[Companies] do not test Microsoft application and OS security much. They assume MS will simply take care of it.
To me, that sounds like companies want Microsoft to worry about the security, and they'll worry about the difficulty involved in regression testing and patch installation.
(although I don't exactly see how anyone thought it would be a good idea to have it be suggested by using apes)
Because another point of the movie is that nuclear war is bad. And it's a lot easier to believe that nuclear war would make apes intelligent enough (and big enough) to rule the planet (and speak English) than just about any other animal. They're also the easiest costumes for actors to dress up in. (And the ape costumes in the original were not at all easy.)
Also, the apes represent the oppressors in the movie. Which would equate them with White America, not Black America. So I don't see how anyone could be offended by claiming that the film tries to equate black people with monkeys, when quite the opposite is true.
Close. A guy at the company I work at was fired* because he was tying his shoes behind a woman who felt uncomfortable with that. (Although the general consensus is that her feelings were not unfounded. And his response when confronted by security did not help his case.)
* Technically he was not fired. He was a contractor, and was told by security that he was not allowed to come onto the premises.
A child predator was caught using a car to lure his victims. Parents and politicians are outraged that cars can be used for such nefarious purposes, and have promised to enact legislation to limit the use of cars. Some are calling for an out-right ban on cars.
Both of you have conducted a good debate, even though I think you've both missed a few things. Just one major point I'd like to add.
There are 4 considerations when looking at a storage medium (ignoring connectivity, reliability, and compatibility issues): capacity, speed (sustained and latency), size, and price. Look at how hard drives and flash RAM (and microchip memory in general) have progressed over the years. You've shown that flash RAM is already smaller in size than a comparable-capacity hard drive.
Hard drives have probably increased in capacity faster over the long term, but haven't gotten all that much faster. It's to the point where CPUs have to wait 1000s of cycles to get data from a hard drive. (And even RAM access is up into the 10s of cycles.) I think this will soon (next 3 years) drive a move to an additional memory type in the memory access scheme (CPU -> L1 cache -> L2 cache -> RAM -> HD). I think we'll have systems with flash RAM between the RAM and hard drives. Or maybe I'm wrong, and we'll just continue with even larger hard disk caches, or system RAM will get replaced with flash RAM.
Even capacity-wise, in recent years, typical flash RAM devices have moved from 64 MB to 1 GB in about 2 years, staying about the same price and size. That's a factor of 16, or a factor of 4 every year. (Maybe my memory [pun not intended] is off, and 64 MB was common 4 years ago -- that'd still be doubling every year.) Hard drives are doubling in capacity at a rate of probably more than 18 months. I think this will eventually lead to flash RAM surpassing hard disks in terms of capacity/price. At current trends, I'd predict this to happen in about 2013. That would put both technologies at about 6 TB for $200. (I should really get some real numbers and graph the trends to make a more accurate prediction.) However, I actually expect hard drive capacities to slow their rate of increase, so it could happen earlier.
One minor point I'd like to get back to is the low number of writes allowed to flash RAM devices. You mentioned that you couldn't write to the whole disk in 15 years. But the real problem is that certain portions of a hard drive get written to a lot more often than others. For example, the FAT tables on FAT filesystems probably get written to 100K times in less than a year. (Granted, NTFS and UNIX file systems are better, but they still have to make frequent changes to common branching points in the directory tree.) I'm not sure if there are algorithms to rotate these writes through different memory cells (perhaps you were alluding to that -- I didn't quite follow what you meant). But even if there are, I doubt that they're good enough to reduce writes enough for it not to be an issue with some types of access patterns. Do you have any good pointers to info that would correct my mis-understanding?;)
Most text editors treat an uninterrupted string of text characters as a single step when it comes to Undo. For example, let's say I type "How are you, Bob?" and cursor up, then type "Dear Bob". The first Undo would remove "Dear Bob". The second Undo would remove "How are you, Bob?". This is what most people expect, as we think in words and sentence fragments.
I don't think it makes sense to look at the trade balance for a category when trying to determine which category is the "most important". Otherwise, you'd conclude that we "lost money" on most of our trade. But there are a lot of industries that made a lot of money on exports in those categories. Should we "give up" on the categories we're "losing" in?
Use the links at the top of the page to go to the Goods page, and you'll see that manufactured goods, agricultural goods, and travel (on the services) page are larger than the royalties category. Hence, IP royalties are #4.
I've started using DTV on my Mac. It's basically Internet TV done mostly right, where anyone can create a "channel". I've found many good videocasts, some of them fiction. My favorite is Bush's Speech Writer. Turns out it's a commercial production by Comedy Central, but who cares? I don't know about audiocasts though. I'd suspect NPR would have some.
Get a (live-in) significant other that only speaks the language you want to learn. You'll learn fast, or die trying. Better yet, move in with her whole family.
if you allow anyone to use a trademark for any reason you risk loosing the trademark because it can become a generic term
Not true. There has to be a likelihood of confusion or deception. But your reason number 2 may be valid. It's a grey area though -- whether just because it's a gene associated with cancer is enough to turn the case in Nintendo's favor is questionable. The added "fame" of the Pokemon trademark might be enough to throw the case into their favor though.
That sounds like a good way to do it. Although if you're writing a script, I'd consider that to be equivalent to the command-line. Without the interactive testing ability.
NT is (was) a microkernel. How/why the fsck did any drivers get into kernel-land anyway? The whole point of a microkernel is to have all your drivers run in user-space.
Having the graphics subsystem outside of the kernel is not special to Unix.
On the contrary, Windows is perhaps the only OS that would have even considered putting the entire GUI in the kernel space.
And it is much better than say the linux kernel
Actually, in the 1980s and 1990s, it was believed by most that microkernels were superior to monolithic kernels. (See Linus' argument with Andrew Tannenbaum.) But that thinking has largely been shown to be false in the real world in the past decade or so. And with dynamic module loading, as in Linux, it becomes less of an issue. Not to mention that when you go and throw an entire GUI subsystem in kernel space, you're left with something almost, but not quite, entirely unlike a microkernel.
It's the only thing that could still make Windows crash, a driver crash.
That's ridiculous. (And it makes you sound like a Microsoft yes-boy.) Any code running in kernel space (or with full hardware access) can hard-crash your system. You can be pretty sure that even a microkernel has some bugs in it. And the number of bugs will be fairly proportional to the amount of code in the system. So moving drivers out of the kernel is probably a good thing, but is by no means guaranteed to prevent all hard system crashes. Not to mention the fact that the way the Win32 subsystem works, if you crash it, you've effectively killed the machine anyway.
I don't think most Windows users, even the "pros" realize how much resource is involved in something as simple as moving the mouse, moving a window around or resizing it.
Back in college, we had a VAX with 2 CPUs that had roughly 10-20 people logged in via VT220s at one time. The CPUs would typically be at a few percent. A few times, I was watching the CPU utilization while I ran the text editor. I found that if I held a key down for long enough, it would max out one of the CPUs. My guess is that the program was written poorly and had to loop through every character on the line each time a new character was added at the end.
Anyway, the point is that CPU usage can be surprising. There's no way a terminal running at 9600 baud should have been able to max out that CPU just with typing in a text editor. But it did.
Ah, yes. Just what we all want. Command-line administration of Active Directory and Exchange.
Acutally, you should want that. Let's say HR sends you a list of 100 new employees, along with user IDs. Do you really want to have to type each of those in (or cut and paste each one) and click on all the settings for each user? Maybe things have changed since I last had to admin Windows networks, but that's what you had to do. Assuming that a GUI is the Right Tool for all admin tasks is ridiculous. Especially when you end up with large batches of work.
Perhaps insurance does have an analogy here. Perhaps it's due to the fact that the more clients your system is serving, the more likely you'll need to make a call to their support staff. In a way, support is very much like insurance. Of course, this is assuming that they don't have separate pricing for support of the software.
I'd really like to find someone who could explain the pros and cons of Rails versus Nitro. I've looked at Nitro (and used Rails a bit) and I find Og (Nitro's SQL-to-model portion) to be quite a bit better than ActiveRecord (Rails' SQL-to-model portion). The only reason I'm not really considering using Nitro is the momentum/community that Rails has. For example, I was able to buy a book about Rails, and several more are on the way. I doubt that there are any Nitro books on the way any time soon. I'm also considering using Og with Rails.
Ah. I don't know how to tell what repository it comes from either. I'm just getting started with Debian, but I'm really digging it. (I've been using Linux since 1994.) But it does do some things a lot differently than other Linux distros. And Ubuntu looks great too - I can't wait to try it on a desktop soon.
As far as what to do next, I recommend the Pragmatic Programmers' Rails book. It's very well-written, and walks you through a moderate-sized project. I also found that I didn't really need to learn Ruby to get started. I quickly went through the appendix of the book to get the basics of Ruby, then started reading the book. I don't think I've missed too much. The code is easy enough to understand, as long as you're familiar with object-oriented programming concepts.
I'm not a big fan of Prototype either. (Although I have yet to find ANY good extensive JavaScript library that I do like.) But Rails doesn't require that you use Prototype. In fact, unless you use the AJAX-specific "tags", you don't need Prototype at all. (Frankly, I'm not a fan of the RHTML either -- I'd much rather use something like ZPT/TAL or Kid that web designers can pull into Dreamweaver.)
I'm using a different way to do my JavaScript. I'm a big fan of unobtrusive JavaScript. So I just assign specific classes to elements I want to have certain behaviors. Then the JavaScript code just looks for the specific classes, and applies the behaviors to those elements.
Slashdot Mirror
You kind of proved the point:
[Companies] do not test Microsoft application and OS security much. They assume MS will simply take care of it.
To me, that sounds like companies want Microsoft to worry about the security, and they'll worry about the difficulty involved in regression testing and patch installation.
(although I don't exactly see how anyone thought it would be a good idea to have it be suggested by using apes)
Because another point of the movie is that nuclear war is bad. And it's a lot easier to believe that nuclear war would make apes intelligent enough (and big enough) to rule the planet (and speak English) than just about any other animal. They're also the easiest costumes for actors to dress up in. (And the ape costumes in the original were not at all easy.)
Also, the apes represent the oppressors in the movie. Which would equate them with White America, not Black America. So I don't see how anyone could be offended by claiming that the film tries to equate black people with monkeys, when quite the opposite is true.
Close. A guy at the company I work at was fired* because he was tying his shoes behind a woman who felt uncomfortable with that. (Although the general consensus is that her feelings were not unfounded. And his response when confronted by security did not help his case.)
* Technically he was not fired. He was a contractor, and was told by security that he was not allowed to come onto the premises.
He didn't say that black soldiers didn't experience racism. He just said that the French wanted to see their black cocks, not their tails.
(I hope moderators read the context 3 levels up, or this is going to look WAY off-topic!)
A child predator was caught using a car to lure his victims. Parents and politicians are outraged that cars can be used for such nefarious purposes, and have promised to enact legislation to limit the use of cars. Some are calling for an out-right ban on cars.
Looks like I'm a slow reader. Here's a link to Evelyn Wood's.
OK, I found that you've actually done all the trend analysis on your web site. I'd like to retract my 4th paragraph. ;)
Both of you have conducted a good debate, even though I think you've both missed a few things. Just one major point I'd like to add.
;)
There are 4 considerations when looking at a storage medium (ignoring connectivity, reliability, and compatibility issues): capacity, speed (sustained and latency), size, and price. Look at how hard drives and flash RAM (and microchip memory in general) have progressed over the years. You've shown that flash RAM is already smaller in size than a comparable-capacity hard drive.
Hard drives have probably increased in capacity faster over the long term, but haven't gotten all that much faster. It's to the point where CPUs have to wait 1000s of cycles to get data from a hard drive. (And even RAM access is up into the 10s of cycles.) I think this will soon (next 3 years) drive a move to an additional memory type in the memory access scheme (CPU -> L1 cache -> L2 cache -> RAM -> HD). I think we'll have systems with flash RAM between the RAM and hard drives. Or maybe I'm wrong, and we'll just continue with even larger hard disk caches, or system RAM will get replaced with flash RAM.
Even capacity-wise, in recent years, typical flash RAM devices have moved from 64 MB to 1 GB in about 2 years, staying about the same price and size. That's a factor of 16, or a factor of 4 every year. (Maybe my memory [pun not intended] is off, and 64 MB was common 4 years ago -- that'd still be doubling every year.) Hard drives are doubling in capacity at a rate of probably more than 18 months. I think this will eventually lead to flash RAM surpassing hard disks in terms of capacity/price. At current trends, I'd predict this to happen in about 2013. That would put both technologies at about 6 TB for $200. (I should really get some real numbers and graph the trends to make a more accurate prediction.) However, I actually expect hard drive capacities to slow their rate of increase, so it could happen earlier.
One minor point I'd like to get back to is the low number of writes allowed to flash RAM devices. You mentioned that you couldn't write to the whole disk in 15 years. But the real problem is that certain portions of a hard drive get written to a lot more often than others. For example, the FAT tables on FAT filesystems probably get written to 100K times in less than a year. (Granted, NTFS and UNIX file systems are better, but they still have to make frequent changes to common branching points in the directory tree.) I'm not sure if there are algorithms to rotate these writes through different memory cells (perhaps you were alluding to that -- I didn't quite follow what you meant). But even if there are, I doubt that they're good enough to reduce writes enough for it not to be an issue with some types of access patterns. Do you have any good pointers to info that would correct my mis-understanding?
Most text editors treat an uninterrupted string of text characters as a single step when it comes to Undo. For example, let's say I type "How are you, Bob?" and cursor up, then type "Dear Bob". The first Undo would remove "Dear Bob". The second Undo would remove "How are you, Bob?". This is what most people expect, as we think in words and sentence fragments.
Here's the Goods page.
I don't think it makes sense to look at the trade balance for a category when trying to determine which category is the "most important". Otherwise, you'd conclude that we "lost money" on most of our trade. But there are a lot of industries that made a lot of money on exports in those categories. Should we "give up" on the categories we're "losing" in?
Use the links at the top of the page to go to the Goods page, and you'll see that manufactured goods, agricultural goods, and travel (on the services) page are larger than the royalties category. Hence, IP royalties are #4.
I've started using DTV on my Mac. It's basically Internet TV done mostly right, where anyone can create a "channel". I've found many good videocasts, some of them fiction. My favorite is Bush's Speech Writer. Turns out it's a commercial production by Comedy Central, but who cares? I don't know about audiocasts though. I'd suspect NPR would have some.
Can I buy stock in your "Happy Fun Ball" company (Wacky Products Incorporated)?
Um, any shareholder can typically file a petition to have a proposal put to a stockholder ballot. It's apparently known as shareholder activism.
Get a (live-in) significant other that only speaks the language you want to learn. You'll learn fast, or die trying. Better yet, move in with her whole family.
if you allow anyone to use a trademark for any reason you risk loosing the trademark because it can become a generic term
Not true. There has to be a likelihood of confusion or deception. But your reason number 2 may be valid. It's a grey area though -- whether just because it's a gene associated with cancer is enough to turn the case in Nintendo's favor is questionable. The added "fame" of the Pokemon trademark might be enough to throw the case into their favor though.
That sounds like a good way to do it. Although if you're writing a script, I'd consider that to be equivalent to the command-line. Without the interactive testing ability.
they're also moving all the drivers in userland
NT is (was) a microkernel. How/why the fsck did any drivers get into kernel-land anyway? The whole point of a microkernel is to have all your drivers run in user-space.
Having the graphics subsystem outside of the kernel is not special to Unix.
On the contrary, Windows is perhaps the only OS that would have even considered putting the entire GUI in the kernel space.
And it is much better than say the linux kernel
Actually, in the 1980s and 1990s, it was believed by most that microkernels were superior to monolithic kernels. (See Linus' argument with Andrew Tannenbaum.) But that thinking has largely been shown to be false in the real world in the past decade or so. And with dynamic module loading, as in Linux, it becomes less of an issue. Not to mention that when you go and throw an entire GUI subsystem in kernel space, you're left with something almost, but not quite, entirely unlike a microkernel.
It's the only thing that could still make Windows crash, a driver crash.
That's ridiculous. (And it makes you sound like a Microsoft yes-boy.) Any code running in kernel space (or with full hardware access) can hard-crash your system. You can be pretty sure that even a microkernel has some bugs in it. And the number of bugs will be fairly proportional to the amount of code in the system. So moving drivers out of the kernel is probably a good thing, but is by no means guaranteed to prevent all hard system crashes. Not to mention the fact that the way the Win32 subsystem works, if you crash it, you've effectively killed the machine anyway.
I don't think most Windows users, even the "pros" realize how much resource is involved in something as simple as moving the mouse, moving a window around or resizing it.
Back in college, we had a VAX with 2 CPUs that had roughly 10-20 people logged in via VT220s at one time. The CPUs would typically be at a few percent. A few times, I was watching the CPU utilization while I ran the text editor. I found that if I held a key down for long enough, it would max out one of the CPUs. My guess is that the program was written poorly and had to loop through every character on the line each time a new character was added at the end.
Anyway, the point is that CPU usage can be surprising. There's no way a terminal running at 9600 baud should have been able to max out that CPU just with typing in a text editor. But it did.
Ah, yes. Just what we all want. Command-line administration of Active Directory and Exchange.
Acutally, you should want that. Let's say HR sends you a list of 100 new employees, along with user IDs. Do you really want to have to type each of those in (or cut and paste each one) and click on all the settings for each user? Maybe things have changed since I last had to admin Windows networks, but that's what you had to do. Assuming that a GUI is the Right Tool for all admin tasks is ridiculous. Especially when you end up with large batches of work.
Perhaps insurance does have an analogy here. Perhaps it's due to the fact that the more clients your system is serving, the more likely you'll need to make a call to their support staff. In a way, support is very much like insurance. Of course, this is assuming that they don't have separate pricing for support of the software.
I'd really like to find someone who could explain the pros and cons of Rails versus Nitro. I've looked at Nitro (and used Rails a bit) and I find Og (Nitro's SQL-to-model portion) to be quite a bit better than ActiveRecord (Rails' SQL-to-model portion). The only reason I'm not really considering using Nitro is the momentum/community that Rails has. For example, I was able to buy a book about Rails, and several more are on the way. I doubt that there are any Nitro books on the way any time soon. I'm also considering using Og with Rails.
Ah. I don't know how to tell what repository it comes from either. I'm just getting started with Debian, but I'm really digging it. (I've been using Linux since 1994.) But it does do some things a lot differently than other Linux distros. And Ubuntu looks great too - I can't wait to try it on a desktop soon.
As far as what to do next, I recommend the Pragmatic Programmers' Rails book. It's very well-written, and walks you through a moderate-sized project. I also found that I didn't really need to learn Ruby to get started. I quickly went through the appendix of the book to get the basics of Ruby, then started reading the book. I don't think I've missed too much. The code is easy enough to understand, as long as you're familiar with object-oriented programming concepts.
Honestly, I didn't find that I needed to learn much Ruby to get started using Rails.
I'm not a big fan of Prototype either. (Although I have yet to find ANY good extensive JavaScript library that I do like.) But Rails doesn't require that you use Prototype. In fact, unless you use the AJAX-specific "tags", you don't need Prototype at all. (Frankly, I'm not a fan of the RHTML either -- I'd much rather use something like ZPT/TAL or Kid that web designers can pull into Dreamweaver.)
I'm using a different way to do my JavaScript. I'm a big fan of unobtrusive JavaScript. So I just assign specific classes to elements I want to have certain behaviors. Then the JavaScript code just looks for the specific classes, and applies the behaviors to those elements.