Actually, that evolution occurs on some level is fact and is demonstrated by bacteria, viruses, insects, and probably a few other short lived life forms.
Evolutionary theory extends these observations to apply to living things with longer life spans where we cannot directly observe evolution. It also attempts to explain the mechanisms for evolution. Evolutionary theory is constantly being refined and adjusted because the current dataset is very small relative to the coverage of the theory. We have not unearthed all available fossils, or observed all the life currently on this planet by a long shot everytime something new is found it might not fit the current theory and it has to be adjusted.
Look at how Newtonian Mechanics gave way to Relativity which was extended by Quantum Mechanics. And Quantum Mechanics will eventually give way to an even more unified theory which combines Relativity and Quantum Mechanics. Yet, we still teach Newtonian Mechanics in school. So, the argument that because Evolution has holes, it shouldn't be taught, should logically extend to Newton and Einstein. Yet, I can't imagine anyone saying Newton or Einstein should not be taught in school.
Dastardly
Re:sounds good for robots
on
uCsimm News
·
· Score: 1
What about using the ethernet connection to handle communication between multiple uCsimms in the saem robot. Wired ethernet should be able to handle that. Aren't there 10baseT to wireless converters?
And, consider that some of that source code increase is due to platform specific things like x86, Alpha, SPARC, UltraSparc, MIPS, and PPC ports. And, drivers for various hardware that doesn't overlap, and support for a lot of filesystems.
The only exscuse MS has is drivers.
As for distribution bloat, that is expected, and ihas less to do with Linux and more to do with the increasing amount of Open Source software available to put in a Distribution. So, in a sense distribution bloat is a good thing because it says there is a lot of open source software out there to put in a distribution.
Actually, it isn't that he will get kicked out of school. First, he is going to fail one of his CS clsses because all his scholl work was on the Harvard computer. Second, if he gets charged with computer crimes he could have his computer accounts revoked, which the University can pretty much do at will. As a CS major having your computer account revoked is pretty much equivalent to being kicked out.
Multithreading could help with a single CPU and multiple NICs. A single thread could get blocked waiting for I/O, in this case a second thread could use the CPU to service requests from another NIC. Whether this is really adavantageous depends on a lot of factors including whether the IP code ever does block, and what other things can be done when one IP thread is blocked (what resources are locked during the blocking).
Currently, a serial number in a document might be tagged as or the equivalent proprietary tag. In XML the serial number could be tagged as , which tells us what the data is not what it looks like. A DTD tells the writer what tags he can use and the DTD can be customized.
This should read:
Currently, a serial number in a document might be tagged as "italics" or the equivalent proprietary tag. In XML the serial number could be tagged as "serialnumber" , which tells us what the data is, not just what it looks like.
Yes, word processors can create incompatible XML in the sense that they won't necessarily be able to understand each others tags, or display them in an identical fashion.
The real advantage of XML and, SGML as well, is in large scale document publishing. The advantage of XML is that it decouples content creation from formatting. Think about how mch time is wasted deciding whether something should be bold or bold-italics etc... Especially since the formatting says nothing about the content. For example: Currently, a serial number in a document might be tagged as or the equivalent proprietary tag. In XML the serial number could be tagged as , which tells us what the data is not what it looks like. A DTD tells the writer what tags he can use and the DTD can be customized.
In parallel, some one else (myself) is writing style sheets for formatting the tags for display in a browser, or conversion to html, rtf, pdf, etc. But, the key is that now writers will be able to concentrate on content creation, while some one else deals with formatting. It also enforces consistent formatting across writers.
The other advantage of XML is that cusomized tags give information about the content that can make search engines much smarter.
To relate this back to proprietary file formats, XML is only cross-platform WYSIWYG when the DTD and style sheet is either sent with the document or known to all. A word processor can very easily use obfuscated tags, and an internal style sheet. Which while the file may be saved as XML, and the content is available to anyone. No one without the same program can accurately reproduce the document.
Dastardly
P.S. XML does not require a DTD, but does require a style sheet of some sort for accurate reproduction. A DTD is mostly useful for document creation.
You're joking, right? Have you actually gone to the CEO of any company and told him/her that it's time to shut down marketing? "They can just take the next 2 months off while we re-engineer the back end systems."
You'll be lucky if you're not fired outright. And if they listen to you, they are even crazier than you:-)
Maybe that whould be the punishment for marketing incorrectly projecting demand. The only problem is that marketing tends to hae the CEOs ear more than IT. Therefore, IT gets blamed for only meeting marketings projections. But, that is showing my own prejudices.:-)
I think one of the real problems is a lot of companies are not doing the research that will give them good capacity projections. I believe it was Schwab in the article that said they went from quarterly analysis of capacity vs demand to every couple weeks, plus they have plenty of excess capacity, which is required sice demand can spike in days if not hours, while adding capacity probably has lead times measured in weeks.
First, as others have pointed out 2.4 will be the stable release of the 2.3 development kernel. Second, as no one else has pointed out, part of the kernel release process includes a feature freeze. There is a point during the 2.3 development when Linus will stop accepting new features, and only be accepting bug fixes. Essentially, what Linus is saying is that the feature freeze will occur sooner with fewer new features to debug resulting in more stable, but less featureful(sp?) stable kernel releases. This is probably a good thing.
I fervently hope that this bug is used to repeatedly down the stock market and a few military computers. That would put the spotlight like nothing else on Microsoft's failings.
Yeah right. The only thing that would do is get those people arrested and put in jail for computer crimes. No one blames MS for opening a big hole for viruses in Word and Excel. Instead they go after the guy who wrote Melissa. Doesn't anyone think it is extremely stupid that there is the ability to write word processor viruses??
All that would happen if some one exploited this, is that they would be arrested (if they could be tracked down). MS would not be blamed, and probably Eeye and slashdot would be blamed for publicizing the security whole.
It would be ironic if high end K7 chipsets were delayed because they decided to invest a lot of effort getting Rambus to work, and then the RAM modules don't turn up. I think for the high end, with huge 2nd level caches and enormous bandwidth requirements Rambus may have the edge if the caches take the top off the latency problems, and AMD may have thought the same way. And who would have guessed that an Intel-sponsored technology could fail in the PC space?
Rambus currently has no edge as far as quoted numbers go. Rambus bandwidth numbers are 1.6GB/s at 800MHz. PC266 DDR-SDRAM bandwidth numbers are 2.1GB/s. The only advantage I have heard for Rambus is some ephemeral promise that it offers better future scalability (vaporware). Can anyone show any pre-2001 plans for Rambus to match PC266?
Yes, it does grow the only thing is that it doesn't grow that fast. Consider that a white dwarf (~1 solar mass) is about the size of the earth. A neuton star (~1.5-2.5 solar masses) is about 10 miles wide. A stellar black hole (>3 solar masses) has a Shwarzchild diameter of about 1 mile. Last I checked the largest galactic blackholes (millions [billions?] of solar masses) have diameters no bigger than the diameter of our solar system.
As for the gravitational effects at this distance, the Eta Carinae would have the same effect as before because its mass didn't change (as others have described). Case in point, if some how the Sun were instantaneously replaced with a black hole of identical mass, the only effect on the earth would be that it would become really cold and dark.
The lower limit for a stellar mass black hole is only about 3 solar masses. Eta Carinae is about 100 solar masses. This doesn't mean every star over 3 solar masses will become a black hole. It just means if what is left after the super(or hyper)nova is bigger than 3 solar masses, the remnant will collapse into a black hole.
Small problem with using software rendering is that you will be benchmarking the memory subsystem more than the CPU. A software renderer has to load textures from main memory to the CPU in order to map them onto polygons, this is a very memory intensive process, and will easily overwhelm any calculating speed improvements. A hardware renderer on a the fastest possible 3D card with reams of texture memory would be more indicative of calculating performance than a software renderer, at least for Q2.
I guess this is a performance benchmark, so waiting for everything to reach peak performance, then quoting the numbers has some validity. But, for a complete review why not push the machines even further? Wouldn't it be even more interesting to see how each OS/server combination handles an overload?
Sometimes I feel like a leech, I use all this opensource stuff and don't really give anything back. Well, while you may not have given back anything yet. Users are part of the Open source process as well, even the ones that don't write any code. If you were to run into a bug, would you just sit there and pout. Certainly not, you would go on the net and see if anyone else had seen it. Being an intelligent user you would give a detailed description of your system, OS config, and what you were doing. Then, a couple things might happen. The bug has been seen before, and there is a fix out or on its way. The bug hasn't been seen, but a programmer type or five see your problem and work with you on isolating, fixing and/or working around the bug. Or, improbably, no one is interested in fixing it, but you decide you need it fixed so with the advice of some of these net people you fix the problem yourself. Either way you have contributed to the open source process, though not necessarily as a programmer, but as a tester and QA person.
It is kind of scary that having 10 years at a single job is now a liability instead of an asset. It used to be that corporations invested in their employees and in exchange wanted long term employment. Now they expect you to come in with a full skill set and then (ab)use you for a year or two, then cast you off. Luckily I am working at a place hat doesn't do that. It is just annoying that should I decide I want to do something different it will be harder because I actually like my job.
While guns cannot be blamed as a cause, they are a gating factor in the sense that without guns you can't go on a shooting rampage.
Of course even if guns could be completely eliminated (not likely) you are still left with a person with problems who wants to go out and shoot some one, but at least they don't have a gun. Instead they will come up with something else (bombs, poisons, knives, etc.) It is interesting the way people on both sides of gun control don't seem to realize how guns relate to the crime itself. The NRA says "Guns don't kill people, people kill people", but guns enable a person to shoot some one, and therefore is an enabling factor. While the gun control nuts figure that guns are the cause when they are really just an enabling factor that doesn't do anything to address the fact that a person actually wanted to kill another person. The solution is not simple, but simple solutions is what gets politicians elected.
I think the causes are an OK start, but terribly incomplete. I don't profess to know all the events, feelings, and influences that came together to cause these two kids to decide that shooting up the school was a good idea. The key though is that it wasn't one thing, it wasn't four things, it was probably more like 15 different things including the 4 described above, but not limited to those. The key though is that had any one of those things been missing this would not have happened.
Some might then say the solution is to eliminate one of the causes and that way this will never happen again. This is where gun control comes in. While guns don't cause a person to go out and mass murder. They are a gating factor in that without the guns the person could not act. But, this is too simple a solution, and still leaves us with an outcast, depressed, and anti-social person.
I also have a problem with blaming the violent movies, music, and video games for violence. It is hard to come up with any correlation let alone causation. If you say well there were 100 shootings where the killers liked Marilyn Manson, I will point out that there were 1 million other people who listened to Marilyn Manson and didn't shoot people. The only way to get correlation is if this number is greater than total killers vs the general population. And, that is still only correlation, there is no way to show causation. Listening to Marilyn Manson is more likely an effect just like going out and shooting people is an effect.
The point is there has to be a differentiation between things that correlate and things that might cause a person to be a killer. We also need to know which solutions remove causes and which are simply gating solutions that do nothing for the person.
There is absolutely no proof of this, and can't be unless a statistcally significant population of Dvorak users can be found. But, isn't there the possibility that while speed might not improve, repetetive stress injuries might be reduced by the Dvorak keyboard?
oh, so in order for the "bazaar" to be successful, first I have to most of the heavy lifting and then give the code away for free to the rest of the world.
Not exactly, but you should have something that compiles, runs, and does something useful. It doesn't necesarily have to have every feature, or even be completely bug free. If you don't have something that runs minimally it is real hard to spark enough interest to gain critical mass and have the open source project take off. That bare minimum should not be heavy lifting by any stretch of the imagination.
If the internet were not involved none of these cases of a 14 year old flying out to be with some man they think they love would make anything mor ethan the local news. There seems to be this inherent assumption that there are more perverts, rapists, child molesters, or whatever on the internet than there are in society at large. But, there is no good evidence one way or another that I have found that isn't anecdotal. So, here is my proposed challenge to the Slashdot effect.
Let's find a good study one or the shows one way or the other whether there is a greater or lower occurrence of perverts, rapists, child molesters, with internet ties than without. I bet any difference would be statistically insignificant.
The study also should compensate for the differences between the people who are on the on the internet vs not. For example: the is a minimum affluence level for hte internet where the person has to at least be able to afford a computer and internet connection.
One thing that needs to be brought up is that 40K in one place is different form 40K in another. For example: In Silicon Valley 2 bed, 2 bath apartments go for minimum 2K a month. (At least anywhere you woudl want to live.) So, you are guaranteeed to have a roommate, and still pay almost half your take home in rent.
In other places the same person might go for $800 a month, so with a roomamte that coes to only $400. In which case, 30K is OK.
So, don't jump to conclusions about what amount is reasonable.
Slashdot Mirror
Actually, that evolution occurs on some level is fact and is demonstrated by bacteria, viruses, insects, and probably a few other short lived life forms.
Evolutionary theory extends these observations to apply to living things with longer life spans where we cannot directly observe evolution. It also attempts to explain the mechanisms for evolution. Evolutionary theory is constantly being refined and adjusted because the current dataset is very small relative to the coverage of the theory. We have not unearthed all available fossils, or observed all the life currently on this planet by a long shot everytime something new is found it might not fit the current theory and it has to be adjusted.
Look at how Newtonian Mechanics gave way to Relativity which was extended by Quantum Mechanics. And Quantum Mechanics will eventually give way to an even more unified theory which combines Relativity and Quantum Mechanics. Yet, we still teach Newtonian Mechanics in school. So, the argument that because Evolution has holes, it shouldn't be taught, should logically extend to Newton and Einstein. Yet, I can't imagine anyone saying Newton or Einstein should not be taught in school.
Dastardly
What about using the ethernet connection to handle communication between multiple uCsimms in the saem robot. Wired ethernet should be able to handle that. Aren't there 10baseT to wireless converters?
Anyone out there do Robot Wars? One of these would probably work nicely. ;-)
And, consider that some of that source code increase is due to platform specific things like x86, Alpha, SPARC, UltraSparc, MIPS, and PPC ports. And, drivers for various hardware that doesn't overlap, and support for a lot of filesystems.
The only exscuse MS has is drivers.
As for distribution bloat, that is expected, and ihas less to do with Linux and more to do with the increasing amount of Open Source software available to put in a Distribution. So, in a sense distribution bloat is a good thing because it says there is a lot of open source software out there to put in a distribution.
Actually, it isn't that he will get kicked out of school. First, he is going to fail one of his CS clsses because all his scholl work was on the Harvard computer. Second, if he gets charged with computer crimes he could have his computer accounts revoked, which the University can pretty much do at will. As a CS major having your computer account revoked is pretty much equivalent to being kicked out.
Multithreading could help with a single CPU and multiple NICs. A single thread could get blocked waiting for I/O, in this case a second thread could use the CPU to service requests from another NIC. Whether this is really adavantageous depends on a lot of factors including whether the IP code ever does block, and what other things can be done when one IP thread is blocked (what resources are locked during the blocking).
Dastardly
I believe the military does require certain documents to follow an SGML DTD. I don't remember the name of the DTD, but it is out there.
Dastardly
Oops, fell into an html trap.
Currently, a serial number in a document might be tagged as or the equivalent proprietary tag. In XML the serial number could be tagged as , which tells us what the data is not what it looks like. A DTD tells the writer what tags he can use and the DTD can be customized.
This should read:
Currently, a serial number in a document might be tagged as "italics" or the equivalent proprietary tag. In XML the serial number could be tagged as "serialnumber" , which tells us what the data is, not just what it looks like.
Yes, word processors can create incompatible XML in the sense that they won't necessarily be able to understand each others tags, or display them in an identical fashion.
The real advantage of XML and, SGML as well, is in large scale document publishing. The advantage of XML is that it decouples content creation from formatting. Think about how mch time is wasted deciding whether something should be bold or bold-italics etc... Especially since the formatting says nothing about the content. For example: Currently, a serial number in a document might be tagged as or the equivalent proprietary tag. In XML the serial number could be tagged as , which tells us what the data is not what it looks like. A DTD tells the writer what tags he can use and the DTD can be customized.
In parallel, some one else (myself) is writing style sheets for formatting the tags for display in a browser, or conversion to html, rtf, pdf, etc. But, the key is that now writers will be able to concentrate on content creation, while some one else deals with formatting. It also enforces consistent formatting across writers.
The other advantage of XML is that cusomized tags give information about the content that can make search engines much smarter.
To relate this back to proprietary file formats, XML is only cross-platform WYSIWYG when the DTD and style sheet is either sent with the document or known to all. A word processor can very easily use obfuscated tags, and an internal style sheet. Which while the file may be saved as XML, and the content is available to anyone. No one without the same program can accurately reproduce the document.
Dastardly
P.S. XML does not require a DTD, but does require a style sheet of some sort for accurate reproduction. A DTD is mostly useful for document creation.
You're joking, right? Have you actually gone to the CEO of any company and told him/her that it's time to shut down marketing? "They can just take the next 2 months off while we re-engineer the back end systems."
:-)
:-)
You'll be lucky if you're not fired outright. And if they listen to you, they are even crazier than you
Maybe that whould be the punishment for marketing incorrectly projecting demand. The only problem is that marketing tends to hae the CEOs ear more than IT. Therefore, IT gets blamed for only meeting marketings projections. But, that is showing my own prejudices.
I think one of the real problems is a lot of companies are not doing the research that will give them good capacity projections. I believe it was Schwab in the article that said they went from quarterly analysis of capacity vs demand to every couple weeks, plus they have plenty of excess capacity, which is required sice demand can spike in days if not hours, while adding capacity probably has lead times measured in weeks.
Dastardly
First, as others have pointed out 2.4 will be the stable release of the 2.3 development kernel. Second, as no one else has pointed out, part of the kernel release process includes a feature freeze. There is a point during the 2.3 development when Linus will stop accepting new features, and only be accepting bug fixes. Essentially, what Linus is saying is that the feature freeze will occur sooner with fewer new features to debug resulting in more stable, but less featureful(sp?) stable kernel releases. This is probably a good thing.
I fervently hope that this bug is used to repeatedly down the stock market and a few military computers. That would put the spotlight like nothing else on Microsoft's failings.
Yeah right. The only thing that would do is get those people arrested and put in jail for computer crimes. No one blames MS for opening a big hole for viruses in Word and Excel. Instead they go after the guy who wrote Melissa. Doesn't anyone think it is extremely stupid that there is the ability to write word processor viruses??
All that would happen if some one exploited this, is that they would be arrested (if they could be tracked down). MS would not be blamed, and probably Eeye and slashdot would be blamed for publicizing the security whole.
Dastardly
It would be ironic if high end K7 chipsets were delayed because they decided to invest a lot of effort getting Rambus to work, and then the RAM modules don't turn up. I think for the high end, with huge 2nd level caches and enormous bandwidth requirements Rambus may have the edge if the caches take the top off the latency problems, and AMD may have thought the same way. And who would have guessed that an Intel-sponsored technology could fail in the PC space?
Rambus currently has no edge as far as quoted numbers go. Rambus bandwidth numbers are 1.6GB/s at 800MHz. PC266 DDR-SDRAM bandwidth numbers are 2.1GB/s. The only advantage I have heard for Rambus is some ephemeral promise that it offers better future scalability (vaporware). Can anyone show any pre-2001 plans for Rambus to match PC266?
Dastardly
Yes, it does grow the only thing is that it doesn't grow that fast. Consider that a white dwarf (~1 solar mass) is about the size of the earth. A neuton star (~1.5-2.5 solar masses) is about 10 miles wide. A stellar black hole (>3 solar masses) has a Shwarzchild diameter of about 1 mile. Last I checked the largest galactic blackholes (millions [billions?] of solar masses) have diameters no bigger than the diameter of our solar system.
As for the gravitational effects at this distance, the Eta Carinae would have the same effect as before because its mass didn't change (as others have described). Case in point, if some how the Sun were instantaneously replaced with a black hole of identical mass, the only effect on the earth would be that it would become really cold and dark.
The lower limit for a stellar mass black hole is only about 3 solar masses. Eta Carinae is about 100 solar masses. This doesn't mean every star over 3 solar masses will become a black hole. It just means if what is left after the super(or hyper)nova is bigger than 3 solar masses, the remnant will collapse into a black hole.
Small problem with using software rendering is that you will be benchmarking the memory subsystem more than the CPU. A software renderer has to load textures from main memory to the CPU in order to map them onto polygons, this is a very memory intensive process, and will easily overwhelm any calculating speed improvements. A hardware renderer on a the fastest possible 3D card with reams of texture memory would be more indicative of calculating performance than a software renderer, at least for Q2.
Dastardly
I guess this is a performance benchmark, so waiting for everything to reach peak performance, then quoting the numbers has some validity. But, for a complete review why not push the machines even further? Wouldn't it be even more interesting to see how each OS/server combination handles an overload?
Sometimes I feel like a leech, I use all this opensource stuff and don't really give anything back.
Well, while you may not have given back anything yet. Users are part of the Open source process as well, even the ones that don't write any code. If you were to run into a bug, would you just sit there and pout. Certainly not, you would go on the net and see if anyone else had seen it. Being an intelligent user you would give a detailed description of your system, OS config, and what you were doing. Then, a couple things might happen. The bug has been seen before, and there is a fix out or on its way. The bug hasn't been seen, but a programmer type or five see your problem and work with you on isolating, fixing and/or working around the bug. Or, improbably, no one is interested in fixing it, but you decide you need it fixed so with the advice of some of these net people you fix the problem yourself. Either way you have contributed to the open source process, though not necessarily as a programmer, but as a tester and QA person.
It is kind of scary that having 10 years at a single job is now a liability instead of an asset. It used to be that corporations invested in their employees and in exchange wanted long term employment. Now they expect you to come in with a full skill set and then (ab)use you for a year or two, then cast you off. Luckily I am working at a place hat doesn't do that. It is just annoying that should I decide I want to do something different it will be harder because I actually like my job.
While guns cannot be blamed as a cause, they are a gating factor in the sense that without guns you can't go on a shooting rampage.
Of course even if guns could be completely eliminated (not likely) you are still left with a person with problems who wants to go out and shoot some one, but at least they don't have a gun. Instead they will come up with something else (bombs, poisons, knives, etc.) It is interesting the way people on both sides of gun control don't seem to realize how guns relate to the crime itself. The NRA says "Guns don't kill people, people kill people", but guns enable a person to shoot some one, and therefore is an enabling factor. While the gun control nuts figure that guns are the cause when they are really just an enabling factor that doesn't do anything to address the fact that a person actually wanted to kill another person. The solution is not simple, but simple solutions is what gets politicians elected.
I think the causes are an OK start, but terribly incomplete. I don't profess to know all the events, feelings, and influences that came together to cause these two kids to decide that shooting up the school was a good idea. The key though is that it wasn't one thing, it wasn't four things, it was probably more like 15 different things including the 4 described above, but not limited to those. The key though is that had any one of those things been missing this would not have happened.
Some might then say the solution is to eliminate one of the causes and that way this will never happen again. This is where gun control comes in. While guns don't cause a person to go out and mass murder. They are a gating factor in that without the guns the person could not act. But, this is too simple a solution, and still leaves us with an outcast, depressed, and anti-social person.
I also have a problem with blaming the violent movies, music, and video games for violence. It is hard to come up with any correlation let alone causation. If you say well there were 100 shootings where the killers liked Marilyn Manson, I will point out that there were 1 million other people who listened to Marilyn Manson and didn't shoot people. The only way to get correlation is if this number is greater than total killers vs the general population. And, that is still only correlation, there is no way to show causation. Listening to Marilyn Manson is more likely an effect just like going out and shooting people is an effect.
The point is there has to be a differentiation between things that correlate and things that might cause a person to be a killer. We also need to know which solutions remove causes and which are simply gating solutions that do nothing for the person.
There is absolutely no proof of this, and can't be unless a statistcally significant population of Dvorak users can be found. But, isn't there the possibility that while speed might not improve, repetetive stress injuries might be reduced by the Dvorak keyboard?
oh, so in order for the "bazaar" to be successful, first I have to most of the heavy lifting and then give the code away for free to the rest of the world.
Not exactly, but you should have something that compiles, runs, and does something useful. It doesn't necesarily have to have every feature, or even be completely bug free. If you don't have something that runs minimally it is real hard to spark enough interest to gain critical mass and have the open source project take off. That bare minimum should not be heavy lifting by any stretch of the imagination.
If the internet were not involved none of these cases of a 14 year old flying out to be with some man they think they love would make anything mor ethan the local news. There seems to be this inherent assumption that there are more perverts, rapists, child molesters, or whatever on the internet than there are in society at large. But, there is no good evidence one way or another that I have found that isn't anecdotal. So, here is my proposed challenge to the Slashdot effect.
Let's find a good study one or the shows one way or the other whether there is a greater or lower occurrence of perverts, rapists, child molesters, with internet ties than without. I bet any difference would be statistically insignificant.
The study also should compensate for the differences between the people who are on the on the internet vs not. For example: the is a minimum affluence level for hte internet where the person has to at least be able to afford a computer and internet connection.
One thing that needs to be brought up is that 40K in one place is different form 40K in another. For example: In Silicon Valley 2 bed, 2 bath apartments go for minimum 2K a month. (At least anywhere you woudl want to live.) So, you are guaranteeed to have a roommate, and still pay almost half your take home in rent.
In other places the same person might go for $800 a month, so with a roomamte that coes to only $400. In which case, 30K is OK.
So, don't jump to conclusions about what amount is reasonable.