This may go against the grain here, but I replace my desktop drive about every 12-18 months. As I see it, here are the benefits of doing so:
+1) The drive still has decent resale value at that point, particularly if you sell on a computer forum and not on ebay. This helps reduce the cost of the hard drive update.
+2) Drive capacities are increasing quickly while costs continue to decline. This reduces the cost of the upgrade.
+3) Replacing before the warranty period is up means that the likelihood of experiencing a hard drive failure is low.
+4) While WinXP is a lot better than Win9x, it still doesn't hurt to do a fresh reinstall every 12-18 months. A hard disk replacement is the perfect timing for this.
Of course, there are some valid counter-arguments to these points:
-1) Security. (i.e., somebody could recover your private data.) I run Darik's Boot 'n' Nuke a few times, so I'm not terribly concerned about this. After running such a program, the odds of somebody successfully recovering data on a home budget are pretty low.
-2) You may be replacing too often. Well, I can't do much about that. But good drives don't cost much more than $100-$150 these days. A little peace of mind is worth something, and the regular size/speed upgrades are a nice bonus.
-3) This is no substitute for backups. I completely agree, and make backups of my most critical data to remote servers.
-4) Perhaps this isn't necessary. Perhaps not, but a fresh format is a helpful after 18 months.
Any way around it, I acknowledge that this strategy is a bit more expensive than may be necessary, but it has served me well in the past six years +. I've only had one drive fail in the past, back when I let my drives go well beyond the warranty period. Of course, that drive was a total loss, with no recovery of value to apply to the new drive, and there were some non-recoverable files. In my opinion, preventing problems before they occur is preferable, and getting speed and capacity boosts are just icing on the cake. -- Paul
Strangely enough, I had trouble searching for C++.
For example, C++ jpeg gets turned into C jpeg, and returns a bunch of C code. If you search with quotes, "C++" jpeg, you get "C " jpeg. Search for "devc++", and you get "devc " and information on Devcon international.
This doesn't make the search engine particularly useful for C++ coders.;) -- Paul
The biophysical mechanism in the article makes sense. Reactive oxygen species (i.e., the "free radicals", like superoxide and hydrogen peroxide) are generated in the mitochondria during metabolism. The bulk of the ROS's are scavenged before they can cause harm, but those that aren't neutralized can damage biological molecules (e.g., proteins, lipids, and DNA). On a slightly unrelated note that might make this a bit more concrete, UV rays generate ROS's when they strike biological moleculues in the skin. These ROS's are responsible for a lot of the damage from UV rays, both to collagen (wrinkles) and DNA (carcinogenesis). So there's a very concrete example of free radicals contributing to aging that you can actually see.
The article claims that the "thermostat" is being tricked to reduce metabolism, so that would, in turn, decrease the generation of ROS's. I also wonder if ROS production is temperature-dependent. (Are the molecules less energetic, to the point that it's more difficult to form ROS's?) So, it really does make sense.
The whole SP thing is a throwback to the bad old days of 28.8k modems and CDs by post. Now we can add the fixes as they come along so why bother with a monolithic chunk of code that must be a testing nightmare for MS as well as corporate end users?
Another good aspect of having SP's is that it's a lot easier to say that "Our software requires WinXP SP2 or later" than "Our software requires WinXP with all security updates through at least kb1048482393". Likewise, having fixed "snapshots" in the form of SP's allows a way to specify a known compatible combination of security and feature upgrades. (Tested with WinXP SP2, rather than a selective mix of updates including perhaps some but not all of the updates from SP2.) -- Paul
That's a nice way to intuit it, and it's preceisely what's happening for first-order methods.
In mathy terms, the second-order terms in du/dt are important, the problem is "stiff", and so you have to use a higher-order method to do better. You're right that a first-order method will screw this up regardless of the discretization size, and in fact, it's not unique to astrophysical problems.;-) -- Paul
Write a program using standard Newtonian physics that takes as its input a cluster of 'stars' of various masses.
You'll have to be careful that using good numerics before you conclude too much from that program. If you use simple forward Euler differencing
du/dt = ( u(t+dt)-u(t) )/dt
+ O(dt)
for even a simple two-body system using the inverse-square law, the orbiting object will spiral outward due to accumulated discretization error. It's a good lesson in why you should use something higher-order, such as a good Runge-Kutta discretization.;-) -- Paul
Re:The problem with wireless devices...
on
USB To Go Wireless
·
· Score: 1
On my laptop I use a bluetooth mouse to avoud the hassles of dealing with plugging and unplugging the mouse from the laptop.
And you can solve this problem by leaving your USB mouse and keyboard plugged into a hub on your desktop, so you only have one thing to plug in. As a bonus, you could get a USB-powered laptop cooler with a built-in hub (like this one), so you're not only cleaning up cable clutter, but also reducing heat-related stress on your hard drive. -- Paul
The mechanisms of wound healing aren't too far off from embryonic development. (In fact, many people study angiogenesis by observing embryos.) IIRC, there have been studies on the effects of weightlessness on developing mouse embryos, as gestation time is so short. So, you might look in that direction for your answers. You definitely asked an interesting question!! -- Paul
Besides the fact that dubbing DRM with the propogandistic term "digital restrictions management" doesn't look particularly professional, it seems a bit inviting for people to weasal out of the conditions. "No, we don't use `digital restrictions management' anywhere in this code; only `digital rights management.'"
Similarly, by not referring to DRM by its (current) name, but rather by what it does, I think it helps to future-proof the new GPL. Suppose they kept the explicit term DRM in the language of the license. Then when the next name for DRM comes out, who knows how well the language of the license would work? -- Paul
IANAL, but I suspect that the intent of giving that number was not malicious.
That's a really good point! (IANAL also.) But it's definitely worthwhile being careful in print... Thanks again for bringing up a really good point. -- Paul
The slashdot summary is not only inaccurate, but libelous. By the article, he pleaded guilty to one $542,000 grant. So, he's only been found legally accountable for that amount, not the $2.9 million claimed by the prosecution:
In an agreement with prosecutors, he pleaded guilty in connection with one $542,000 grant; the government said he defrauded federal agencies out of $2.9 million.
This LA Times article from the weekend has a more in-depth look at the grueling process of Chinese university entrance exams, and shows a bit more of the motivation to go to such lengths to cheat.
For example:
hinese college admissions officers don't look at your high school grades, personal interviews, recommendations or essays in making their decisions. They don't make allowances if you don't test well. They won't even cut you slack if your mother died the day before. Everything, countless years of sacrifice and hard work, boils down to this one test. Those who perform miserably have to wait another year to take the exam.
Not a great system from any point of view. Encourages cheating. Discourages creativity, not particularly fair to the students.... -- Paul
Interesting post. I noticed that the "Variable numbers of repeats" looks a lot like microsatellite instability (MSI), where extra copies of short sequences get inserted. Like ATATAT becomes ATATATAT. I'm not entirely up-to-date on what happens as a result, but certainly the encoded protein may lose or alter its function. This type of genetic instability is characteristic of certain types of cancer, like certain colorectal cancers, particularly when a mutation knocks out a crucial repair pathway gene.
Do you have any insight if these truly are similar, or just look that way? (I'm also pretty new to the genetic aspects of cancer, and I'm trying to learn all I can.) Also, do you have any good web references that you could recommend for further learning on these topics? It would be very helpful for this mathematical biologist.:-)
As I understand it, the FDA regulates products sold, not research (aside from clinical trials). You can research any amazing product you want, and foreign companies can manufacture any amazing product they want. But both will run into the FDA at the point where you sell the product in the U.S. No sooner, and no later, same barrier for both. (And only if the FDA applies to that particular product category.)
At this point, you can buy previously top-of-the-line AGP cards for budget prices already. (I know, not current cards, but already plenty fine.) Excellent deals on ebay, computer forums (e.g., www.ocforums.com), etc. An ATi 9800 pro fetches no more than $65 currently, and 6600GT cards are generally under $160. I get excellent Half-Life2 performance with my old GF4 Ti4200 card, and those can be found for $30 or less online!
The budget cards for AGP are the top-of-the-line cards of 1-3 years ago, and they're still perfectly good cards. -- Paul
Part of the problem may be the difference in lifetimes between mice and humans, as well as problems in detecting small tumors.
Anti-angiogenic therapy leads to a hypoxic tumor microenvironment (the tissue surrounding the tumor), which can, in turn, lead a tumor to fragment into smaller tumors. (This has been predicted in mathematical/computer models and verified in some experiments and clinical evidence.)
In a mouse, those small tumors may not have time to grow large enough to detect, whereas in a human, those fragments have more time to do so, leading to recurrence. Or the small tumors may preferentially grow away from the low-O2/low-glucose region to invade nearby tissues.
Other, slow time-scale interactions may also not come into play for short mouse lifespans but may be important on human lifespans.
Of course, the genetic differences are there, too. The problems with the mouse model have always been interesting. -- Paul
True, although locally (at a tissue level), cancer is very much like natural selection: the failure to respond to inhibitory signals (via apoptosis), as well as uncontrolled replication give the cancerous cells a relative survival advantage or fitness over normal cells. In fact, this is how many cellular automata models model the spread of mutations through a tissue, even in tissues that maintain a constant cell population.
Natural selection is particularly useful in understanding resistence to chemo treatments, hypoxia, etc.
If you want, I can refer you to some interesting modeling links on just this very approach, or point you towards some animations of mine on a simple C.A. model. -- Paul
Hmmm, it's an interesting idea, but it seems to give a lot of false positives. (So naturally, it will detect fake papers, if it thinks every paper is fake.)
First thing I tried was some pages on computational oncology website, in particular, my cancer primer, which I wrote in not a short time. Everything I fed was determined to be inauthentic. Perhaps I just write like a robot.:-) I figured that perhaps the detector was more primed for real papers, so I figured it wasn't too big of a deal.
So, next I tried my most recent research paper, and it, too, was determined to be inauthentic, and in fact with less authenticity than my website. So much for the theory of being primed for scientific papers only. This thing is starting to look pretty bogus to me... but an interesting idea, nonetheless. -- Paul
This sounds like a great idea, but I can see some
legitimate causes being harmed. For instance,
Untied.com is a
typo of United, which is used to protest some
labor practices at United Airlines.
I guess the question is, how is MS going to determine
the legitimate misspellings from the illegitimate
misspellings? Certainly United doesn't like the
misspelling above, but it's not anti-consumer like
misspelling a company name and winding up at a
spam site, or worse yet, a phishing site. -- Paul
Personally, I'd be just as concerned about industrial espionage. Nothing like using a well-entrenched business brand to phone home biomedical engineering research... -- Paul
Very good points. I'd definitely like to read the paper in greater detail. The physical assumptions will make or break a simulation, and it's always good to know the limitations of any observed results. I'll be very interested to see if future work confirms these observations...
I definitely appreciate reading your point of view, so thanks for your post. -- Paul
This is just fascinating, and precisely why we do high performance scientific computing. This quote piqued my interest in particular:
The model also shows that the virus coat collapses without its genetic material. This suggests that, when reproducing, the virus builds its coat around the genetic material rather than inserting the genetic material into a complete coat. "We saw something that is truly revolutionary," Schulten says.
So, by doing this simulation of a tiny span of time, the team was able to get new insight into the process of viral replication that would be extremely difficult to come by with experimental techniques. It also is fascinating, since we often think of viruses as little static particles that float around until they interact with a cell, and yet the simulation showed the surface pulsing. Very cool! -- Paul
The next question: Why are we starting with mice if we can't always use promising developments on humans? Wouldn't that be a huge waste of initial effort and expectations?
That's a great question. In part, it's a matter of ethics: you can't try out new ideas on human beings. Also, mice breed and grow quickly, which makes them faster to try new ideas on. But as stated, they aren't a great predictor. Another interesting thought (and one I don't have much insight on) is that perhaps some ideas that don't work out for mice might actually work in humans but are prematurely rejected. (i.e., if false positives are possible, shouldn't false negatives also be possible?)
This touches on my work, in part; I'm a mathematician working on increasingly detailed computer models of cancer to see if we can eventually get a better and faster model than the mouse model. It's also a lot easier to control the experimental conditions on a computer.:)
If you're interested in these kinds of questions, I'd recommend also checking out some BusinessWeek articles from about a year ago, where they talked about the state of cancer research. Their conclusion was that the biggest roadblock today is the mouse model. I don't remember the exact citation, but I could dig for it if you are interested. -- Paul
Slashdot Mirror
This may go against the grain here, but I replace my desktop drive about every 12-18 months. As I see it, here are the benefits of doing so: +1) The drive still has decent resale value at that point, particularly if you sell on a computer forum and not on ebay. This helps reduce the cost of the hard drive update. +2) Drive capacities are increasing quickly while costs continue to decline. This reduces the cost of the upgrade. +3) Replacing before the warranty period is up means that the likelihood of experiencing a hard drive failure is low. +4) While WinXP is a lot better than Win9x, it still doesn't hurt to do a fresh reinstall every 12-18 months. A hard disk replacement is the perfect timing for this. Of course, there are some valid counter-arguments to these points: -1) Security. (i.e., somebody could recover your private data.) I run Darik's Boot 'n' Nuke a few times, so I'm not terribly concerned about this. After running such a program, the odds of somebody successfully recovering data on a home budget are pretty low. -2) You may be replacing too often. Well, I can't do much about that. But good drives don't cost much more than $100-$150 these days. A little peace of mind is worth something, and the regular size/speed upgrades are a nice bonus. -3) This is no substitute for backups. I completely agree, and make backups of my most critical data to remote servers. -4) Perhaps this isn't necessary. Perhaps not, but a fresh format is a helpful after 18 months. Any way around it, I acknowledge that this strategy is a bit more expensive than may be necessary, but it has served me well in the past six years +. I've only had one drive fail in the past, back when I let my drives go well beyond the warranty period. Of course, that drive was a total loss, with no recovery of value to apply to the new drive, and there were some non-recoverable files. In my opinion, preventing problems before they occur is preferable, and getting speed and capacity boosts are just icing on the cake. -- Paul
Strangely enough, I had trouble searching for C++.
For example, C++ jpeg gets turned into C jpeg, and returns a bunch of C code. If you search with quotes, "C++" jpeg, you get "C " jpeg. Search for "devc++", and you get "devc " and information on Devcon international.
This doesn't make the search engine particularly useful for C++ coders. ;) -- Paul
The biophysical mechanism in the article makes sense. Reactive oxygen species (i.e., the "free radicals", like superoxide and hydrogen peroxide) are generated in the mitochondria during metabolism. The bulk of the ROS's are scavenged before they can cause harm, but those that aren't neutralized can damage biological molecules (e.g., proteins, lipids, and DNA). On a slightly unrelated note that might make this a bit more concrete, UV rays generate ROS's when they strike biological moleculues in the skin. These ROS's are responsible for a lot of the damage from UV rays, both to collagen (wrinkles) and DNA (carcinogenesis). So there's a very concrete example of free radicals contributing to aging that you can actually see.
The article claims that the "thermostat" is being tricked to reduce metabolism, so that would, in turn, decrease the generation of ROS's. I also wonder if ROS production is temperature-dependent. (Are the molecules less energetic, to the point that it's more difficult to form ROS's?) So, it really does make sense.
Interesting, stuff, though. -- Paul
Another good aspect of having SP's is that it's a lot easier to say that "Our software requires WinXP SP2 or later" than "Our software requires WinXP with all security updates through at least kb1048482393". Likewise, having fixed "snapshots" in the form of SP's allows a way to specify a known compatible combination of security and feature upgrades. (Tested with WinXP SP2, rather than a selective mix of updates including perhaps some but not all of the updates from SP2.) -- Paul
That's a nice way to intuit it, and it's preceisely what's happening for first-order methods.
In mathy terms, the second-order terms in du/dt are important, the problem is "stiff", and so you have to use a higher-order method to do better. You're right that a first-order method will screw this up regardless of the discretization size, and in fact, it's not unique to astrophysical problems. ;-) -- Paul
PS: I should have mentioned an example when I wrote this post. See problem 2 in this homework solution from a numerical analysis class. -- Paul
You'll have to be careful that using good numerics before you conclude too much from that program. If you use simple forward Euler differencing
du/dt = ( u(t+dt)-u(t) )/dt + O(dt)
for even a simple two-body system using the inverse-square law, the orbiting object will spiral outward due to accumulated discretization error. It's a good lesson in why you should use something higher-order, such as a good Runge-Kutta discretization. ;-) -- Paul
And you can solve this problem by leaving your USB mouse and keyboard plugged into a hub on your desktop, so you only have one thing to plug in. As a bonus, you could get a USB-powered laptop cooler with a built-in hub (like this one), so you're not only cleaning up cable clutter, but also reducing heat-related stress on your hard drive. -- Paul
The mechanisms of wound healing aren't too far off from embryonic development. (In fact, many people study angiogenesis by observing embryos.) IIRC, there have been studies on the effects of weightlessness on developing mouse embryos, as gestation time is so short. So, you might look in that direction for your answers. You definitely asked an interesting question!! -- Paul
Send me a note when you get it up. I may be able to work a cite into my dissertation later this fall. :-) -- Paul
Agreed!
Besides the fact that dubbing DRM with the propogandistic term "digital restrictions management" doesn't look particularly professional, it seems a bit inviting for people to weasal out of the conditions. "No, we don't use `digital restrictions management' anywhere in this code; only `digital rights management.'"
Similarly, by not referring to DRM by its (current) name, but rather by what it does, I think it helps to future-proof the new GPL. Suppose they kept the explicit term DRM in the language of the license. Then when the next name for DRM comes out, who knows how well the language of the license would work? -- Paul
IANAL, but I suspect that the intent of giving that number was not malicious.
That's a really good point! (IANAL also.) But it's definitely worthwhile being careful in print ... Thanks again for bringing up a really good point. -- Paul
The slashdot summary is not only inaccurate, but libelous. By the article, he pleaded guilty to one $542,000 grant. So, he's only been found legally accountable for that amount, not the $2.9 million claimed by the prosecution:
In an agreement with prosecutors, he pleaded guilty in connection with one $542,000 grant; the government said he defrauded federal agencies out of $2.9 million.
This LA Times article from the weekend has a more in-depth look at the grueling process of Chinese university entrance exams, and shows a bit more of the motivation to go to such lengths to cheat.
For example:
hinese college admissions officers don't look at your high school grades, personal interviews, recommendations or essays in making their decisions. They don't make allowances if you don't test well. They won't even cut you slack if your mother died the day before. Everything, countless years of sacrifice and hard work, boils down to this one test. Those who perform miserably have to wait another year to take the exam.
Not a great system from any point of view. Encourages cheating. Discourages creativity, not particularly fair to the students .... -- Paul
Interesting post. I noticed that the "Variable numbers of repeats" looks a lot like microsatellite instability (MSI), where extra copies of short sequences get inserted. Like ATATAT becomes ATATATAT. I'm not entirely up-to-date on what happens as a result, but certainly the encoded protein may lose or alter its function. This type of genetic instability is characteristic of certain types of cancer, like certain colorectal cancers, particularly when a mutation knocks out a crucial repair pathway gene.
Do you have any insight if these truly are similar, or just look that way? (I'm also pretty new to the genetic aspects of cancer, and I'm trying to learn all I can.) Also, do you have any good web references that you could recommend for further learning on these topics? It would be very helpful for this mathematical biologist. :-)
Thanks again for an interesting post! -- Paul
As I understand it, the FDA regulates products sold, not research (aside from clinical trials). You can research any amazing product you want, and foreign companies can manufacture any amazing product they want. But both will run into the FDA at the point where you sell the product in the U.S. No sooner, and no later, same barrier for both. (And only if the FDA applies to that particular product category.)
In a word, that's a bogus argument. -- Paul
Where are the AGP versions of these cards?
At this point, you can buy previously top-of-the-line AGP cards for budget prices already. (I know, not current cards, but already plenty fine.) Excellent deals on ebay, computer forums (e.g., www.ocforums.com), etc. An ATi 9800 pro fetches no more than $65 currently, and 6600GT cards are generally under $160. I get excellent Half-Life2 performance with my old GF4 Ti4200 card, and those can be found for $30 or less online!
The budget cards for AGP are the top-of-the-line cards of 1-3 years ago, and they're still perfectly good cards. -- Paul
Part of the problem may be the difference in lifetimes between mice and humans, as well as problems in detecting small tumors.
Anti-angiogenic therapy leads to a hypoxic tumor microenvironment (the tissue surrounding the tumor), which can, in turn, lead a tumor to fragment into smaller tumors. (This has been predicted in mathematical/computer models and verified in some experiments and clinical evidence.)
In a mouse, those small tumors may not have time to grow large enough to detect, whereas in a human, those fragments have more time to do so, leading to recurrence. Or the small tumors may preferentially grow away from the low-O2/low-glucose region to invade nearby tissues.
Other, slow time-scale interactions may also not come into play for short mouse lifespans but may be important on human lifespans.
Of course, the genetic differences are there, too. The problems with the mouse model have always been interesting. -- Paul
Cancer is not a by-product of evolution
True, although locally (at a tissue level), cancer is very much like natural selection: the failure to respond to inhibitory signals (via apoptosis), as well as uncontrolled replication give the cancerous cells a relative survival advantage or fitness over normal cells. In fact, this is how many cellular automata models model the spread of mutations through a tissue, even in tissues that maintain a constant cell population.
Natural selection is particularly useful in understanding resistence to chemo treatments, hypoxia, etc.
If you want, I can refer you to some interesting modeling links on just this very approach, or point you towards some animations of mine on a simple C.A. model. -- Paul
Hmmm, it's an interesting idea, but it seems to give a lot of false positives. (So naturally, it will detect fake papers, if it thinks every paper is fake.)
First thing I tried was some pages on computational oncology website, in particular, my cancer primer, which I wrote in not a short time. Everything I fed was determined to be inauthentic. Perhaps I just write like a robot. :-) I figured that perhaps the detector was more primed for real papers, so I figured it wasn't too big of a deal.
So, next I tried my most recent research paper, and it, too, was determined to be inauthentic, and in fact with less authenticity than my website. So much for the theory of being primed for scientific papers only. This thing is starting to look pretty bogus to me ... but an interesting idea, nonetheless. -- Paul
This sounds like a great idea, but I can see some legitimate causes being harmed. For instance, Untied.com is a typo of United, which is used to protest some labor practices at United Airlines.
I guess the question is, how is MS going to determine the legitimate misspellings from the illegitimate misspellings? Certainly United doesn't like the misspelling above, but it's not anti-consumer like misspelling a company name and winding up at a spam site, or worse yet, a phishing site. -- Paul
Personally, I'd be just as concerned about industrial espionage. Nothing like using a well-entrenched business brand to phone home biomedical engineering research ... -- Paul
Very good points. I'd definitely like to read the paper in greater detail. The physical assumptions will make or break a simulation, and it's always good to know the limitations of any observed results. I'll be very interested to see if future work confirms these observations...
I definitely appreciate reading your point of view, so thanks for your post. -- Paul
This is just fascinating, and precisely why we do high performance scientific computing. This quote piqued my interest in particular:
The model also shows that the virus coat collapses without its genetic material. This suggests that, when reproducing, the virus builds its coat around the genetic material rather than inserting the genetic material into a complete coat. "We saw something that is truly revolutionary," Schulten says.
So, by doing this simulation of a tiny span of time, the team was able to get new insight into the process of viral replication that would be extremely difficult to come by with experimental techniques. It also is fascinating, since we often think of viruses as little static particles that float around until they interact with a cell, and yet the simulation showed the surface pulsing. Very cool! -- Paul
The next question: Why are we starting with mice if we can't always use promising developments on humans? Wouldn't that be a huge waste of initial effort and expectations?
That's a great question. In part, it's a matter of ethics: you can't try out new ideas on human beings. Also, mice breed and grow quickly, which makes them faster to try new ideas on. But as stated, they aren't a great predictor. Another interesting thought (and one I don't have much insight on) is that perhaps some ideas that don't work out for mice might actually work in humans but are prematurely rejected. (i.e., if false positives are possible, shouldn't false negatives also be possible?)
This touches on my work, in part; I'm a mathematician working on increasingly detailed computer models of cancer to see if we can eventually get a better and faster model than the mouse model. It's also a lot easier to control the experimental conditions on a computer. :)
If you're interested in these kinds of questions, I'd recommend also checking out some BusinessWeek articles from about a year ago, where they talked about the state of cancer research. Their conclusion was that the biggest roadblock today is the mouse model. I don't remember the exact citation, but I could dig for it if you are interested. -- Paul