The facts in general are contended, but it is clear that Juan's regime did not last much past Eva's death (although he did return later), and that his government was heavily reliant on the personalities and personal popularity of the leaders. This contrasts with the Roman Empire, which lasted through a number of leaders with a single continuing government,
Right, like your boss is going to let you waltz in at 10:30 because your brain-wave monitor didn't deem you rested until after 9. (And if you've had enough rest, don't you wake up naturally? Why use a device to do it for you?)
Sure, why not. You just have to stay late. Why pay people to sit around half-asleep in the morning? For that matter, the office will probably be half empty until noon Toronto time when the people in Vancouver start getting there. I've already worked for an east-coast company with a west-coast customer, and I was the one working the customer's hours.
I personally wake up much better with an alarm clock than without. I have a tendency to get too much sleep and not be alert during the day if nothing tells me to get up.
I'm not sure how Google looks greedy for selling for $85 what people are willing to buy for $100. Aside from the few odd small transactions, the stock hasn't been over $105; so the hype hasn't generated a going rate with more profit than the loss Google took on selling the IPO shares low.
I suspect the odd small transactions are what they were trying to reduce. If they'd set the price at $100, people would probably have only bought 63% of their successful bids, which would mean that there would be fewer shares available on the market, and there is a higher chance at any given moment that the number of shares available for a reasonable price is insufficient just because some traders happen to be reading the newspaper.
The 9/11 terrorists were dressed normally and travelling with real passports in their actual names. I'd probably have a hard time flying if I was trying to use an actual Al Queda member's passport, but that's just because I probably wouldn't match the picture at all.
Potential threats look just like ordinary citizens. Ordinary citizens seem to have a harder time getting on planes simply because there are more of them to have trouble and complain.
I wouldn't be surprised if they decided that the auction was unreasonably unpopular, and they'd do better selling their shares on the open market that buyers are more comfortable with. Since the stock hasn't traded as low as 85, they were clearly right.
It's perfectly justified for investors to decide what they believe their stock to be worth, and not sell it for less in the auction. Alternatively, they could have just bid on it, but that would probably have made people more upset.
I think Linux is actually more like the Roman Empire than you seem to imply, while Microsoft is more like one of the 20th century dictatorships.
The Roman Empire managed to be so large because it used a decentralized system of command. Provincial governors could act within their provinces without any oversight from Rome, but could call on Rome's military and financial power if needed. In return, they sent their profits back to the capital. Furthermore, allotment of governorships was based on the governor's record as seen by what they sent to Rome and what assistance they required in maintaining control. The governor could therefore maintain control of the province with the backing of a huge empire; Rome reaped the profits of the huge empire; but the emperor didn't have to think about the whole empire, because he could trust Spain's governor to deal with Spain or be replaced with someone who could. Furthermore, the senatorial class would manage itself in choosing who got control of what. Also, people from various parts of the government could get noticed by the senate and thereby put into more powerful positions. This system allowed Rome to rule a huge area for centuries, across many emperors and even dynasties.
Linux development is done in much this way. Linus doesn't do particularly much in running the project, relative to its size; most of the work is done by others who do it to show their productivity and merit, and who get rewarded by having their code put into the mainline and by having others contribute work to their subsystems, both of which increase the developer's reputation and ability to affect the design of the kernel.
Microsoft, on the other hand, is much more centralized, and the chain of command is more strict. This limits their size and flexibility and the ability of rising stars to affect direction. The loss of critical Microsoft executives would probably have much the effect that the death of Eva Peron had on Peronist Argentina than the death of Julius Caesar had on Rome.
Chances are, Microsoft won't last more than 50 years, while Linux will be around in 500. Not, of course, that that matters much to people under Microsoft's thumb at the moment.
Google doesn't seem to me to have any concern for shareholder confidence. I mean, their prospectus practically says, "We don't care what you investors think. We're going to run the company however we feel like, and you can kindly take your investment elsewhere if you aren't comfortable with that."
Then you'll find out how long the user of an unpatched system survives cut in half?
Re:Biologically speaking, how...
on
RGB to become RGBCMY
·
· Score: 3, Interesting
The human brain rarely says something isn't the right color. There's a huge amount of slop in the brain needed to produce the perception of stable colors of objects under different lighting conditions (if you light a room with light blue light, your eyes will adjust and report the usual colors of objects, even though the light reaching your eyes from them is obviously different).
The real issue is that, since the curves overlap, the green phosphor triggers the red cone to a certain extent, so green plus blue is cyan plus a bit of red, or a bit less cyan plus a bit of white. So the most pure cyan you can trigger in the eye with an RGB screen is less pure than the most pure cyan you get find in the real world. Purple is more of a mess (since the brain is actually making up colors for combinations that aren't generated by any pure wavelengths, and faking the idea that red is next to violet). But it all comes down to limits on the saturation of different colors due to not being able to keep from stimulating some cone or other.
Re:Uses existing signal and price is right.
on
RGB to become RGBCMY
·
· Score: 1
The issue with RGB is that it is limited in how much it can fool your eyes. Outside your body, in the real world, there is a continuous range of wavelengths. Your eye has receptors each of which is sensitive in different degrees to different wavelengths. The three color-based types are centered at red, green, and blue, but they overlap in sensitivity. That means that your red receptors respond to a certain extent to green light. If your image is made up exclusively of red, green, and blue light, less than the full range of output from the eye is possible, because you can't get the effect of pure cyan, which triggers the red receptor less than half and half green and blue.
Worse, you can't produce a real violet, which is characterized by stimulating the green receptor less than it is stimulated by pure blue. Your impression of purple on screen comes from your brain subtracting white from red and lots of blue to get blue with a deficit of green plus some white; that means that as you get further into the violet range in perception, you get less and less possible saturation and more perception of white.
Essentially, everything's fine toward the low-saturation (greyish) end, but getting very pure colors needs more light sources of different wavelengths.
It probably wouldn't be hard to produce scripts which would extract the configuration and build system from a linux distribution, so that you could distribute only that part initially, and get other files as needed. (Actually, you'd probably do it by the directory to keep things simple)
On the other hand, you'd have to download multiple things in order to build your kernel, and you'd have to download parts after you'd configured. The kernel site would have to either have a lot of little tar files (which wouldn't compress as well) or build tar files of your desired list on the fly (which increases the server load).
I doubt it would actually be worthwhile, but a patch to the build system to list the directories needed for a particular configuration would probably be welcome.
The BBC's FAQ says they started before Theora had gotten particularly far, and decided to continue as a separate codec because variety is good in codecs. Considering that there's already both Vorbis and FLAC for audio, it makes sense to have multiple video codecs with different characteristics, too.
They're working on it with a different packaging format presently, but they expect that someone will do Ogg Dirac at some point.
I have to agree with what (he says) Google does: hire Python programmers for Java development jobs. Regardless of what language your project is in, you'll do better if you hire developers who are motivated to try out (and learn from) tools that they don't need for their job and who know a wide variety of ways code can be structured.
On the other hand, I'm not impressed by Python. You can't apply any static reasoning to the code, because other code can dynamically change the code you're looking at. There's no way to determine if a variable is undeclared without actually running the program and getting to that line, because some other code could add it. Other code could even read a string from the user, and add a field with that name to an existing object. It's relatively hard to test Python code, because you have to exercise every line of code to find out if you have a simple typo.
Java does a reasonably good job (and a better job in 5) of making code that couldn't be right not compile. Better programmers make fewer design errors, but they don't make any fewer typos, so the verification that Java provides is, if anything, more important to better programmers.
It seems like most Java code is designed by bad designers and written by bad programmers, but that doesn't mean that Java is bad; it could be that a Python program is always more broken than an equivalent Java program, so only really good programmers can get Python programs to work at all.
for (SomeType x : collection) {
do_something(x); }
The differences are that the looping construct in Java isn't a member of the collection, but a keyword; and that you declare the element type with the variable. I think these are both advantages for making errors easy to detect.
The main downside to Java is that Sun is extremely slow in the core parts of the language (which is why it took so long to get the nice version of "for"), and pretty careless in the non-core library. On the other hand, they did manage to get a nice set of features together for version 5 that wouldn't have been so nice if they weren't all done together.
I think it would be neat to live in a world where everyone uses Linux, including the people who use Mac or Windows. I'd actually really like to see a completely non-*nix userspace with a Linux kernel. CONFIG_FS_DIRECTORY_SEPARATOR, anyone?
It should really be up to the browser to give combo-boxes for text inputs. In fact, Firefox makes slashdot's subject field a combo-box (except for lacking a button to pop up a dropdown of all of the suggestions). The document should, however, be able to give a list of suggested values which could be added to the browser's suggestions.
In fact, XForms supports "open selections" which is the form model equivalent to combo-boxes (keeping in mind that, for accessibility, the browser may get this input in some other way).
If this case gets dismissed, SCO can sue for copyright infringement. In fact, suing for slander of title is a weird thing to do in such a situation; it would be a bit like complaining that the guy who stole your car isn't maintaining it well. Slander of title is normally used when somebody else is telling people he owns the house you have the deed to; making it stick depends on there being some clear and unambiguous certification of ownership that you have and the other person does not, which pretty much means that it can never be used for copyrights unless the copyright dispute has already been settled and the loser is failing to abide by the outcome.
I'd guess that they're worried that the ownership of the copyrights wouldn't be decided in this trial anyway, since (as this motion points out) it is not necessarily a direct issue. Novell doesn't want to argue about the copyright situation and then have the court decide the case without a determination of the ownership of the copyrights, since that would just delay things further.
The court could, in fact, listen to a complete argument of who owns the copyrights, and then rule simply that SCO didn't have an uncontended claim to them. Novell wants to make sure that, if the judge is simply going to rule on whether SCO's claim is uncontended, Novell doesn't spend a long time in court arguing something that doesn't get ruled on.
My current favorite free software project is readline. It's 146k, and it's the best thing I've seen for getting text from the user. I'm only sorry that it only has a termcap interface, because that means that it's painful to type anything in a GUI.
It also has an insane range of features. It has keyboard macros, custom keymaps, four kinds of history search, application-specific tab completion, etc. If you don't like how backspace works, you can switch it to a bunch of different things.
If Microsoft had been keeping up with the state of the art, you could hit the up arrow in the IP address entry boxes to get IP addresses you've used recently, or hit ctrl-R to search backward through them for matches against the pattern you remember.
I wouldn't be surprised if someone actually challenged the triple damages law on the basis that it makes it undesireable for inventors to look at patents, and therefore that the system violates the patent office's constitutional mandate. Of course, it would probably not be argued in front of the present Supreme Court, since they didn't seem to go for a similar argument in Eldred v. Ashcroft.
My guess is that the initial products that this article is about will be essentially today's SD cards, except with higher throughput. As memory density continues to increase like it has, it will eventually pass the SD-card limits and ucard will be necessary.
I wouldn't be surprised if probe storage was to the SD format what microdrives are to the CF format; a new storage technology packaged in a standard package.
The reason that Apple didn't get a lot of market share was that they didn't price the systems right for that. The Mac was never sold as something that everyone would have, unlike either the PC or the Apple II. It was sold as something that could keep a company in the Fortune 500 with 4% of the market. Apple went for a strategy which could be (and was) successful with a very small segment of the market. Microsoft and a number of other companies went with strategies which demand a monopoly; of course, only one managed it.
If your plan is low margins and high volume, you have to beat everyone else who has this plan. If your plan is high margins and low volume, there's a lot more room for competition. Of course, in a market with a successful company of the first type and a number of successful companies of the second type, the first one has almost all of the market share, but that doesn't matter all that much. And as your margins get higher, the market share you need drops.
Apple probably could have done better by continuing the Apple II line until it could be folded into the Mac line, thereby keeping a foot in the low-end market and providing an upgrade path. They'd also have done better in the business market if they hadn't already orphaned a system, which makes users have to face the fact that they're using a closed system. But licensing the Mac to other companies would just have driven down the margins and made them need more market share.
Slashdot Mirror
The facts in general are contended, but it is clear that Juan's regime did not last much past Eva's death (although he did return later), and that his government was heavily reliant on the personalities and personal popularity of the leaders. This contrasts with the Roman Empire, which lasted through a number of leaders with a single continuing government,
Right, like your boss is going to let you waltz in at 10:30 because your brain-wave monitor didn't deem you rested until after 9. (And if you've had enough rest, don't you wake up naturally? Why use a device to do it for you?)
Sure, why not. You just have to stay late. Why pay people to sit around half-asleep in the morning? For that matter, the office will probably be half empty until noon Toronto time when the people in Vancouver start getting there. I've already worked for an east-coast company with a west-coast customer, and I was the one working the customer's hours.
I personally wake up much better with an alarm clock than without. I have a tendency to get too much sleep and not be alert during the day if nothing tells me to get up.
I'm not sure how Google looks greedy for selling for $85 what people are willing to buy for $100. Aside from the few odd small transactions, the stock hasn't been over $105; so the hype hasn't generated a going rate with more profit than the loss Google took on selling the IPO shares low.
I suspect the odd small transactions are what they were trying to reduce. If they'd set the price at $100, people would probably have only bought 63% of their successful bids, which would mean that there would be fewer shares available on the market, and there is a higher chance at any given moment that the number of shares available for a reasonable price is insufficient just because some traders happen to be reading the newspaper.
The 9/11 terrorists were dressed normally and travelling with real passports in their actual names. I'd probably have a hard time flying if I was trying to use an actual Al Queda member's passport, but that's just because I probably wouldn't match the picture at all.
Potential threats look just like ordinary citizens. Ordinary citizens seem to have a harder time getting on planes simply because there are more of them to have trouble and complain.
I wouldn't be surprised if they decided that the auction was unreasonably unpopular, and they'd do better selling their shares on the open market that buyers are more comfortable with. Since the stock hasn't traded as low as 85, they were clearly right.
It's perfectly justified for investors to decide what they believe their stock to be worth, and not sell it for less in the auction. Alternatively, they could have just bid on it, but that would probably have made people more upset.
I think Linux is actually more like the Roman Empire than you seem to imply, while Microsoft is more like one of the 20th century dictatorships.
The Roman Empire managed to be so large because it used a decentralized system of command. Provincial governors could act within their provinces without any oversight from Rome, but could call on Rome's military and financial power if needed. In return, they sent their profits back to the capital. Furthermore, allotment of governorships was based on the governor's record as seen by what they sent to Rome and what assistance they required in maintaining control. The governor could therefore maintain control of the province with the backing of a huge empire; Rome reaped the profits of the huge empire; but the emperor didn't have to think about the whole empire, because he could trust Spain's governor to deal with Spain or be replaced with someone who could. Furthermore, the senatorial class would manage itself in choosing who got control of what. Also, people from various parts of the government could get noticed by the senate and thereby put into more powerful positions. This system allowed Rome to rule a huge area for centuries, across many emperors and even dynasties.
Linux development is done in much this way. Linus doesn't do particularly much in running the project, relative to its size; most of the work is done by others who do it to show their productivity and merit, and who get rewarded by having their code put into the mainline and by having others contribute work to their subsystems, both of which increase the developer's reputation and ability to affect the design of the kernel.
Microsoft, on the other hand, is much more centralized, and the chain of command is more strict. This limits their size and flexibility and the ability of rising stars to affect direction. The loss of critical Microsoft executives would probably have much the effect that the death of Eva Peron had on Peronist Argentina than the death of Julius Caesar had on Rome.
Chances are, Microsoft won't last more than 50 years, while Linux will be around in 500. Not, of course, that that matters much to people under Microsoft's thumb at the moment.
Google doesn't seem to me to have any concern for shareholder confidence. I mean, their prospectus practically says, "We don't care what you investors think. We're going to run the company however we feel like, and you can kindly take your investment elsewhere if you aren't comfortable with that."
GPS is the thing right after FPS, obviously.
At least now, when you get the impulse to throttle Windows, it will work (326863)...
Then you'll find out how long the user of an unpatched system survives cut in half?
The human brain rarely says something isn't the right color. There's a huge amount of slop in the brain needed to produce the perception of stable colors of objects under different lighting conditions (if you light a room with light blue light, your eyes will adjust and report the usual colors of objects, even though the light reaching your eyes from them is obviously different).
The real issue is that, since the curves overlap, the green phosphor triggers the red cone to a certain extent, so green plus blue is cyan plus a bit of red, or a bit less cyan plus a bit of white. So the most pure cyan you can trigger in the eye with an RGB screen is less pure than the most pure cyan you get find in the real world. Purple is more of a mess (since the brain is actually making up colors for combinations that aren't generated by any pure wavelengths, and faking the idea that red is next to violet). But it all comes down to limits on the saturation of different colors due to not being able to keep from stimulating some cone or other.
The issue with RGB is that it is limited in how much it can fool your eyes. Outside your body, in the real world, there is a continuous range of wavelengths. Your eye has receptors each of which is sensitive in different degrees to different wavelengths. The three color-based types are centered at red, green, and blue, but they overlap in sensitivity. That means that your red receptors respond to a certain extent to green light. If your image is made up exclusively of red, green, and blue light, less than the full range of output from the eye is possible, because you can't get the effect of pure cyan, which triggers the red receptor less than half and half green and blue.
Worse, you can't produce a real violet, which is characterized by stimulating the green receptor less than it is stimulated by pure blue. Your impression of purple on screen comes from your brain subtracting white from red and lots of blue to get blue with a deficit of green plus some white; that means that as you get further into the violet range in perception, you get less and less possible saturation and more perception of white.
Essentially, everything's fine toward the low-saturation (greyish) end, but getting very pure colors needs more light sources of different wavelengths.
It probably wouldn't be hard to produce scripts which would extract the configuration and build system from a linux distribution, so that you could distribute only that part initially, and get other files as needed. (Actually, you'd probably do it by the directory to keep things simple)
On the other hand, you'd have to download multiple things in order to build your kernel, and you'd have to download parts after you'd configured. The kernel site would have to either have a lot of little tar files (which wouldn't compress as well) or build tar files of your desired list on the fly (which increases the server load).
I doubt it would actually be worthwhile, but a patch to the build system to list the directories needed for a particular configuration would probably be welcome.
The BBC's FAQ says they started before Theora had gotten particularly far, and decided to continue as a separate codec because variety is good in codecs. Considering that there's already both Vorbis and FLAC for audio, it makes sense to have multiple video codecs with different characteristics, too.
They're working on it with a different packaging format presently, but they expect that someone will do Ogg Dirac at some point.
I have to agree with what (he says) Google does: hire Python programmers for Java development jobs. Regardless of what language your project is in, you'll do better if you hire developers who are motivated to try out (and learn from) tools that they don't need for their job and who know a wide variety of ways code can be structured.
On the other hand, I'm not impressed by Python. You can't apply any static reasoning to the code, because other code can dynamically change the code you're looking at. There's no way to determine if a variable is undeclared without actually running the program and getting to that line, because some other code could add it. Other code could even read a string from the user, and add a field with that name to an existing object. It's relatively hard to test Python code, because you have to exercise every line of code to find out if you have a simple typo.
Java does a reasonably good job (and a better job in 5) of making code that couldn't be right not compile. Better programmers make fewer design errors, but they don't make any fewer typos, so the verification that Java provides is, if anything, more important to better programmers.
It seems like most Java code is designed by bad designers and written by bad programmers, but that doesn't mean that Java is bad; it could be that a Python program is always more broken than an equivalent Java program, so only really good programmers can get Python programs to work at all.
The main downside to Java is that Sun is extremely slow in the core parts of the language (which is why it took so long to get the nice version of "for"), and pretty careless in the non-core library. On the other hand, they did manage to get a nice set of features together for version 5 that wouldn't have been so nice if they weren't all done together.
I think it would be neat to live in a world where everyone uses Linux, including the people who use Mac or Windows. I'd actually really like to see a completely non-*nix userspace with a Linux kernel. CONFIG_FS_DIRECTORY_SEPARATOR, anyone?
Maybe they spent too much time looking at the handbook...
(You know you come from a family of engineers when "chemical rubber" makes you think of a company name that was changed before you were born)
It should really be up to the browser to give combo-boxes for text inputs. In fact, Firefox makes slashdot's subject field a combo-box (except for lacking a button to pop up a dropdown of all of the suggestions). The document should, however, be able to give a list of suggested values which could be added to the browser's suggestions.
In fact, XForms supports "open selections" which is the form model equivalent to combo-boxes (keeping in mind that, for accessibility, the browser may get this input in some other way).
If this case gets dismissed, SCO can sue for copyright infringement. In fact, suing for slander of title is a weird thing to do in such a situation; it would be a bit like complaining that the guy who stole your car isn't maintaining it well. Slander of title is normally used when somebody else is telling people he owns the house you have the deed to; making it stick depends on there being some clear and unambiguous certification of ownership that you have and the other person does not, which pretty much means that it can never be used for copyrights unless the copyright dispute has already been settled and the loser is failing to abide by the outcome.
I'd guess that they're worried that the ownership of the copyrights wouldn't be decided in this trial anyway, since (as this motion points out) it is not necessarily a direct issue. Novell doesn't want to argue about the copyright situation and then have the court decide the case without a determination of the ownership of the copyrights, since that would just delay things further.
The court could, in fact, listen to a complete argument of who owns the copyrights, and then rule simply that SCO didn't have an uncontended claim to them. Novell wants to make sure that, if the judge is simply going to rule on whether SCO's claim is uncontended, Novell doesn't spend a long time in court arguing something that doesn't get ruled on.
My current favorite free software project is readline. It's 146k, and it's the best thing I've seen for getting text from the user. I'm only sorry that it only has a termcap interface, because that means that it's painful to type anything in a GUI.
It also has an insane range of features. It has keyboard macros, custom keymaps, four kinds of history search, application-specific tab completion, etc. If you don't like how backspace works, you can switch it to a bunch of different things.
If Microsoft had been keeping up with the state of the art, you could hit the up arrow in the IP address entry boxes to get IP addresses you've used recently, or hit ctrl-R to search backward through them for matches against the pattern you remember.
I wouldn't be surprised if someone actually challenged the triple damages law on the basis that it makes it undesireable for inventors to look at patents, and therefore that the system violates the patent office's constitutional mandate. Of course, it would probably not be argued in front of the present Supreme Court, since they didn't seem to go for a similar argument in Eldred v. Ashcroft.
My guess is that the initial products that this article is about will be essentially today's SD cards, except with higher throughput. As memory density continues to increase like it has, it will eventually pass the SD-card limits and ucard will be necessary.
I wouldn't be surprised if probe storage was to the SD format what microdrives are to the CF format; a new storage technology packaged in a standard package.
The reason that Apple didn't get a lot of market share was that they didn't price the systems right for that. The Mac was never sold as something that everyone would have, unlike either the PC or the Apple II. It was sold as something that could keep a company in the Fortune 500 with 4% of the market. Apple went for a strategy which could be (and was) successful with a very small segment of the market. Microsoft and a number of other companies went with strategies which demand a monopoly; of course, only one managed it.
If your plan is low margins and high volume, you have to beat everyone else who has this plan. If your plan is high margins and low volume, there's a lot more room for competition. Of course, in a market with a successful company of the first type and a number of successful companies of the second type, the first one has almost all of the market share, but that doesn't matter all that much. And as your margins get higher, the market share you need drops.
Apple probably could have done better by continuing the Apple II line until it could be folded into the Mac line, thereby keeping a foot in the low-end market and providing an upgrade path. They'd also have done better in the business market if they hadn't already orphaned a system, which makes users have to face the fact that they're using a closed system. But licensing the Mac to other companies would just have driven down the margins and made them need more market share.