Just as an aside, I believe that most of the content is not royalty-free, but photo.net has some really high-class photography, if you're just wanting to browse for interesting photographs.
That's a bit hyperbolic, but is exactly why we don't want the UN running DNS.
I don't want the UN running DNS because (a) the US has done a reasonable job thus far, IMHO, and the practical drawbacks of possibly screwing up DNS are greater than any ideological concerns for me and (b) it's not as if, if the US and the UN come into serious conflict over this, that the UN couldn't set up alternate roots in the future -- but right now, I don't see the need.
There was a section that went out of the way to highlight what appeared to br one womans unease about the privacy problems caused by Google's ability to store the results of a users searches - with no mention of the fact that in most cases all Gooogle will have is an IP address, or even that using Google isn't compulsory.
Two points.
First, I'd like to say that any search engine (or website, or whatever) is likely to do this. I recognize that it's kind of spooky to consider what kind of a profile someone like Google could build up on you, given how pervasive Google is -- which is why I wholeheartedly support Google giving the finger to the feds in general when it comes to their users' privacy. This may be a problem, but it's a search engine problem, not a Google problem.
Frankly, I think that we need tougher restrictions present on law enforcement obtaining search engine data. There are obviously practical problems inherent in defining what a "search engine" is, but hear me out. Traditionally, law enforcement could maybe get a warrant to start tapping a phone or search a house (and, incidently, they have to notify people that they *searched* the house, if they do so). I believe that LE can request phone records (though I don't know how far back, and in any event, this is at least somewhat limited information).
On the other hand, search engine data contains an entire history of what people have done on their computer for maybe years. This is absolutely unprecedented. It can be a snapshot spanning *years*. I think that there is too much incentive to grab data for some other claimed purpose and then abuse it -- it would clearly be very useful for political reasons.
I also worry about the chilling effects on thought -- it is as objectionable to me as feds being able to obtain library reading lists (worse, secretly). I want people to be able to read and educate themselves on things without worrying about whether or not that reading might be used against them at some time in the future -- if a lawyer wants to read about communist ideology, I don't think that that should eventually be used to prohibit him from becoming a Supreme Court Justice, for example.
I could see restrictions where LE cannot request data older than $N years, and possibly must go through a more substantial review process than a typical tap or search warrant (in which a judge determines that seizing search engine records is not only *useful* to an investigation, but that there is no other, less invasive, way to perform the request). Furthermore, I think that there should be a requirement to notify the person whose data was seized (in much the same way that house searches currently require notification). This provides some disincentive for "fishing trips".
Second, the woman being concerned was on BBC -- I'm guessing that she's European. European data privacy generally differs from US data privacy in that in the US, the government is often more limited in the personal data that they can obtain, but in Europe, corporations are often more limited in how they can handle personal data. Her concerns were probably about what Google (or someone buying the information from Google, or someone buying the information from them) could do, not with the government.
Oh, and one more followup. This applies to *application* code only, not to library code. I'm currently maintaining a cross-platform abstraction library. In this library always check malloc() results and propagate errors upwards (even though, on a default Linux setup, this code will never be executed). In this code, it makes sense to handle allocation failures and hand them back out, because it's fairly straightforward to propagate the error upwards and because the application using the library may be doing that init() phase where it needs allocations to be checked.
I would point out that from an educational aspect, Bush may be one of the least useful presidents to comment on because no one yet knows the long term outcome of his policies and actions.
That may be the most convincing argument I've heard yet, actually.:-)
Hmm. My statement was a little unclear. I wasn't calling "outsourcing" a technology, but referring to the technology that allows outsourcing (the Internet).
They are documents that they freely handed out, and things that they said in class. That's why there's nothing wrong with it.
It's legal to take pictures of someone in a public area. How would you feel if someone was filming you 24/7 every time you walked out of your house? Would it change how you act?
I think the point is that a public university should, on the whole, be POV neutral.
If, using the best facts and analysis that I have available, I conclude that Bush is not being an effective President, is it not neutral to make that statement in a political science class?
Good point. Unfortunately, vouchers seem to have stalled, thanks to the teachers union. Woohoo. Shame I can't dismantle the teachers' union on my own.
The people I see mostly supporting vouchers are people who advocate Christian ideology being taught in schools -- while a public school isn't allowed to do so, a private school using vouchers is.
These days, actual conservatives tend to vote for independent or libertarian candidates. They don't vote for the Democrats, and they sure don't vote for the Republicans, because neither party truly represents the views and ideals of conservatism.
I think that you're roping too much under the banner of "conservative". At the *very* least, you need to split fiscal and social values (and it's still not as if you can categorize everyone perfectly).
You can be socially liberal or conservative. The archetypal social conservative advocates gun rights, laws against abortion, Christianity in schools, and laws against gay marriage.
You can be fiscally liberal or conservative. The archetypal social conservative advocates limited spending, limited defense, and limited government regulation of business.
Personally, I think that a "legal" axis is also a useful idea, where legal conservatives favor traditional, more limited powers of the federal government, though I've never heard of someone using the term. It seems to be tied to fiscal conservativism.
The current Republican party is socially conservative. It is fiscally liberal. I would say that it is legally liberal.
The current Democratic party is socially liberal. It is fiscally liberal. It is legally liberal.
From what I understand of the Libertarian Party, it is socially liberal, fiscally conservative, and legally conservative.
The Constitution Party is socially conservative (*extremely* so), fiscally conservative, and legally conservative.
I tend to be socially liberal across the board (with the exception of gun rights and capital punishment). I thought that I was fiscally conservative, but one day I asked myself "okay, what branches of the federal government would you eliminate", and couldn't convince myself to eliminate anything significant -- what do you want to get rid of? The Park Service? The IRS? I do favor running a surplus (we have a *lot* of national debt to pay off), and would support whatever taxation is necessary to do so. I'm definitely legally conservative -- I think that a lot of federal powers should revert back to the states. This is all pretty complicated -- I can't really describe myself as *just* conservative or liberal. Similarly, you saying that the GOP is not conservative isn't really a meaningful statement.
The only way I ace political science classes is by parroting the pinko commie crap that the professor advocates right back to him in my assignments. Works every time.
Have you actually *tried* presenting a well-supported argument that opposes what you believe to be his views, or just assumed that he's out to get you?
I'm not talking about ranting; I mean seriously addressing points. I know an awful lot of people who assumed that professors were out to get them because of $RANDOM_PERSONAL_CHARACTERISTIC_OF_PERSON when in fact, they just weren't really doing very good work for the class.
The credit companies are data brokers. They are fairly incompetent and close enough to being full time criminals.
Do you have data to support them being incompetent? I remember one professor who had worked in statistical analysis for credit card firms for some time. It seemed that they were pretty sophisticated to me (but then again, I'm not a statistican).
Unfortunately, it's prohibitively expensive to try and retrofit this into a prexisting home/building.
I don't know much about air ductwork, but surely there's some way to access the stuff for maintenance, cleaning out vermin/water, etc?
And if you do that, as long as the ductwork is an acyclic graph, I can't imagine why it would be that expensive. It would presumably take a motorized panel in the duct, wiring run to the control system, and a sensor...but I can't imagine that being prohbitively expensive.
This is the worst advice ever posted on Slashdot. Congratulations. The scary part is that it got modded "Insightful".
It certainly illustrates the diversity of views.
It is not impossible to cleanly back out failure states. It is difficult.
I agree that it is not impossible. However, I will argue that it is not worth development time. Your software will never be perfectly bug-free, so you have to figure out where to spend your time working on it. Checking malloc() and writing code to back out of allocations is just not an effective use of that time.
Don't confuse the two. Your job as a programmer is not to make your life easy, it's to write programs that function correctly. Handling memory errors correctly (say, not losing the user's states and corrupting their current working file due to allocation error) is part of the program's role. Advice like this leads to the type of apps that die mysteriously for no apparent reason, leaving no trace or debug message of what went wrong.
Wrap malloc() with abort(), or something similar. If abort() fires, your program has a bug.
But handling most error cases correctly (obviously I try for total coverage) is better than handling none of them.
Maybe. Or it might lead to cascading problems, making the original bug harder to track down. I think that pretty much unilaterally all the large applications that I used to run on platforms with hard memory limits, like classic Mac OS, would die in one way or another if they ran out of memory. Maybe you'd see a series of three error dialogs come up about a lack of memory, and then the program would crash. [shrug]
This is an incorrect assertion and a terrible programming practice. Imaging your Linux box is running 1000 processes, 1 sshd and 999 apache instances. If traffic spikes up, and the apaches start consuming more memory, the system may not have enough to give them, and they may gracefully handle it. But if sshd was coded using your standards, it would crash if someone tried to initiate a connection because it did not check its malloc return. I think we can all agree that under no circumstances should increased memory use by apache (not a bug) cause sshd to crash.
This is why I said "almost all circumstances" -- you're right that without an OOM killer, something like this could happen. The idea is that a typical desktop system *will* start to break under memory pressure. Probably processes will start crashing -- even if all memory allocation errors are handled and the actions backed out, things like ssh will become unusable because they have no memory to work with. Unless the system's memory usage starts dropping, even perfect software is going to sit there, immobilized.
Now, yes, the OOM killer isn't perfect, but the idea is that memory exhaustion should never be reached on a system unless a bug is present, and then the OOM killer blows away that process.
My point is that chances are pretty good that either a software package is allocating memory in a loop or has made a huge allocation if memory is exhausted -- so that package is probably the one at fault, so even without an OOM killer, abort()ing in the event of memory exhaustion is probably not unreasonable.
And how exactly does the GIMP know how much memory is available to it at any time? It may be running on a box with 16 megs of RAM, or 16 gigs. The system may have memory allocated to other things right now that was available five seconds ago. I'd like to see portable code that can predict when nontrivial resource allocation will fail.
It doesn't -- it might throw up a warning dialog when one isn't necessary, and I suppose that if you were running it on a box with 16 megs of RAM, it might fail. But you also aren't going to make the GIMP bulletproof, and this is a pretty reasonable strategy.
How else can you explain college costs going up, but graduate aptitude going down?
The fact that it's much harder to get a decent job without a college degree than it was, say, fifty years ago.
This means more demand -- there's your higher prices, the free market at work.
That also would explain a decrease in the average aptitude (if such exists), as instead of just taking the best available, say, n% students and having them go to college, n*5% of people are getting degrees.
I'm not sure we'll see any real changes until we remove the federal funding of education from all education, especially the college grants and loans that the government seems to happy to dole out.
We spend money on grants, student loans, and scholarships because education is an investment with a pretty good turnaround.
If Joe wants to work at Jim-Bob's carwash, that's fine, but if Joe doesn't want to and has the potential to be a really good mathematician, I'd hate to see that go to waste because he couldn't fund college.
If you think that eliminating federal education subsidies is likely to produce a better society, I'm interested in what factor you think is more overriding than an educated society.
The Nazis didn't start out in control of the government. They and the groups that they sprang from (nationlist right-wingers with a good deal of support from the military) started out by intimidating opposition and those who spoke against them.
Why would you want to take it *off* the ebook reader?
I mean, I read lengthy HTML and text files regularly on my computer -- right now, I'm reading a free copy of Journey to the West. I have a bunch of scripts that render them to PDF and make them look really nice (two columns of text, antialising), so that I can just tap spacebar or backspace in xpdf to read them. Half of the problem with reading long amounts of text on a computer is that the readers are abysmal (web browsers are not suitable), and this solves the problem.
I've *never* been unhappy that I have to use PDF -- I still have the master.
I'm suspicious that some of the people here are just worried that they won't be able to pirate commercial ebooks. Don't worry -- nobody's ever managed to effectively lock down a consumer electronics device. I'm quite sure that someone will crack this in short order, and in any event, books can be obtained in other formats.
Slashdot Mirror
Just as an aside, I believe that most of the content is not royalty-free, but photo.net has some really high-class photography, if you're just wanting to browse for interesting photographs.
That's a bit hyperbolic, but is exactly why we don't want the UN running DNS.
I don't want the UN running DNS because (a) the US has done a reasonable job thus far, IMHO, and the practical drawbacks of possibly screwing up DNS are greater than any ideological concerns for me and (b) it's not as if, if the US and the UN come into serious conflict over this, that the UN couldn't set up alternate roots in the future -- but right now, I don't see the need.
There was a section that went out of the way to highlight what appeared to br one womans unease about the privacy problems caused by Google's ability to store the results of a users searches - with no mention of the fact that in most cases all Gooogle will have is an IP address, or even that using Google isn't compulsory.
Two points.
First, I'd like to say that any search engine (or website, or whatever) is likely to do this. I recognize that it's kind of spooky to consider what kind of a profile someone like Google could build up on you, given how pervasive Google is -- which is why I wholeheartedly support Google giving the finger to the feds in general when it comes to their users' privacy. This may be a problem, but it's a search engine problem, not a Google problem.
Frankly, I think that we need tougher restrictions present on law enforcement obtaining search engine data. There are obviously practical problems inherent in defining what a "search engine" is, but hear me out. Traditionally, law enforcement could maybe get a warrant to start tapping a phone or search a house (and, incidently, they have to notify people that they *searched* the house, if they do so). I believe that LE can request phone records (though I don't know how far back, and in any event, this is at least somewhat limited information).
On the other hand, search engine data contains an entire history of what people have done on their computer for maybe years. This is absolutely unprecedented. It can be a snapshot spanning *years*. I think that there is too much incentive to grab data for some other claimed purpose and then abuse it -- it would clearly be very useful for political reasons.
I also worry about the chilling effects on thought -- it is as objectionable to me as feds being able to obtain library reading lists (worse, secretly). I want people to be able to read and educate themselves on things without worrying about whether or not that reading might be used against them at some time in the future -- if a lawyer wants to read about communist ideology, I don't think that that should eventually be used to prohibit him from becoming a Supreme Court Justice, for example.
I could see restrictions where LE cannot request data older than $N years, and possibly must go through a more substantial review process than a typical tap or search warrant (in which a judge determines that seizing search engine records is not only *useful* to an investigation, but that there is no other, less invasive, way to perform the request). Furthermore, I think that there should be a requirement to notify the person whose data was seized (in much the same way that house searches currently require notification). This provides some disincentive for "fishing trips".
Second, the woman being concerned was on BBC -- I'm guessing that she's European. European data privacy generally differs from US data privacy in that in the US, the government is often more limited in the personal data that they can obtain, but in Europe, corporations are often more limited in how they can handle personal data. Her concerns were probably about what Google (or someone buying the information from Google, or someone buying the information from them) could do, not with the government.
Oh, and one more followup. This applies to *application* code only, not to library code. I'm currently maintaining a cross-platform abstraction library. In this library always check malloc() results and propagate errors upwards (even though, on a default Linux setup, this code will never be executed). In this code, it makes sense to handle allocation failures and hand them back out, because it's fairly straightforward to propagate the error upwards and because the application using the library may be doing that init() phase where it needs allocations to be checked.
I would point out that from an educational aspect, Bush may be one of the least useful presidents to comment on because no one yet knows the long term outcome of his policies and actions.
:-)
That may be the most convincing argument I've heard yet, actually.
Hmm. My statement was a little unclear. I wasn't calling "outsourcing" a technology, but referring to the technology that allows outsourcing (the Internet).
They are documents that they freely handed out, and things that they said in class. That's why there's nothing wrong with it.
It's legal to take pictures of someone in a public area. How would you feel if someone was filming you 24/7 every time you walked out of your house? Would it change how you act?
I think the point is that a public university should, on the whole, be POV neutral.
If, using the best facts and analysis that I have available, I conclude that Bush is not being an effective President, is it not neutral to make that statement in a political science class?
Good point. Unfortunately, vouchers seem to have stalled, thanks to the teachers union. Woohoo. Shame I can't dismantle the teachers' union on my own.
The people I see mostly supporting vouchers are people who advocate Christian ideology being taught in schools -- while a public school isn't allowed to do so, a private school using vouchers is.
the summary makes it clear that he's finding out professors from either side who presents one sided arguments in the classroom
The summary makes it clear that that is what he *says* that he's doing.
His web page does not strike me as a that of a group that has particularly neutral goals in mind...
These days, actual conservatives tend to vote for independent or libertarian candidates. They don't vote for the Democrats, and they sure don't vote for the Republicans, because neither party truly represents the views and ideals of conservatism.
I think that you're roping too much under the banner of "conservative". At the *very* least, you need to split fiscal and social values (and it's still not as if you can categorize everyone perfectly).
You can be socially liberal or conservative. The archetypal social conservative advocates gun rights, laws against abortion, Christianity in schools, and laws against gay marriage.
You can be fiscally liberal or conservative. The archetypal social conservative advocates limited spending, limited defense, and limited government regulation of business.
Personally, I think that a "legal" axis is also a useful idea, where legal conservatives favor traditional, more limited powers of the federal government, though I've never heard of someone using the term. It seems to be tied to fiscal conservativism.
The current Republican party is socially conservative. It is fiscally liberal. I would say that it is legally liberal.
The current Democratic party is socially liberal. It is fiscally liberal. It is legally liberal.
From what I understand of the Libertarian Party, it is socially liberal, fiscally conservative, and legally conservative.
The Constitution Party is socially conservative (*extremely* so), fiscally conservative, and legally conservative.
I tend to be socially liberal across the board (with the exception of gun rights and capital punishment). I thought that I was fiscally conservative, but one day I asked myself "okay, what branches of the federal government would you eliminate", and couldn't convince myself to eliminate anything significant -- what do you want to get rid of? The Park Service? The IRS? I do favor running a surplus (we have a *lot* of national debt to pay off), and would support whatever taxation is necessary to do so. I'm definitely legally conservative -- I think that a lot of federal powers should revert back to the states. This is all pretty complicated -- I can't really describe myself as *just* conservative or liberal. Similarly, you saying that the GOP is not conservative isn't really a meaningful statement.
The only way I ace political science classes is by parroting the pinko commie crap that the professor advocates right back to him in my assignments. Works every time.
Have you actually *tried* presenting a well-supported argument that opposes what you believe to be his views, or just assumed that he's out to get you?
I'm not talking about ranting; I mean seriously addressing points. I know an awful lot of people who assumed that professors were out to get them because of $RANDOM_PERSONAL_CHARACTERISTIC_OF_PERSON when in fact, they just weren't really doing very good work for the class.
The credit companies are data brokers. They are fairly incompetent and close enough to being full time criminals.
Do you have data to support them being incompetent? I remember one professor who had worked in statistical analysis for credit card firms for some time. It seemed that they were pretty sophisticated to me (but then again, I'm not a statistican).
Unfortunately, it's prohibitively expensive to try and retrofit this into a prexisting home/building.
I don't know much about air ductwork, but surely there's some way to access the stuff for maintenance, cleaning out vermin/water, etc?
And if you do that, as long as the ductwork is an acyclic graph, I can't imagine why it would be that expensive. It would presumably take a motorized panel in the duct, wiring run to the control system, and a sensor...but I can't imagine that being prohbitively expensive.
This is the worst advice ever posted on Slashdot. Congratulations. The scary part is that it got modded "Insightful".
It certainly illustrates the diversity of views.
It is not impossible to cleanly back out failure states. It is difficult.
I agree that it is not impossible. However, I will argue that it is not worth development time. Your software will never be perfectly bug-free, so you have to figure out where to spend your time working on it. Checking malloc() and writing code to back out of allocations is just not an effective use of that time.
Don't confuse the two. Your job as a programmer is not to make your life easy, it's to write programs that function correctly. Handling memory errors correctly (say, not losing the user's states and corrupting their current working file due to allocation error) is part of the program's role. Advice like this leads to the type of apps that die mysteriously for no apparent reason, leaving no trace or debug message of what went wrong.
Wrap malloc() with abort(), or something similar. If abort() fires, your program has a bug.
But handling most error cases correctly (obviously I try for total coverage) is better than handling none of them.
Maybe. Or it might lead to cascading problems, making the original bug harder to track down. I think that pretty much unilaterally all the large applications that I used to run on platforms with hard memory limits, like classic Mac OS, would die in one way or another if they ran out of memory. Maybe you'd see a series of three error dialogs come up about a lack of memory, and then the program would crash. [shrug]
This is an incorrect assertion and a terrible programming practice. Imaging your Linux box is running 1000 processes, 1 sshd and 999 apache instances. If traffic spikes up, and the apaches start consuming more memory, the system may not have enough to give them, and they may gracefully handle it. But if sshd was coded using your standards, it would crash if someone tried to initiate a connection because it did not check its malloc return. I think we can all agree that under no circumstances should increased memory use by apache (not a bug) cause sshd to crash.
This is why I said "almost all circumstances" -- you're right that without an OOM killer, something like this could happen. The idea is that a typical desktop system *will* start to break under memory pressure. Probably processes will start crashing -- even if all memory allocation errors are handled and the actions backed out, things like ssh will become unusable because they have no memory to work with. Unless the system's memory usage starts dropping, even perfect software is going to sit there, immobilized.
Now, yes, the OOM killer isn't perfect, but the idea is that memory exhaustion should never be reached on a system unless a bug is present, and then the OOM killer blows away that process.
My point is that chances are pretty good that either a software package is allocating memory in a loop or has made a huge allocation if memory is exhausted -- so that package is probably the one at fault, so even without an OOM killer, abort()ing in the event of memory exhaustion is probably not unreasonable.
And how exactly does the GIMP know how much memory is available to it at any time? It may be running on a box with 16 megs of RAM, or 16 gigs. The system may have memory allocated to other things right now that was available five seconds ago. I'd like to see portable code that can predict when nontrivial resource allocation will fail.
It doesn't -- it might throw up a warning dialog when one isn't necessary, and I suppose that if you were running it on a box with 16 megs of RAM, it might fail. But you also aren't going to make the GIMP bulletproof, and this is a pretty reasonable strategy.
How else can you explain college costs going up, but graduate aptitude going down?
The fact that it's much harder to get a decent job without a college degree than it was, say, fifty years ago.
This means more demand -- there's your higher prices, the free market at work.
That also would explain a decrease in the average aptitude (if such exists), as instead of just taking the best available, say, n% students and having them go to college, n*5% of people are getting degrees.
It probably tells you that they live in a world that is seniority based instead of merit based.
You think that *academia* is excessively seniority-based rather than merit-based?
You need to work at a large company for a while...
Libertarians have liberal social values.
Those who invoke Godwin's Law often do so because they lack a counterargument.
I'll bet in your ideal world, people don't get out of college with much by the way of analytic thinking or debating ability.
I'm not sure we'll see any real changes until we remove the federal funding of education from all education, especially the college grants and loans that the government seems to happy to dole out.
We spend money on grants, student loans, and scholarships because education is an investment with a pretty good turnaround.
If Joe wants to work at Jim-Bob's carwash, that's fine, but if Joe doesn't want to and has the potential to be a really good mathematician, I'd hate to see that go to waste because he couldn't fund college.
If you think that eliminating federal education subsidies is likely to produce a better society, I'm interested in what factor you think is more overriding than an educated society.
Academia is considered to be more liberal than conservative, or at least it's presented as such.
You ever wonder whether it might just not be a coincidence that the more learned members of society disagree with conservative values?
The Nazis didn't start out in control of the government. They and the groups that they sprang from (nationlist right-wingers with a good deal of support from the military) started out by intimidating opposition and those who spoke against them.
Why would you ever carry a credit card balance month-to-month, though? It's a pretty damn expensive way to borrow money...
Why would you want to take it *off* the ebook reader?
I mean, I read lengthy HTML and text files regularly on my computer -- right now, I'm reading a free copy of Journey to the West. I have a bunch of scripts that render them to PDF and make them look really nice (two columns of text, antialising), so that I can just tap spacebar or backspace in xpdf to read them. Half of the problem with reading long amounts of text on a computer is that the readers are abysmal (web browsers are not suitable), and this solves the problem.
I've *never* been unhappy that I have to use PDF -- I still have the master.
I'm suspicious that some of the people here are just worried that they won't be able to pirate commercial ebooks. Don't worry -- nobody's ever managed to effectively lock down a consumer electronics device. I'm quite sure that someone will crack this in short order, and in any event, books can be obtained in other formats.