As press releases tend to do, this leaves much to be desired for folks who are familiar with the discipline. As I read it, it seems to imply that the main driver is phrase-matching. What does it do with phrases it hasn't seen before? The problem is solved by throwing lots of data at it -- how much data is needed for a reasonable system? How well does it generalize to text outside the domains of the training data?
Incidentally, had my brother been a girl, he was in serious danger of being named Rosetta Stone.
The Mercury article mentioned opposition to cloning and I just had a strike of insight.
Sci-fi "designer babies" ieas tend to assume that parents (with the help of scientists) can determine exactly how their baby will turn out. They just select a default set of genes, add a few they like, and then send it to the manufacturer (er, womb).
Now, as any software engineer will tell you, no code worth anything works the first time. I make embarrasing mistakes on quick 3-line perl scripts, and anything with high complexity requires months (if not years) of coordinated planning, development, extensive testing, and lots and lots of debugging.
Surely constructing a designer person is as complex as, say, Microsoft Office (if not as complex as a Mars Rover). Office had a few bugs, and Microsoft releases patches from time to time. But people are different. Testing a person requires a lot more resource expenditure than running a shell script. You can't just code up a quick prototype and throw it away. And once the person is out in the wild, updates are nigh impossible. I sure don't know how to patch a 7-year-old.
That's not to say the gold standard isn't buggy. Sex has produced some pretty lousy people, in both hardware and software. But evolution is an ongoing develop/test/debug cycle on the scale of millions of years. I don't know many parents who want to wait that long.
I don't know enough about what Canada has to offer, so this is limited to the U.S.
When folks around here say they're "going backpacking" they usually mean they'll be hiking in the wildnerness with just what they can carry on their back. Such trips rarely include visits to bookstores, musea, and other geek centers. Such trips are best in mountainous areas -- I can't imagine backpacking in North Dakota, for instance -- and can be done on a pretty low daily budget (but make sure you invest in high quality boots, tents, etc.). Some folks have mentioned the Appalachian Trail, which spans from Vermont to Georgia. On the other side of the country are lots of swell backpacking areas from the Rocky Mountains west. The national parks in Utah and Arizona (Canyonlands, Staircase, Zion, Grand Canyon, etc.) are especially popular for such trips, though if you've spent much time in the outback you may be sick of such a climate (though the terrain here is more impressive). Almost any national park or national forest is a good backpacking experience, and entrance fees (especially if you get a year-long pass) are astonishingly cheap.
Unfortunately, you'll be arriving at the tail end of good backpacking season. Beginning in late September you can't trust in a lack of snow anywhere inland in the northern two thirds of the country, though places like southern Arizona are quite enjoyable. Unless you're staying until late next spring, you should hit any outdoor areas in the north first and work your way south.
Unfortunatey, U.S. public transportaion leaves much to be desired. There's nothing like a Eurail pass, and Amtrak stops mostly in larger cities, which is sad, because trains played such a large part in building America. Greyhound has excellent coverage and fairly reasonable rates, but if you're going to a lot of places, your pocketbook could take a big hit. Finally, hitchhiking across the country is probably no longer a viable plan, but it may be invaluable in a pinch. Hitchhikers are, generally, assumed to be dangerous until proven otherwise. On the plus side, most cities have a bus system decent enough for tourists to enjoy the town.
Unless you have access to a car, my advice is to pick a handfull of (relatively small) areas you want to visit and then figure out what all the great things to do there are.
Some geeky things in my neck of the woods (Boulder, Colorado): National Institute of Standards and Technology (home of the most accurate clock in this hemisphere) is right next to a branch of National Oceanographic and Atmospheric Association and a beautiful two-hour mountain hike away from the National Center for Atmospheric Research. They've all got free tours and such, though I haven't taken one since security got tightened after 9/11/01. About 10 miles south of town is the National Renewable Energy Laboratory and more beautiful mountain hike areas. 30 miles or so to the north west is Rocky Mountain National Park, which gets pretty cold in September and later. Denver, CO has Forney Museum of Tranportation and also the nation's only major airport built in the last 20 years, so it's full of neat engineering bits.
Your post sounds quite ambitious, and there's no way you can really experience all of what's neat in America in even a year, so find some of it and enjoy the hell out of that! Cheerio!
There's an organization working on stationary bike-powered computers with Internet access for rural areas without electricity. I see no reason the same concept can't be used to save power costs and keep in shape on this side of the pond.
As a class project, a friend and I built a music recognition database. You can read our paper.
The general approach is fairly straightforward. You extract a set of "features" (typically several Mel Frequency Cepstral Coefficients, or MFCCs) from each sample of the song, say 10ms. You then pick several (say, 16) arbitrary points and iteratively generate that many "average" feature vectors, along with their weights so that they all sum to a one vector. This data is turned into a Hidden Markov Model (HMM). To see what audio you have, you run it through each of the possible HMMs and see which produces the greatest likelihood.
This method is typically applied to speaker recognition, where a linear search through HMMs is reasonable. This obviously isn't the case when you know about hundreds of thousands of songs, so a large part of the challenge is narrowing the field of HMMs to check (which is one of the focuses in our paper). Relatable, who were working with Napster a long time ago, have clusters that can classify 1,000 songs per second; I'm pretty sure they use this technique.
This technique has several important features. First, it doesn't depend on any properties of files themselves. Checksums would be trivial to beat, looking at a file's length could be circumvented by inserting silence, etc. Since this creates an average of sample data, a song would need to be changed quite a bit to fail to match. (The system is robust to, for instance, changes in bitrate, slowing the music down, and rearranging bits of the song or putting it in reverse.) We didn't have enough "derivative" music to test how it handles sampled music vs. the original -- it depends how much is changed.
Finally, this sort of system is useful for much more than song identification. You can build a model for an artist or genre and determine how to classify the song. One of my focuses in the paper is unsupervised genre classification -- my tests indicated some fairly reasonable groupings. This technique could be used for music recommendation -- "You like Dropkick Murphys? Well, they sound like Flogging Molly, so you might want to check them out."
Sren Kierkegaard was engaged to Regina, but wrote that it was vetoed by God, even though he still loved her. He went on to be an influential existentialist with his writings on keeping faith, giving up in dispair, and so on. Perhaps God knew SK could only amount to anything by not being married.
Of course, Yaweh isn't married either, and look at all he's created!
The reason this is 3.5 rather than version 4 is the size of the changes. 3 to 3.5 is not anything like 2 to 3. 3.5 is a balancing act, fixing the broken and strengthening the weak. Just like MacOS 7.5 was much like 7.0, just better, you don't need to relearn very much. If you know 3, you'll only need the 3.5 books for reference. If you don't know 3, 3.5 is a good place to start.
There must be scores of legitimate reasons for airplanes to go over city centers.
* Most major sports stadia are near downtown, and planes towing advertisements often circle during the game. * Even more annoying, stadium events often have flyovers. (Hmmm... maybe keeping those away is a good thing.) * Helicopters can safely land downtown. This is a Good Thing(tm) when someone must be airlifted to a hospital. Of course, I don't think this proposal applies to helicopters, so what will terrorists use? * Many airports are fairly near city centers. Soft walls would impede take off and landing. * Others?
Now that I think about it, you could kill a LOT of people by crashing a plane into an airport terminal. We'd better throw up soft walls, anti-aircraft missles, and other anti-plane measures at all the nation's airports!
If you have to try say 3 times to send a message [over a 5 day period or so] you're ability to mass send 100million emails is really squashed.
It's hardly squashed, it just takes a little longer.
Legitimate people first time sending won't really mind the few day wait and most MTAs will try for upto a month.
I don't think I've seen a bounce that tried for more than five days. And while instantaneous may be a poor expectation, I often legitimately expect within-an-hour times. The most obvious case of this is automatic responses. A less obvious case is when you know the recipient will be available.
For instance, a web visitor submits a form. I happen to be at my computer, and ask for more info. The visitor is still available, so responds quickly. I now have to dink around online for an hour to get a piece of mail that shouldn't take more than a few minutes.
Another case: I phone a friend and ask him to look at the source code I'm having trouble with. He has to wait an hour to receive it, then I have to wait an hour to see his response. Perhaps this could be better solved by instant messaging, but I don't have an IM client at work.
Finally, the timing of carbon copied messages could get wonky. If several people are conversing via email, each sender/recipient pair creates a new timeout. If some of the participants have previously corresponded, their messages will arrive quickly, and they may respond immediately. Others wait for more than an hour for the original message, but may receive some replies before the original. This becomes vastly confusing to follow and significantly inconvenient.
I like the Greylisting approach, and these issues could be circumvented by delivering all mail to a user when the MTA thinks she's reading her email (when she's POPed recently or is logged on to a system with low idle time, etc.).
One other note -- 4 hours seems too small of an initial window, especially if there's some sort of attack.
World population will hit 7 Billion by 2011. So all those extra tech guys will be working for Total Information Awareness keeping tabs on all these new terrorists.
As a TA, I graded a homework assignment which, for whatever reason, left in the debugging cout's "PISS" "SHIT" and "DAMN" which were used as loop markers. The program was annoyed on a regular basis.
Tau Beta Pi, the Engineering Honor Society, has struggled with this question as well. What disciplines should be considered "engineering?" At last year's convention, we approved (pending ratification due on Tuesday) following the guidelines set forth in ABET's criteria for Engineering programs. ABET (Accreditation Board for Engineering and Technology) accredits four types of programs -- engineering, engineering technology, computer science, and applied math. 's move was to limit membership to those programs which fall in the domain of the first.
ABET recently added accreditation guidelines for Software Engineering, but as yet, no such programs have yet been accredited. (That is, their pop-up menu for "search by program" doesn't list Software Engineering.) ABET also accredits Computer Engineering, which is usually about hardware, though this includes some programs with "Computer Science" in their names (such as Berkeley's EECS program).
The requirements for Software Engineering are: 1. Curriculum The curriculum must provide both breadth and depth across the range of engineering and computer science topics implied by the title and objectives of the program. The program must demonstrate that graduates have: the ability to analyze, design, verify, validate, implement, apply, and maintain software systems; the ability to appropriately apply discrete mathematics, probability and statistics, and relevant topics in computer science and supporting disciplines to complex software systems; and the ability to work in one or more significant application domains. 2. Faculty The program shall demonstrate that those faculty teaching core software engineering material have practical software engineering experience.
This is the sense of "engineering" which is concerned with the profession of engineering -- who can legally say they're an engineer, which is what the Texas legislature is talking about. But the word "engineer" goes beyond accreditation and licensing. Are the only teachers in the world people who have a license to teach? Are you out of line by calling Hippocrates a physician and doctor because he didn't have an M.D.? Of course not.
The second sense of "engineer" is someone who integrates principles of math and science with real-world constraints in the design, creation, or maintenance of some in-context solution. Thus, studying the physical properties of electricity and building a circuit isn't engineering, it's science. But building such a circuit under performance, economic, and other constraints is engineering.
Algorithm development is thus mathematical engineering. When Edsgar Dijkstra first published a shortest path algorithm and provided a complexity bound (the first version was O(n^3) I've heard), mathematicians said "So what? Just list all possible paths and pick the shortest one." Which is, of course, impractical for any large system.
Software Engineers work with performance constraints, economic constraints, time constraints, constraints imposed by existing systems, security constraints, constraints of readability and maintainability, etc. The code you write for a homework assignment which is graded purely on its functional properties isn't (likely) an engineered program. But programs you download (especially large ones) have been engineered (or they suck).
Engineering doesn't have to be technical, though. City planners can be said to be Social Engineers. Some people claim the title of Financial Engineer. The folks at Kodak are Image Engineers. Gutenberg was one of the world's first Publishing Engineer.
"Engineering," like "guardian" has a legal definition and a common sense definition. Engineering, in English, is an approach and a mindset.
Once or twice I've been glad that there's a Windows lab in the Engineering Center. Sometimes you just need to run software that won't run with the (BSD) devil.
When I majored in presidentin' in college, I had no idea how much time I'd spend meetin' with boring people who want this that and the other thing. Nobody told me I'd have to read laws and get briefed all about trade with Zambiabwezie.
I thought I just had to sign papers and wave to crowds and have my secret service agents foil assassination attempts.
It seems that TiVo recommendations are volumes more humorous than the shows themselves.
Most machine learning algorithms take quite a bit of data for anything resembling acceptable accuracy. And if TiVo recommends based on everythong from genre to year to actors, you have to watch a LOT of TV before TiVo knows you as well as the clerk at your local independent video store. It very quickly becomes a large game of 6 Degrees of Pregnant Gay Korean War Porn Stars.
One reasons your friends give you better book recommendations than Amazon is that they know something about your life outside your literary habits. Did you buy a book about southwestern cooking because you grew up in New Mexico or because you have a bumper bean crop this year? Did you buy a Bible because your last one wore out, or because you wanted to make Marilyn Manson fan art?
They say American kids spend more time watching TV than they do in school. For perfectly tuned TiVo recommendations, you'd probably have to give up sleep, work, and slashdot.
Not only does the webmaster need to pay for bandwidth hosting and other incurred costs, so does the user.
I wonder if it's feasable to collect statistics on the percentages of bandwidth and time used in viewing advertisements. It wouldn't be too hard to do a localized study for a limited set of advertising sites (doubleclick and co.). Can I send the webmaster a bill? (They say 36% of email now is spam. If 36% of the web bytes you were served were ads, would you be inclined to keep surfing?)
To all those Slashdot readers intending to use Anti-Theft's software (hehe), take note of the following: * To successfully require popup-ad viewing, you'll need to make your site worthless without JavaScript enabled. * Anyone who turns off filters in order to view your site isn't going to click through pop-ups. * Anyone who can't (or doesn't want to) turn off filtering won't view your adds anyway... or your content. But will incur bandwidth costs of several loads (one (or more) for the attempted page, two or three for the tests it runs, one for the redirect). * I don't imagine folks viewing your site with a PDA, cellphone, etc. will be inclined in your favor.
Part of the whole theory of advertising is that you throw it out there and hope a rather small fraction of the viewers will purchase your product. Scores of millions of people see the newest SUV during the Super Bowl. How many buy it? Not many.
(Incidentally, you don't have to quit your browser to view the page. You probably just have to reload.)
A Mandrake box and a playlist aren't enough to start a good radio station. You're a college student, so use that to your advantage -- get your university involved. THE most important piece of running a radio station is people power, and that's one of a university's great resources. Start a radio club, if there isn't one already. Promote it -- it shouldn't be hard to have folks itchin' to be DJs, but what you really need is folks to do behind-the-scenes stuff. Calling people, filling out forms, promoting, doing tech, etc. Running a radio station is a lot of work, but also a lot of fun. The best radio stations are run in large part by volunteers, 'cause then it's the love of the station and the music that drives them, not maximizing profits and targeting the right markets. They also tend to be commercial free, so you get to listen to more radio in the same one-hour period. Community and college radio is the open source of broadcasting.
Next to people power you need money, space, and equipment. This is where your university administration comes in. Figure out how to get the university to pay for it -- they may have money lying around for stuff like that or you might be able to get students to vote to raise student fees for a radio station. Be prepared to ask listeners for donations, too. Your university can probably give you space to broadcast from and have offices in.
As folks have pointed out, you can get an educational liscence and deal with a lot less crap. Running it through the university means it'll also have a chance to survive once you and your friends graduate. (Make sure to get community members involved too, 'cause students are notoriously busy, flaky, and absent for significant portions of time.)
It's possible that your university turns down your proposal. KGNU, a well-known community station, was started after the University of Colorado turned down a proposal for a radio station, but most schools will support such a project. CU does now, in the form of KVCU/Radio 1190.
You must be prepared to work really hard without tangible results for a while. You must be able to present clear and reasoned proposals. You must inspire lots of help. You have to REALLY love what you're doing. It's like starting a business, but maybe an order of magnitude harder.
If we haven't scared you off yet, the first thing to do is talk to folks from other college and community stations. Get their tips and tricks. Figure out what they didn't forsee and make sure you have a plan for it.
Slashdot Mirror
As press releases tend to do, this leaves much to be desired for folks who are familiar with the discipline. As I read it, it seems to imply that the main driver is phrase-matching. What does it do with phrases it hasn't seen before? The problem is solved by throwing lots of data at it -- how much data is needed for a reasonable system? How well does it generalize to text outside the domains of the training data?
Incidentally, had my brother been a girl, he was in serious danger of being named Rosetta Stone.
-- Trevor Stone, aka Flwyd
The Mercury article mentioned opposition to cloning and I just had a strike of insight.
Sci-fi "designer babies" ieas tend to assume that parents (with the help of scientists) can determine exactly how their baby will turn out. They just select a default set of genes, add a few they like, and then send it to the manufacturer (er, womb).
Now, as any software engineer will tell you, no code worth anything works the first time. I make embarrasing mistakes on quick 3-line perl scripts, and anything with high complexity requires months (if not years) of coordinated planning, development, extensive testing, and lots and lots of debugging.
Surely constructing a designer person is as complex as, say, Microsoft Office (if not as complex as a Mars Rover). Office had a few bugs, and Microsoft releases patches from time to time. But people are different. Testing a person requires a lot more resource expenditure than running a shell script. You can't just code up a quick prototype and throw it away. And once the person is out in the wild, updates are nigh impossible. I sure don't know how to patch a 7-year-old.
That's not to say the gold standard isn't buggy. Sex has produced some pretty lousy people, in both hardware and software. But evolution is an ongoing develop/test/debug cycle on the scale of millions of years. I don't know many parents who want to wait that long.
I don't know enough about what Canada has to offer, so this is limited to the U.S.
When folks around here say they're "going backpacking" they usually mean they'll be hiking in the wildnerness with just what they can carry on their back. Such trips rarely include visits to bookstores, musea, and other geek centers. Such trips are best in mountainous areas -- I can't imagine backpacking in North Dakota, for instance -- and can be done on a pretty low daily budget (but make sure you invest in high quality boots, tents, etc.). Some folks have mentioned the Appalachian Trail, which spans from Vermont to Georgia. On the other side of the country are lots of swell backpacking areas from the Rocky Mountains west. The national parks in Utah and Arizona (Canyonlands, Staircase, Zion, Grand Canyon, etc.) are especially popular for such trips, though if you've spent much time in the outback you may be sick of such a climate (though the terrain here is more impressive). Almost any national park or national forest is a good backpacking experience, and entrance fees (especially if you get a year-long pass) are astonishingly cheap.
Unfortunately, you'll be arriving at the tail end of good backpacking season. Beginning in late September you can't trust in a lack of snow anywhere inland in the northern two thirds of the country, though places like southern Arizona are quite enjoyable. Unless you're staying until late next spring, you should hit any outdoor areas in the north first and work your way south.
Unfortunatey, U.S. public transportaion leaves much to be desired. There's nothing like a Eurail pass, and Amtrak stops mostly in larger cities, which is sad, because trains played such a large part in building America. Greyhound has excellent coverage and fairly reasonable rates, but if you're going to a lot of places, your pocketbook could take a big hit. Finally, hitchhiking across the country is probably no longer a viable plan, but it may be invaluable in a pinch. Hitchhikers are, generally, assumed to be dangerous until proven otherwise. On the plus side, most cities have a bus system decent enough for tourists to enjoy the town.
Unless you have access to a car, my advice is to pick a handfull of (relatively small) areas you want to visit and then figure out what all the great things to do there are.
Some geeky things in my neck of the woods (Boulder, Colorado): National Institute of Standards and Technology (home of the most accurate clock in this hemisphere) is right next to a branch of National Oceanographic and Atmospheric Association and a beautiful two-hour mountain hike away from the National Center for Atmospheric Research. They've all got free tours and such, though I haven't taken one since security got tightened after 9/11/01. About 10 miles south of town is the National Renewable Energy Laboratory and more beautiful mountain hike areas. 30 miles or so to the north west is Rocky Mountain National Park, which gets pretty cold in September and later. Denver, CO has Forney Museum of Tranportation and also the nation's only major airport built in the last 20 years, so it's full of neat engineering bits.
Your post sounds quite ambitious, and there's no way you can really experience all of what's neat in America in even a year, so find some of it and enjoy the hell out of that! Cheerio!
I've been transplanting my tongue into other people's mouths for years.
There's an organization working on stationary bike-powered computers with Internet access for rural areas without electricity. I see no reason the same concept can't be used to save power costs and keep in shape on this side of the pond.
How will they come up with enough funny jokes to present at each login prompt? I don't think /usr/games/fortune is up to snuff.
As a class project, a friend and I built a music recognition database. You can read our paper.
The general approach is fairly straightforward. You extract a set of "features" (typically several Mel Frequency Cepstral Coefficients, or MFCCs) from each sample of the song, say 10ms. You then pick several (say, 16) arbitrary points and iteratively generate that many "average" feature vectors, along with their weights so that they all sum to a one vector. This data is turned into a Hidden Markov Model (HMM). To see what audio you have, you run it through each of the possible HMMs and see which produces the greatest likelihood.
This method is typically applied to speaker recognition, where a linear search through HMMs is reasonable. This obviously isn't the case when you know about hundreds of thousands of songs, so a large part of the challenge is narrowing the field of HMMs to check (which is one of the focuses in our paper). Relatable, who were working with Napster a long time ago, have clusters that can classify 1,000 songs per second; I'm pretty sure they use this technique.
This technique has several important features. First, it doesn't depend on any properties of files themselves. Checksums would be trivial to beat, looking at a file's length could be circumvented by inserting silence, etc. Since this creates an average of sample data, a song would need to be changed quite a bit to fail to match. (The system is robust to, for instance, changes in bitrate, slowing the music down, and rearranging bits of the song or putting it in reverse.) We didn't have enough "derivative" music to test how it handles sampled music vs. the original -- it depends how much is changed.
Finally, this sort of system is useful for much more than song identification. You can build a model for an artist or genre and determine how to classify the song. One of my focuses in the paper is unsupervised genre classification -- my tests indicated some fairly reasonable groupings. This technique could be used for music recommendation -- "You like Dropkick Murphys? Well, they sound like Flogging Molly, so you might want to check them out."
Of course, Yaweh isn't married either, and look at all he's created!
The reason this is 3.5 rather than version 4 is the size of the changes. 3 to 3.5 is not anything like 2 to 3. 3.5 is a balancing act, fixing the broken and strengthening the weak. Just like MacOS 7.5 was much like 7.0, just better, you don't need to relearn very much. If you know 3, you'll only need the 3.5 books for reference. If you don't know 3, 3.5 is a good place to start.
Thieves sued banks for forcing robbery to occur by not giving away free money.
I guess this means I'm going to be giving my English instructor a bottle of champagne instead of a research paper this Fall.
It's just as well. He'll be happier, and you'll get a better grade.
There must be scores of legitimate reasons for airplanes to go over city centers.
* Most major sports stadia are near downtown, and planes towing advertisements often circle during the game.
* Even more annoying, stadium events often have flyovers. (Hmmm... maybe keeping those away is a good thing.)
* Helicopters can safely land downtown. This is a Good Thing(tm) when someone must be airlifted to a hospital. Of course, I don't think this proposal applies to helicopters, so what will terrorists use?
* Many airports are fairly near city centers. Soft walls would impede take off and landing.
* Others?
Now that I think about it, you could kill a LOT of people by crashing a plane into an airport terminal. We'd better throw up soft walls, anti-aircraft missles, and other anti-plane measures at all the nation's airports!
Two hours was the striking distance for the roving bombers in Dr. Stragelove, or How I Learned to Stop Worrying and Love the Bomb.
Running a laptop which has a liquid as poth input and output sounds like a bad idea. I have enough trouble trying to keep my soda away.
If you have to try say 3 times to send a message [over a 5 day period or so] you're ability to mass send 100million emails is really squashed.
It's hardly squashed, it just takes a little longer.
Legitimate people first time sending won't really mind the few day wait and most MTAs will try for upto a month.
I don't think I've seen a bounce that tried for more than five days. And while instantaneous may be a poor expectation, I often legitimately expect within-an-hour times. The most obvious case of this is automatic responses. A less obvious case is when you know the recipient will be available.
For instance, a web visitor submits a form. I happen to be at my computer, and ask for more info. The visitor is still available, so responds quickly. I now have to dink around online for an hour to get a piece of mail that shouldn't take more than a few minutes.
Another case: I phone a friend and ask him to look at the source code I'm having trouble with. He has to wait an hour to receive it, then I have to wait an hour to see his response. Perhaps this could be better solved by instant messaging, but I don't have an IM client at work.
Finally, the timing of carbon copied messages could get wonky. If several people are conversing via email, each sender/recipient pair creates a new timeout. If some of the participants have previously corresponded, their messages will arrive quickly, and they may respond immediately. Others wait for more than an hour for the original message, but may receive some replies before the original. This becomes vastly confusing to follow and significantly inconvenient.
I like the Greylisting approach, and these issues could be circumvented by delivering all mail to a user when the MTA thinks she's reading her email (when she's POPed recently or is logged on to a system with low idle time, etc.).
One other note -- 4 hours seems too small of an initial window, especially if there's some sort of attack.
World population will hit 7 Billion by 2011. So all those extra tech guys will be working for Total Information Awareness keeping tabs on all these new terrorists.
Where do I sign up?
As a TA, I graded a homework assignment which, for whatever reason, left in the debugging cout's "PISS" "SHIT" and "DAMN" which were used as loop markers. The program was annoyed on a regular basis.
Tau Beta Pi, the Engineering Honor Society, has struggled with this question as well. What disciplines should be considered "engineering?" At last year's convention, we approved (pending ratification due on Tuesday) following the guidelines set forth in ABET's criteria for Engineering programs. ABET (Accreditation Board for Engineering and Technology) accredits four types of programs -- engineering, engineering technology, computer science, and applied math. 's move was to limit membership to those programs which fall in the domain of the first.
ABET recently added accreditation guidelines for Software Engineering, but as yet, no such programs have yet been accredited. (That is, their pop-up menu for "search by program" doesn't list Software Engineering.) ABET also accredits Computer Engineering, which is usually about hardware, though this includes some programs with "Computer Science" in their names (such as Berkeley's EECS program).
The requirements for Software Engineering are:
1. Curriculum
The curriculum must provide both breadth and depth across the range of engineering and computer science topics implied by the title and objectives of the program. The program must demonstrate that graduates have: the ability to analyze, design, verify, validate, implement, apply, and maintain software systems; the ability to appropriately apply discrete mathematics, probability and statistics, and relevant topics in computer science and supporting disciplines to complex software systems; and the ability to work in one or more significant application domains.
2. Faculty
The program shall demonstrate that those faculty teaching core software engineering material have practical software engineering experience.
This is the sense of "engineering" which is concerned with the profession of engineering -- who can legally say they're an engineer, which is what the Texas legislature is talking about. But the word "engineer" goes beyond accreditation and licensing. Are the only teachers in the world people who have a license to teach? Are you out of line by calling Hippocrates a physician and doctor because he didn't have an M.D.? Of course not.
The second sense of "engineer" is someone who integrates principles of math and science with real-world constraints in the design, creation, or maintenance of some in-context solution. Thus, studying the physical properties of electricity and building a circuit isn't engineering, it's science. But building such a circuit under performance, economic, and other constraints is engineering.
Algorithm development is thus mathematical engineering. When Edsgar Dijkstra first published a shortest path algorithm and provided a complexity bound (the first version was O(n^3) I've heard), mathematicians said "So what? Just list all possible paths and pick the shortest one." Which is, of course, impractical for any large system.
Software Engineers work with performance constraints, economic constraints, time constraints, constraints imposed by existing systems, security constraints, constraints of readability and maintainability, etc. The code you write for a homework assignment which is graded purely on its functional properties isn't (likely) an engineered program. But programs you download (especially large ones) have been engineered (or they suck).
Engineering doesn't have to be technical, though. City planners can be said to be Social Engineers. Some people claim the title of Financial Engineer. The folks at Kodak are Image Engineers. Gutenberg was one of the world's first Publishing Engineer.
"Engineering," like "guardian" has a legal definition and a common sense definition. Engineering, in English, is an approach and a mindset.
...that will let people copy between 30 and 100 hours of music onto a single disc.
I was getting sick of changing CDs on road trips from Miami to Anchorage.
I am embarassed to admit this, but...
Once or twice I've been glad that there's a Windows lab in the Engineering Center. Sometimes you just need to run software that won't run with the (BSD) devil.
I'll go sit in the corner and feel shame, now.
From your perspective at the ISP, what's the most problematic type of spam, and how do you deal with it? How would you like it dealt with?
When I majored in presidentin' in college, I had no idea how much time I'd spend meetin' with boring people who want this that and the other thing. Nobody told me I'd have to read laws and get briefed all about trade with Zambiabwezie.
I thought I just had to sign papers and wave to crowds and have my secret service agents foil assassination attempts.
Presidentin' is hard!
-- George W.
It seems that TiVo recommendations are volumes more humorous than the shows themselves.
Most machine learning algorithms take quite a bit of data for anything resembling acceptable accuracy. And if TiVo recommends based on everythong from genre to year to actors, you have to watch a LOT of TV before TiVo knows you as well as the clerk at your local independent video store. It very quickly becomes a large game of 6 Degrees of Pregnant Gay Korean War Porn Stars.
One reasons your friends give you better book recommendations than Amazon is that they know something about your life outside your literary habits. Did you buy a book about southwestern cooking because you grew up in New Mexico or because you have a bumper bean crop this year? Did you buy a Bible because your last one wore out, or because you wanted to make Marilyn Manson fan art?
They say American kids spend more time watching TV than they do in school. For perfectly tuned TiVo recommendations, you'd probably have to give up sleep, work, and slashdot.
Not only does the webmaster need to pay for bandwidth hosting and other incurred costs, so does the user.
I wonder if it's feasable to collect statistics on the percentages of bandwidth and time used in viewing advertisements. It wouldn't be too hard to do a localized study for a limited set of advertising sites (doubleclick and co.). Can I send the webmaster a bill?
(They say 36% of email now is spam. If 36% of the web bytes you were served were ads, would you be inclined to keep surfing?)
To all those Slashdot readers intending to use Anti-Theft's software (hehe), take note of the following:
* To successfully require popup-ad viewing, you'll need to make your site worthless without JavaScript enabled.
* Anyone who turns off filters in order to view your site isn't going to click through pop-ups.
* Anyone who can't (or doesn't want to) turn off filtering won't view your adds anyway... or your content. But will incur bandwidth costs of several loads (one (or more) for the attempted page, two or three for the tests it runs, one for the redirect).
* I don't imagine folks viewing your site with a PDA, cellphone, etc. will be inclined in your favor.
Part of the whole theory of advertising is that you throw it out there and hope a rather small fraction of the viewers will purchase your product. Scores of millions of people see the newest SUV during the Super Bowl. How many buy it? Not many.
(Incidentally, you don't have to quit your browser to view the page. You probably just have to reload.)
A Mandrake box and a playlist aren't enough to start a good radio station. You're a college student, so use that to your advantage -- get your university involved. THE most important piece of running a radio station is people power, and that's one of a university's great resources. Start a radio club, if there isn't one already. Promote it -- it shouldn't be hard to have folks itchin' to be DJs, but what you really need is folks to do behind-the-scenes stuff. Calling people, filling out forms, promoting, doing tech, etc. Running a radio station is a lot of work, but also a lot of fun. The best radio stations are run in large part by volunteers, 'cause then it's the love of the station and the music that drives them, not maximizing profits and targeting the right markets. They also tend to be commercial free, so you get to listen to more radio in the same one-hour period. Community and college radio is the open source of broadcasting. Next to people power you need money, space, and equipment. This is where your university administration comes in. Figure out how to get the university to pay for it -- they may have money lying around for stuff like that or you might be able to get students to vote to raise student fees for a radio station. Be prepared to ask listeners for donations, too. Your university can probably give you space to broadcast from and have offices in. As folks have pointed out, you can get an educational liscence and deal with a lot less crap. Running it through the university means it'll also have a chance to survive once you and your friends graduate. (Make sure to get community members involved too, 'cause students are notoriously busy, flaky, and absent for significant portions of time.) It's possible that your university turns down your proposal. KGNU, a well-known community station, was started after the University of Colorado turned down a proposal for a radio station, but most schools will support such a project. CU does now, in the form of KVCU/Radio 1190. You must be prepared to work really hard without tangible results for a while. You must be able to present clear and reasoned proposals. You must inspire lots of help. You have to REALLY love what you're doing. It's like starting a business, but maybe an order of magnitude harder. If we haven't scared you off yet, the first thing to do is talk to folks from other college and community stations. Get their tips and tricks. Figure out what they didn't forsee and make sure you have a plan for it.