People who believe that should try the Implicit Association Test for race. It measures your unconscious biases; you can see how well they correspond to your conscious ones.
Saying it's only a class issue is sweeping some things under the rug. People use race to guess class, and what social class people are in is affected by race, especially historically but also in the present. One of the ways that is mediated is through the unconscious bias that things like the IAT reveals.
I resent that. I didn't have anything handed to me on a silver platter and I'm as white as it gets.
I'm sure you mean this sincerely, but I promise you that you've had a good dose of white male privilege, just like me. I used to say the same thing you did, but a variety of experiences made me realize that I was like a fish who never noticed the water. Here are a couple:
After high school, I ended up living in South America for a year. As an American overseas, initial reactions to me were bimodal: either I was presumed a patrón, and therefore deserving of excess respect for my race, gender, and wealth (because all Americans are, from there perspective, wealthy) or I was dirt on the floor because I represented, in their minds, the oppressive Yanqui colonialist government. When I came back both were attenuated, but both reactions to my presumed status still existed, especially in cities like Chicago with a fairly mixed ethnic balance.
More recently, I went, with otherwise no change, from being a guy with a ponytail and a goatee to a guy with a shaved head and a goatee. The change in reaction was dramatic: people were scared of me, and would cross the street to get away from me. A black friend of mine remarked, "Ah, now you know what it's like." And he's right: although he's better looking, better dressed, more visibly well off, nicer, and comes from a better family, there's no question that strangers treated me, on average, better, especially once I grew some hair back.
I wish there were a modern version of Black Like Me. It wouldn't be as dramatic, but it would help people to see that racism still is with us in a thousand ways.
Does it really matter who (in the sense of ethnic diversity) writes the OSS code or makes a important innovation?
Yes, for two reasons. The first is societal. Remember the first time you went to some geeky gathering? I had a strong feeling of, "My people! I have found them!" That's a harder feeling to get when nobody there looks like you. Consider Ellen Spertus's take on how it feels to be a woman in CS.
The second reason is selfish: If OSS programmers seems to be drawn inordinately from one ethnic group or gender, it's reasonable to suppose that there are people out there who could be perfectly good OSS programmers who aren't. The more of us there are, the more great software we all get. So it's in all of our interests to make sure our community is as open and welcoming as possible.
The web people believe (reasonably or not) that the form fields will be cleaned up by the backend people. [...] The backend people believe (reasonably or not) that the data will be cleaned up by the web people. [...] Then some **** adds a cross-scripting exploit and compromises sensitive information. Who's responsible, the developers or the managers?
Both. Everybody shares responsibility for a quality product. To say it's only the managers means that the programmers are interchangeable cogs in the software development machine. To say it's only the programmers says that the managers are superfluous. Neither should be the case.
Basically the co-worker saw what he did as "I showed them, I'm nobody's fool and they will see how powerful and valuable I am". The net result was that the CEO referred to the co-worker as 'a cancer on the company that should be removed at all costs'.
Heh. Maybe that's why the CEOs read employee mail: so they can root out the complete retards. Of course, the positive spin is that by inviting everybody to email him, he gets a lot of people thinking about the big picture, which is valuable on its own.
Really, I think it's a positive trend. Either the CEO listens only to the people directly under him, and so lives in a mediated-reality bubble, or he casts a wider net. Email from random people with a bee in their bonnet isn't an unbiased sample, but it's better than listening only to a bunch of people who do no actual work and all got their MBAs at the same 5 schools.
Most undergraduate students I've talked to about this idea say something along the lines of "Philosophy has nothing to do with science."
Wow! That demonstrates a pretty stunning ignorance of science itself. As I'm sure you know, all of what is now science used to be called natural philosophy.
Since when did scientists start behaving like fundies?
I thought it was roughly forever. It's the resistance that any revolutionary idea encounters before the revolution happens. But we shouldn't do away with it, as it's also the resistance that keeps out the zillions of bad ideas, and makes us confident that the few that make it through are worth keeping.
It's a shame that some are being jerks about it, but that's what you get for trying to do science with a bunch of monkeys who just have an extra layer of neurons wrapped around the primate core.
Get used to it: Japan and China will own the major technological innovations and discoveries 25 years from now.
At least on the basis of this, I'm not sweating it. Japan has a long history of funding big useless research in the name of industrial policy. A great example is their multi-billion-dollar push for the Fifth-Generation Computer.
Japan will be mining the moon for essential minerals before we ever get there again.
That's an intriguing prediction. You should register it with Long Bets.
For those, like me, who had trouble visualizing the flight path, here's the great cirle route. To my surprise, the most direct route is mainly over land.
sounds like an interesting methodology - but what does FUN stand for? How does it compare with SCRUM and is it Agile?
I know you were mainly kidding, but I think one of the secrets to the success of the assorted Agile methods is that everybody has more fun. If I were talking to a business exec, I'd cover it up by using a more suit-friendly term like "engaged", "motivated" or "morale-building", but from my perspective, fun is what it is.
In any event, I think questions like this are more helpful for management than they are for programmers or anybody else with a similar profession.
I don't know about that. I'd use the term "enjoyable" instead of "fun", but I strive to structure the projects I work on so that everybody is enjoying it. Like most people, I'm much more productive when I'm enjoying things, so there's a big financial benefit to arranging things so that people are eager to come to work every day. And I think it's every employee's responsibility to push for an environment that's maximally productive.
I'm sure many more people would have a cleaner house if cleaning their house was fun.
This works for me. We do an hour of cleaning together every Sunday while listening to Car Talk. Both of those make it fun in a way that doing solo chores wasn't, and our house is much cleaner.
This is a trivial example, but my point is that in a complex Java project, readable and elegant code bears no correlation to fast and efficient code. I believe this is why Java is slow.
I disagree. Most of the ugly systems I see are also slow. In my sample, at least, design clarity is strongly correlated with speed.
Good OO style can come at a moderate resource cost, although this article points out that JVMs get better and better at turning common OO structures into resource-optimal structures at run-time. But good OO style also gives you increased flexibility and maintainability. When you get to the point of optimizing, a clean design makes it much easier to find slow points and make them fast easily.
Here's my process:
make it work
make it right
make it fast
First, I get something working. Then I refactor the design to be clear and easy to understand. Only then do I worry much about speed. Most of the time I can get the speed I want without design distortion. On the rare occasions when I get speed by, say, breaking encapsulation, I do the minimum necessary and mark it clearly so that when the next generation of the JVM makes my tweaks obsolete, somebody can easily put things back.
But really, a surprising number of my performance issues go away at the design cleanup stage. A lot of performance problems are really design problems.
The copying collector sounds really fast indeed, but I can immediately see two problems: The first one is the need for a huge amount of memory. [...] The second is that performance is going to suck when garbage collection is performed.
I think that's more an effect of his simplified explanations. There are some pretty obvious optimizations that I'm sure Sun has taken. One of them he mentions: by doing thread-local pools, you parallelize things. That also means smaller buffers, so the cleanups are less likely to be noticeable, and easier to schedule when the thread is idle or blocked.
Further, he only mentions the copying GC when talking about young-generation GC. Since it only has to hold long-lived objects for a little while, I'd bet it will be a relatively small fraction of total memory. My guess is that they use something else that's more complicated for longer-lived objects. But they can afford it: because they've handled 95% of the allocation/deallocation cheaply, that gives you a lot more juice to handle the long-lived stuff.
It would seem that the optimal way of dealing with this is restricting the amount of memory available to the application, otherwise any app can grow to the maximum size allowed by the VM, whether it needs it or not. But this sounds rather crappy to me, now every developer needs to figure out an right limit for the application.
I do find that long-running 1.4 JVMs will slowly creep up in usage, even though there should be no need. In practice it's a little annoying, but I can live with it. I'd rather we had more knobs to twist, especially from inside the JVM. E.g. being able to say, "right now you can do GC for 50 ms". Or to hint to it about how happy I am to trade RAM usage vs CPU usage.
I'd also like more collaboration between the JVM and the host OS. If a system has a half-gig of RAM free and is currently under heavy CPU load, the JVM can be pretty slack for a while, and there's no need to throw OutOfMemory errors. But if the system is RAM-constrained and starting to swap, it's be great if the JVM automatically started freeing RAM and giving it back. Doing GC across swapped memory would suck much worse than a very aggressive GC.
There are lots of games done in Java, mainly for mobile phones through J2ME.
Note also the MMORPG Puzzle Pirates which is Java both on desktop and server. It's not cutting edge in the sense of pushing the boundaries of the latest graphics cards, but then I wouldn't do that in Java anyhow. But Puzzle Pirates shows you can make a commercially and critically successful modern game in Java.
Wow, that's really shocking. Until you actually look at the Detlef paper and realize that it was published in 1994, 11 years ago!! Who knows, maybe things have improved a bit in 11 years. The author certainly thinks Java is getting better; maybe it's possible that C/C++ compilers have improved as well.
So let me get this straight. You're accusing somebody of spreading FUD and your sole evidence is, uh, nothing? What do the U and D stand for again, hmmmm?
A JVM runs much heavier on the system, and when I run Netbeans, it is continuously on the verge of eating my 1.2 GB powerbook alive, in fact I have to frequently restart Netbeans to get memory back.
I don't know if that's an Apple problem or a Netbeans problem, but I don't have similar issues on an Intel Linux box using IntelliJ IDEA with a half-gig of RAM.
Try writing a Java program that eats less than 32k.
Why would that ever matter? Other than when doing embedded work, I can't remember the last time that that 32k mattered to me. I just bought some RAM yesterday, and 32k of that RAM cost me about 0.4 cents.
My current computer has 250,000 times the RAM my first computer did 25 years ago. My allowance that week was $10; if my salary had scaled similarly I'd be making $130 million a year, more than most Fortune 500 CEOs. That's no excuse for needless waste, but I can't take my money or my RAM with me, so I might as well enjoy it while I'm here.
The fact is that Java programs include more intelligence about programming from the compiler and the JVM. Which means that the same programmer budget can produce more product, or a lower budget can produce the same product.
Yes! And I don't think that means you should hire less competent programmers. All the Java programmers I want to work with are perfectly capable of working in other languages. The difference with Java is that we can spend our obsessive powers and more of our intellectual energy on things that matter.
Not that Java is perfect or anything; I have a lot of issues with it. But computers keep getting faster, while people stay about the same. Anything that can be cheaply and effectively automated should be, and for most projects I see, memory allocation is now firmly in that category.
Imagine the same songs, but done with a good metal band.
This is getting pretty off-topic, but you should check out Cookie Mongoloid; they do speed metal covers of Sesame Street songs. They're a lot of fun live.
A problem -- there is no way they will be hiring enough professionals to grade every song out there that I might be interested in. If they get a sufficient following, I see labels paying to have their songs indexed... good luck to the independent musicians out there.
I'd say it depends a lot on how their revenues look. But it also depends a lot on their per-song expenses.
There are a lot of underpaid musicians out there, and my guess is their experts can tag things pretty quickly. And they might not even have to pay: I knew people in college who would do this kind of tagging for free in exchange for free music and unfettered access to a huge music database like this. IMDB, CDDB, Flickr, and Wikipedia are all good examples of large datasets built with very little paid staff time.
Slashdot Mirror
It's not a race issue; it's a class issue.
People who believe that should try the Implicit Association Test for race. It measures your unconscious biases; you can see how well they correspond to your conscious ones.
Saying it's only a class issue is sweeping some things under the rug. People use race to guess class, and what social class people are in is affected by race, especially historically but also in the present. One of the ways that is mediated is through the unconscious bias that things like the IAT reveals.
I resent that. I didn't have anything handed to me on a silver platter and I'm as white as it gets.
I'm sure you mean this sincerely, but I promise you that you've had a good dose of white male privilege, just like me. I used to say the same thing you did, but a variety of experiences made me realize that I was like a fish who never noticed the water. Here are a couple:
After high school, I ended up living in South America for a year. As an American overseas, initial reactions to me were bimodal: either I was presumed a patrón, and therefore deserving of excess respect for my race, gender, and wealth (because all Americans are, from there perspective, wealthy) or I was dirt on the floor because I represented, in their minds, the oppressive Yanqui colonialist government. When I came back both were attenuated, but both reactions to my presumed status still existed, especially in cities like Chicago with a fairly mixed ethnic balance.
More recently, I went, with otherwise no change, from being a guy with a ponytail and a goatee to a guy with a shaved head and a goatee. The change in reaction was dramatic: people were scared of me, and would cross the street to get away from me. A black friend of mine remarked, "Ah, now you know what it's like." And he's right: although he's better looking, better dressed, more visibly well off, nicer, and comes from a better family, there's no question that strangers treated me, on average, better, especially once I grew some hair back.
I wish there were a modern version of Black Like Me. It wouldn't be as dramatic, but it would help people to see that racism still is with us in a thousand ways.
Does it really matter who (in the sense of ethnic diversity) writes the OSS code or makes a important innovation?
Yes, for two reasons. The first is societal. Remember the first time you went to some geeky gathering? I had a strong feeling of, "My people! I have found them!" That's a harder feeling to get when nobody there looks like you. Consider Ellen Spertus's take on how it feels to be a woman in CS.
The second reason is selfish: If OSS programmers seems to be drawn inordinately from one ethnic group or gender, it's reasonable to suppose that there are people out there who could be perfectly good OSS programmers who aren't. The more of us there are, the more great software we all get. So it's in all of our interests to make sure our community is as open and welcoming as possible.
Again the Murrays of the world jump in to explain why the black "race" is better suited for basketball than others (e.g. white men can't jump).
Do you know if Murray has any data on whether white people are genetically more prone to using minimal and/or dubious data to justify old prejudices?
Sadly, I'm guessing it's a human universal.
The web people believe (reasonably or not) that the form fields will be cleaned up by the backend people. [...] The backend people believe (reasonably or not) that the data will be cleaned up by the web people. [...] Then some **** adds a cross-scripting exploit and compromises sensitive information. Who's responsible, the developers or the managers?
Both. Everybody shares responsibility for a quality product. To say it's only the managers means that the programmers are interchangeable cogs in the software development machine. To say it's only the programmers says that the managers are superfluous. Neither should be the case.
If you want people to try to work on this, try posting an image.
I second that! If you're worried about copyright liability, just post the first and last copule megabytes of the disc.
Basically the co-worker saw what he did as "I showed them, I'm nobody's fool and they will see how powerful and valuable I am". The net result was that the CEO referred to the co-worker as 'a cancer on the company that should be removed at all costs'.
Heh. Maybe that's why the CEOs read employee mail: so they can root out the complete retards. Of course, the positive spin is that by inviting everybody to email him, he gets a lot of people thinking about the big picture, which is valuable on its own.
Really, I think it's a positive trend. Either the CEO listens only to the people directly under him, and so lives in a mediated-reality bubble, or he casts a wider net. Email from random people with a bee in their bonnet isn't an unbiased sample, but it's better than listening only to a bunch of people who do no actual work and all got their MBAs at the same 5 schools.
Most undergraduate students I've talked to about this idea say something along the lines of "Philosophy has nothing to do with science."
Wow! That demonstrates a pretty stunning ignorance of science itself. As I'm sure you know, all of what is now science used to be called natural philosophy.
Since when did scientists start behaving like fundies?
I thought it was roughly forever. It's the resistance that any revolutionary idea encounters before the revolution happens. But we shouldn't do away with it, as it's also the resistance that keeps out the zillions of bad ideas, and makes us confident that the few that make it through are worth keeping.
It's a shame that some are being jerks about it, but that's what you get for trying to do science with a bunch of monkeys who just have an extra layer of neurons wrapped around the primate core.
Get used to it: Japan and China will own the major technological innovations and discoveries 25 years from now.
At least on the basis of this, I'm not sweating it. Japan has a long history of funding big useless research in the name of industrial policy. A great example is their multi-billion-dollar push for the Fifth-Generation Computer.
Japan will be mining the moon for essential minerals before we ever get there again.
That's an intriguing prediction. You should register it with Long Bets.
It turns out there are extensive rules about this known as ETOPS. There's more information available in FAQ at the very cool Great Circle Mapper.
For those, like me, who had trouble visualizing the flight path, here's the great cirle route. To my surprise, the most direct route is mainly over land.
sounds like an interesting methodology - but what does FUN stand for? How does it compare with SCRUM and is it Agile?
I know you were mainly kidding, but I think one of the secrets to the success of the assorted Agile methods is that everybody has more fun. If I were talking to a business exec, I'd cover it up by using a more suit-friendly term like "engaged", "motivated" or "morale-building", but from my perspective, fun is what it is.
In any event, I think questions like this are more helpful for management than they are for programmers or anybody else with a similar profession.
I don't know about that. I'd use the term "enjoyable" instead of "fun", but I strive to structure the projects I work on so that everybody is enjoying it. Like most people, I'm much more productive when I'm enjoying things, so there's a big financial benefit to arranging things so that people are eager to come to work every day. And I think it's every employee's responsibility to push for an environment that's maximally productive.
I'm sure many more people would have a cleaner house if cleaning their house was fun.
This works for me. We do an hour of cleaning together every Sunday while listening to Car Talk. Both of those make it fun in a way that doing solo chores wasn't, and our house is much cleaner.
I disagree. Most of the ugly systems I see are also slow. In my sample, at least, design clarity is strongly correlated with speed.
Good OO style can come at a moderate resource cost, although this article points out that JVMs get better and better at turning common OO structures into resource-optimal structures at run-time. But good OO style also gives you increased flexibility and maintainability. When you get to the point of optimizing, a clean design makes it much easier to find slow points and make them fast easily.
Here's my process:
First, I get something working. Then I refactor the design to be clear and easy to understand. Only then do I worry much about speed. Most of the time I can get the speed I want without design distortion. On the rare occasions when I get speed by, say, breaking encapsulation, I do the minimum necessary and mark it clearly so that when the next generation of the JVM makes my tweaks obsolete, somebody can easily put things back.
But really, a surprising number of my performance issues go away at the design cleanup stage. A lot of performance problems are really design problems.
The copying collector sounds really fast indeed, but I can immediately see two problems: The first one is the need for a huge amount of memory. [...] The second is that performance is going to suck when garbage collection is performed.
I think that's more an effect of his simplified explanations. There are some pretty obvious optimizations that I'm sure Sun has taken. One of them he mentions: by doing thread-local pools, you parallelize things. That also means smaller buffers, so the cleanups are less likely to be noticeable, and easier to schedule when the thread is idle or blocked.
Further, he only mentions the copying GC when talking about young-generation GC. Since it only has to hold long-lived objects for a little while, I'd bet it will be a relatively small fraction of total memory. My guess is that they use something else that's more complicated for longer-lived objects. But they can afford it: because they've handled 95% of the allocation/deallocation cheaply, that gives you a lot more juice to handle the long-lived stuff.
It would seem that the optimal way of dealing with this is restricting the amount of memory available to the application, otherwise any app can grow to the maximum size allowed by the VM, whether it needs it or not. But this sounds rather crappy to me, now every developer needs to figure out an right limit for the application.
I do find that long-running 1.4 JVMs will slowly creep up in usage, even though there should be no need. In practice it's a little annoying, but I can live with it. I'd rather we had more knobs to twist, especially from inside the JVM. E.g. being able to say, "right now you can do GC for 50 ms". Or to hint to it about how happy I am to trade RAM usage vs CPU usage.
I'd also like more collaboration between the JVM and the host OS. If a system has a half-gig of RAM free and is currently under heavy CPU load, the JVM can be pretty slack for a while, and there's no need to throw OutOfMemory errors. But if the system is RAM-constrained and starting to swap, it's be great if the JVM automatically started freeing RAM and giving it back. Doing GC across swapped memory would suck much worse than a very aggressive GC.
There are lots of games done in Java, mainly for mobile phones through J2ME.
Note also the MMORPG Puzzle Pirates which is Java both on desktop and server. It's not cutting edge in the sense of pushing the boundaries of the latest graphics cards, but then I wouldn't do that in Java anyhow. But Puzzle Pirates shows you can make a commercially and critically successful modern game in Java.
Wow, that's really shocking. Until you actually look at the Detlef paper and realize that it was published in 1994, 11 years ago!! Who knows, maybe things have improved a bit in 11 years. The author certainly thinks Java is getting better; maybe it's possible that C/C++ compilers have improved as well.
So let me get this straight. You're accusing somebody of spreading FUD and your sole evidence is, uh, nothing? What do the U and D stand for again, hmmmm?
A JVM runs much heavier on the system, and when I run Netbeans, it is continuously on the verge of eating my 1.2 GB powerbook alive, in fact I have to frequently restart Netbeans to get memory back.
I don't know if that's an Apple problem or a Netbeans problem, but I don't have similar issues on an Intel Linux box using IntelliJ IDEA with a half-gig of RAM.
Try writing a Java program that eats less than 32k.
Why would that ever matter? Other than when doing embedded work, I can't remember the last time that that 32k mattered to me. I just bought some RAM yesterday, and 32k of that RAM cost me about 0.4 cents.
My current computer has 250,000 times the RAM my first computer did 25 years ago. My allowance that week was $10; if my salary had scaled similarly I'd be making $130 million a year, more than most Fortune 500 CEOs. That's no excuse for needless waste, but I can't take my money or my RAM with me, so I might as well enjoy it while I'm here.
The fact is that Java programs include more intelligence about programming from the compiler and the JVM. Which means that the same programmer budget can produce more product, or a lower budget can produce the same product.
Yes! And I don't think that means you should hire less competent programmers. All the Java programmers I want to work with are perfectly capable of working in other languages. The difference with Java is that we can spend our obsessive powers and more of our intellectual energy on things that matter.
Not that Java is perfect or anything; I have a lot of issues with it. But computers keep getting faster, while people stay about the same. Anything that can be cheaply and effectively automated should be, and for most projects I see, memory allocation is now firmly in that category.
Imagine the same songs, but done with a good metal band.
This is getting pretty off-topic, but you should check out Cookie Mongoloid; they do speed metal covers of Sesame Street songs. They're a lot of fun live.
A problem -- there is no way they will be hiring enough professionals to grade every song out there that I might be interested in. If they get a sufficient following, I see labels paying to have their songs indexed... good luck to the independent musicians out there.
I'd say it depends a lot on how their revenues look. But it also depends a lot on their per-song expenses.
There are a lot of underpaid musicians out there, and my guess is their experts can tag things pretty quickly. And they might not even have to pay: I knew people in college who would do this kind of tagging for free in exchange for free music and unfettered access to a huge music database like this. IMDB, CDDB, Flickr, and Wikipedia are all good examples of large datasets built with very little paid staff time.