though it uses an encryption standard.
They need something else.
If they had outlets in planes, then the answer would be yes. I'd bring a MiniATX machine and an LCD monitor. They're less fragile than laptops, and easier to repair.
Not unless you deploy it. Make sure it does something.
Put it on CPAN. How about:
Crypt::Insecure
Drop-in replacement for Crypt::Random. Demonstrates common vulnerabilities in OSS with ten intentional security holes.
He pretty much gave them free reign. ANY OSS at all!
Have you seen CPAN? Half of that code is something someone hacked up in a day! And what about all those sourceforge projects that have one developer and less than 10000 lines?
Meanwhile, almost every piece of code that this class is looking at is stuff that's already had a once over - heck, probably even been looked over thousands of times. No wonder they couldn't find any bugs. They were looking in the houses, not the motels.
I agree. Just the other day, I heard about a company went under when they fired their star programmer for leaving a disk labeled "main project-2004" on his desk (which was used in a presentation to the venture capitalists), instead of marking it "Sesame street - hardcore pr0n." He just didn't have a marker.
To make a long story short, the company didn't get the investment they needed due to a bad presentation, and they went under.
I don't think Fujitsu is really known for it's marking...shouldn't they have merged with another company?
The problem that the original poster understood (and that you're not getting) is that robotics is hard compared to programming. There's simply too much knowledge required to make something that actually does something.
As he was saying, you get a lot of bang for your buck with programming. You can make something that'll impress people within a day's work.
There is an exception, though - I think that if you stick to lego mindstorms or other canned robotic solutions (so that in general what you're talking about is programming robots rather than building them), then it become more possible - as long as you stay away from things like "navigating a maze" or "picking up an object using vision" which are the kind of things you get in college and professional level robotic competitions.
Fighting robots are WAY too expensive for high schools to actually be able to do (and I'm speaking as a friend of someone who made one that competed in battlebots), while simultaneously teaching very little about robotics, and a lot about remote control cars.
I'm not sure you're right about the calculators...I was known as "the calculator magi" in my school. People would come to me to get games, programs, and other things for the calculator, and I think I impressed a few. Sure, I didn't become popular just because I was good with my calculator, but I did meet a lot of people that way that I probably wouldn't have met otherwise. Oh, and I even got a scholarship for one of the programs I wrote for it. There are bound to be a few like me.
Ah...what metric are you using to determine the fractal dimension?
If you're talking about box count, this is not as accurate a metric as you are claiming. If you're talking about exact Capacity Dimension, then you do have a point: it can be used. It is also very difficult to determine precisely in almost all cases. It is difficult to prove that something actually has such a dimension other than by empirically showing that the box count is consistent enough to imply a fractal nature.
Box count is a fairly useful metric, actually. However, a better one that seems to show the same information while having little to do with fractals is edge density. It's also faster to compute.
A PDA with this kind of drive in it could be used to store 120 or so movies (well...mine only gets 320x240 resolution, and I'm assuming good compression like DIVX).
They already have video units like this, but for some reason they think that if your PDA does this then it doesn't have to be able to do anything else.
Which may or may not have an effect on battery life. After all, the new drives could consume more power.
Maybe the other way around?
Jungle Book came out in 1967.
developing field
It's been developing for over thirty years - over a hundred, depending on how you count it (mathematical description of Cantor dust? Mandlebrot's ideas? What?). There are lots of "applications." We can at least say that a lot of discoveries have been made, but that none of them have been very applicable to anything.
This development speed is a lot slower than almost any other field that's only thirty years old. Now I'm willing to admit that development efforts could have been slow because no one's been clever enough to make use of it - after all, number theory didn't really take off until cryptography did, and that depended on the power of computers.
However, unlike number theory, Fractal math has many, many admirers - people like myself, at least, as well as brilliant mathematicians. Its fun because it's an art, too, and so there are a lot who have made it a hobby. Considering the amount of man-hours put into fractals, I would expect more output if they were genuinely very useful.
I think, rather, that very few things actually conform to fractal patterns, which makes their uses much more limited.
No! Bad monkey!
They have these things called "math books" filled with problems. The advantage there is that you can work with pencil and paper and actually struggle through problems.
Need help? A better place - one with actual information about how things work, rather than just how to solve certain kinds of problems - is Mathworld. This one will stick with you, too.
I still use it, and I'm getting my Masters in Engineering.
"Hey, those fractals look pretty cool, I think I'll sign up for this advanced fractals course."
Me, circa 1999.
"Holy flying fluke, Batman, where did these equations come from?!?"
Same era, but it was still interesting. I didn't give up, and I learned a lot about them.
One thing I learned is that knowledge of fractals is almost useless. Fractal compression doesn't compete very well with Wavelet compression (used now), and there's virtually no other area to which they have been applied successfully.
The only exception to this is the plasma fractal, which is used in landscape and cloud generation. It makes a decent PRNG, but stream ciphers are faster.
So...I kind of wish I hadn't seen those shiny colors. They distracted me from learning stuff that I would be able to make use of later.
It is a decent Matlab clone now. The missing piece? Libraries. Matlab has lots and lots of libraries that do tons of stuff.
Well, Octave does too thanks to Octave-forge. There's even a symbolic engine (as this is probably the most popular thing used). There are now a few areas that Octave rises above Matlab in performance - albiet very few.
The biggest lacking area now is Simulink. Matlab could be used to run hardware through a simulation engine. That's not going to happen with Octave until someone needs to do it.
But for school apps, and for algorithms in general, Octave is more than good enough.
Actually, for the past few years, I've been a TA for the Freshman Engineering course at Purdue University. The majority of this is course is about learning Matlab.
I can attest that everything we have ever had the students learn will work in both Matlab and Octave+Octaveforge.
As to price, that is not what they charge educational institutions. It's more like $100 per license. Also, most specialized engineering applications are actually quite expensive; this is nothing new (Autocad, Synopsis, and Cadence are some of the others I've come in contact with at school).
Law #4: When you post late at night after studying for many hours at the very end of the semester when you've fried your brain, don't expect it to be perfect.
Okay, so I meant the former. I'm surpised that post ended up coherent at all. I don't know what I was thinking posting in that state of mind.