The rest of the problem is habit, and water will work there to. Whenever you feel like you need a cup of coffee or a can of soda, drink some water instead.
Decaf can be a helpful crutch, too. You can simply substitute it for full-strength coffee whenever you'd drink that. This is especially good for people who actually enjoy coffee for more than its caffeine content. An important note, though, is that decaf is not completely caffeine free; it contains something like 2-5% of the caffeine content of normal coffee, so if it's important to you to cut out caffeine completely it won't work.
(Further note: the correct spelling in caffeine. This violates the normal English rule of I before E except after C or sounding like AY. This is because "ine" is a suffix used by chemists to indicate alkyloids, the class of compounds to which caffeine belongs. So "caffeine" means the alkyloid found in "caffe" or coffee. This is similar to cocaine (found in coca), nicotine (found in nicotina), and many others.)
The purpose of taint checking is more as a debugging tool than an absolute check on illegitimate data. This fits with the general Perl view that the function of the language is to assist the programmer rather than to constrain him. Thus if you turn on taint checking, Perl will stop you from doing things directly using tainted data, but it lets you "launder" the data by running it through a regexp. This isn't a perfect solution, but anything that was radically better would be all-but impossible to code. It won't stop a lazy programmer from using:
$var =~ m/^(.*)$/; $var = $1;
to bypass the checking (false laziness!), but that's not its purpose. What it will do is to catch many non-obvious mistakes by a programmer who's actually trying to write safe code.
Actually he didn't miss that point at all. This was mentioned in the very end of the article under the heading "Conclusion: Common Sense Prevails". Some of the comments there include:
Over and above the clear test results, our price-performance analysis clearly shows that the added performance of CPUs in the upper bracket bears no sensible relation to the extra price....
In the gaming sector, many processor makers are dogged by the fact that only a few programs need really fast CPUs. One reason for this development is the displacement of graphics-intensive operations to the graphics card; another is the ongoing tense competition between AMD and Intel that long ago outstripped the requirements of modern standard software in terms of performance....
Novices should certainly consider the AthlonXP 2600+ or 2800+, since a serviceable platform with 512 MB of memory is inexpensive and will do nicely for the next 18-24 months....
The AMD Athlon64 FX and Intel's Pentium 4 Extreme Edition are still status symbols for the computing jet set. After all, you can pick up a complete and high-performance system for between $750 and $1,000, which as our benchmarks show, also offer a superior price/performance ratio.
That certainly sounds like somebody who understands that most ordinary users will get all the performance they need by buying a cheaper processor, especially one of the notably cheaper AMD models.
Why should the Linux ones have to? Linus isn't a party to the USL/BSDI lawsuit, so he isn't bound by the terms of the settlement. There's no reason to think that there are any valid copyrights to files involved in the suit (USL settled because the judge was threatening to rule that their copyrights were invalid) so there's no need for any copyright notice unless it's forced by some other legal reason, like the settlement. And that's assuming that the files actually are copied and haven't been re-implemented from scratch using the POSIX standards.
If you try opening the directory that you get from yum.conf in a browser, it takes you to the directory that I described. I'm not sure why they've done things that way, but I found that I got better results if I rewrote my yum.conf to use the re-directed directory. That, at least, always seems to exist.
I also checked, and you can add the Development directory to your yum.conf and download the latest, greatest, least stable stuff onto your computer. I added:
[fedora-development]
name=Fedora Development Packages (Development)
baseurl=http://download.fedora.redhat.com/pub/fedo ra/linux/core/development/$basearch
gpgcheck=1
to my yum.conf and it gave me the option of updating to 2.6.0.
Not unreasonable actions in themselves, but in those circumstances it killed them.
The reason that they were fatal is that their competitor wasn't playing by the rules. If Netscape had decided to give their product away to buy marketshare, Microsoft would simply have kept up their practice of giving theirs away free. Netscape needed to start charging for their products some time, because that was their only source of revenue. It's not as though they could keep giving their products away for long enough to run Microsoft through its cash reserves. Whenever they did start charging, MS would still be there and able to undercut them with a free competing product, and all of their expensively purchased market share would evaporate. The only question was when Netscape was going to start charging, not what would happen when they did.
Anti-trust law forbids what Microsoft did specifically to give companies like Netscape a chance. You can't have a free market without competetion, and you can't have competetion if a monopolist can destroy competitors whenever they try to enter the market. Netscape lost because they could only win with prompt anti-trust enforcement, and the government stepped in too late to save them.
My mistake. It's not under updates/testing. It's under development. Also, FWIW, I found that I just got a 404 if I tried using the addresses in the original yum.conf. I got much better results when I updated mine to include (for updates-testing as an example):
I'm not 100% certain that the equivalent directory for development (which would be/pub/fedora/linux/core/development/$basearch) is a valid yum channel, but it ought to work if you want to keep on the bleeding edge of things. That looks to be roughly the Fedora equivalent of Rawhide, with newer goodies. It does include kernel-2.6.0-0.1.14.i686.rpm.
It's already available on the testing channel for Fedora. If you have the updates-testing in your yum configuration, you can update to 2.6 with a yum update.
School shouldn't be about filling your head with facts, but about encouraging you to study things that you're interested in.
That's probably true, but there are two obvious limits. One is that you need some knowledge base just to know what things are out there that you might be interested in studying. You'll never become interested in the surgeon who perfected heart transplants if you never learn that it's possible to transplant organs from one person to another. You won't become interested in Chuck Yeager if you don't have the idea that supersonic flight didn't exsist at one time (or you don't know that it exists today).
The other is that there are some really important facts that everyone needs to know, whether they're boring or not. Everyone needs to know enough about the law to avoid getting thrown in jail, for instance. It's also important to know how the government works, where one's society came from (socially and politically), enough of a common culture to have the same references as everyone else, enough about science to make informed choices, etc. That stuff may or may not be boring, but forcing people to learn it is a critical function of schools.
Having seen 2 GB USB memory keys starting to become available, I have to wonder what the great advantage is of the microdrive.
Price. Pricewatch lists the cheapest 2GB USB memory key as going for $514 and the cheapest 2GB microdrive as going for $195. In 4 GB sizes I'd expect the microdrive to have an even bigger advantage, but there's no listing for 4 GB USB memory keys, probably because they're too expensive for anyone to think about them.
Actually, IBM (now Toshiba) microdrives are distinctly cheaper than solid-state media of the same storage capacity- at least in the 2 and 4 GB sizes that they're selling these days. A quick check at Pricewatch shows 2 GB microdrives (in Compact Flash packaging) selling for as little as $195, while the cheapest solid state 2 GB CF is $430. 4 GB sizes are not listed at Pricewatch, but the price advantage for Microdrives is likely to be even greater there. When I checked at Ritz Camera, the 4 GB microdrive was only about 50% more than the 2GB, while the 4 GB solid state CF was more than double the price of 2 GB.
There are recommendations to not format flash to ext3 because of frequent rewrites to the same sectors, which could cause the flash cells to end-of-life pretty quickly, but hard disks don't generally have that problem. (OTOH, neither do vfat systems...not sure about ext2.)
And Linux has a solution specifically designed to deal with this problem. There's a JFFS (journaling flash file system) that's designed specifically for use on flash memory devices. I'm not 100% certain, but I'm pretty sure that the designers took the problem of flash wearing out with frequent re-writes into account when designing the filesystem.
There is the tiny problem of the price. The 2 GB is going for $700 and the 4 GB for $1600. The equivalent microdrives are $350 and $500. I don't know about you, but I can think of plenty of things to do with the $350 to $1100 I'd save by going with the hard drive over flash memory.
It's already available, at least for more expensive cameras. You can get an IBM microdrive in a Compact Flash Type II form factor, which is a bit thicker but otherwise dimensionally compatible with the regular Compact Flash cards. Less expensive cameras aren't designed to accept both Type I and Type II, but many of the high-end ones- including all of the Digital SLRs, AFAIK- are. The extra capacity is obviously really useful when dealing with a 6+ megapixel camera that may want to save pictures in raw (i.e. not compressed) format. The availablility of hard drive storage is one of the key things that keeps Compact Flash relevant; it's bigger and clunkier than other card types, but at the very high end it can hold way more than any of the others.
Linus is literally in the minority in this case in that he is a member of Finland's small Swedish-speaking population. He apparently grew up speaking Swedish at home, so I assume that he was fluent in both Swedish and Finnish before learning English.
At least Marcello has the excuse that he's not a native English speaker. I guess you could say the same thing about Bush, though; he speaks Texan. It's interesting, though, that Linus is so much more articulate than Darl McBride even though English must be at least his third language (after Swedish and Finnish).
Somebody else pointed out an additional dimension, which is that IBM feels that its integrity has been impugned. A key part of SCO's suit is that IBM has failed to live up to its contractual obligations. That's a big deal in business, because nobody will want to deal with you if they don't believe that you'll hold up your side of the bargain. IBM clearly has a very strong business reason to want their reputation cleared, which is a driving factor behind the Lanham Act violation countersuit. I sincerely doubt that they would accept any deal that didn't clear up that issue unequivocally.
I'm not sure if that's true, or at least if it's true in the sense that you mean it. Part of the defensive value of IBM's suit is from making an example of SCO. IBM wants to prove to anyone else who might be thinking of launching a nuisance suit that it's a really, really bad idea. If the result is just a draw, or even a small win in IBM's favor (like SCO agreeing to pay legal fees) then there isn't sufficient example. If, OTOH, the final result is smoking craters where SCO, Canopy, et. al. used to be, it sends the much stronger message that suing IBM is a good way to lose everything.
Re:Controls everything?
on
AOL's $299 PC
·
· Score: 3, Insightful
As a result of what they have learned they believe that offering a really cheap computer will attract/retain members.
I think that this is really the big point. ISPs have surprisingly good customer loyalty when you consider how little difference there is between one dialup service and another. I guess that a big part of it is just the hassle of switching email addresses. In any case, AOL knows that many of those people who sign up for one year of service will continue to get service from AOL for the forseeable future, bringing in a lot more revenue than the subsidy on the computer. It may be an especially good idea if Microsoft is succeeding in attracting newbies away from AOL by making it easy to sign up for MSN from Windows.
Re:Deja vu, MSN
on
AOL's $299 PC
·
· Score: 5, Informative
I think that you're remembering incorrectly. ISTR that the thing that really killed the MSN rebate business was that it had legal problems in some jurisdictions. In particular, California law made is such that people could sign up, drop the MSN service immediately, and not have to return the $400. When customers started to abuse this in droves (i.e. not long after somebody figured it out), Microsoft decided to can the idea.
What we need is a better package management system that even grandma could use. Is UserLinux going to try and solve that problem or just slap a frontend on apt and call it good to go?
What's wrong with apt or portage (or yum for that matter) other than their UI? They do an excellent job of the essential function of a package-management system, which is resolving dependencies and installing or uninstalling the packages that need to be added or removed. The biggest issues for a good UI would be presenting a reasonable list of possible packages to be installed (which actually is a biggie) and configuring the packages when you're done installing them (which is also a biggie). (I actually think that portage will always have problems with the "Grandma test" because of compile times, but I'll set that aside for the time being.)
As a practical matter, I think that the issue of presenting a list of packages (at least in "Grandma mode") is solvable by using a coarse granularity of packages. If the options are "Basic Desktop", "Business Programs", "Software Development", "Mail Server", etc. rather than individual programs then it should be possible to list them all in a reasonably sized UI. Configuration is harder, but Bruce has already mentioned that it's likely to be one of the big thrusts of the development process. A combination of sane defaults (so that most users don't have to change anything) and maybe auto-running the appropriate GUI based configuration utility on installation might help to solve this kind of problem.
Most users want to install a named program, like "OpenOffice.org" or "Mozilla", and could not care less what libraries that will bring in.
I think that most users don't even think at that level. Most people think at the level of basic function rather than specific programs. If they're setting up a desktop, they want to get a basic desktop environment, email, web browser, word processor, spreadsheet, presentation program, utilities, etc. without having to say so. So the system should just have a "Standard Desktop" option that will install all that stuff without the user needing to fiddle over each and every program in the group. You could have a fallback system that would let you install package-by-package if you needed, but I doubt that most users would use it much.
It seems to me that an obvious example of this would be your idea of having more automated configuration. You suggested having a sort of configuration metapackage that had the primary package as a dependency and then did the work of configuring the primary package after it was installed. Debian might or might not want such a thing as part of their system, so it might be necessary to maintain it separately from the main Debian tree.
I'd assume that some of the ease of installation could be handled by a similar use of metapackages. Each function that a system might fill would have its own metapackage that didn't do anything except for having a set of dependencies that would cause the system to install appropriate functionality. A "desktop" package would have a basic desktop environment, productivity software, and the like, while a "Windows Server" package would include SAMBA plus configuration tools. This would work well with the idea of minimizing the number of different tools for a given job; if you provide only one MTA then you can make a lot more assumptions about setting up a mail server than if you had several choices.
You may need to save it to your hard-drive and view it using an image viewer if your browser cannot open it successfully.
I'm guessing that web browsers probably have some limits on the size of images that they're capable of showing. I'd guess that it's probably easier to program an image viewer including hard limits on the size of the images. After all, how many people are going to try viewing images bigger than a few thousand pixels wide in their web browsers?
It may be true that there are many selfish reasons to pursue a project. I'll admit that most of my own programming is done to solve my own particular problems. But Linus and RMS could very easily have decided to keep their programs to themselves, or to sell what they had made to the highest bidder. Many other people have done just that, and some of them have gone on to make lots of money by doing so. What made Linus and (especially) RMS different is that they decided to share their work with others. The decision to share may not have been motivated by anything so high as bettering humanity, but it was a fundamentally altruistic one.
Slashdot Mirror
Decaf can be a helpful crutch, too. You can simply substitute it for full-strength coffee whenever you'd drink that. This is especially good for people who actually enjoy coffee for more than its caffeine content. An important note, though, is that decaf is not completely caffeine free; it contains something like 2-5% of the caffeine content of normal coffee, so if it's important to you to cut out caffeine completely it won't work.
(Further note: the correct spelling in caffeine. This violates the normal English rule of I before E except after C or sounding like AY. This is because "ine" is a suffix used by chemists to indicate alkyloids, the class of compounds to which caffeine belongs. So "caffeine" means the alkyloid found in "caffe" or coffee. This is similar to cocaine (found in coca), nicotine (found in nicotina), and many others.)
The purpose of taint checking is more as a debugging tool than an absolute check on illegitimate data. This fits with the general Perl view that the function of the language is to assist the programmer rather than to constrain him. Thus if you turn on taint checking, Perl will stop you from doing things directly using tainted data, but it lets you "launder" the data by running it through a regexp. This isn't a perfect solution, but anything that was radically better would be all-but impossible to code. It won't stop a lazy programmer from using:
to bypass the checking (false laziness!), but that's not its purpose. What it will do is to catch many non-obvious mistakes by a programmer who's actually trying to write safe code.
Actually he didn't miss that point at all. This was mentioned in the very end of the article under the heading "Conclusion: Common Sense Prevails". Some of the comments there include:
That certainly sounds like somebody who understands that most ordinary users will get all the performance they need by buying a cheaper processor, especially one of the notably cheaper AMD models.
Why should the Linux ones have to? Linus isn't a party to the USL/BSDI lawsuit, so he isn't bound by the terms of the settlement. There's no reason to think that there are any valid copyrights to files involved in the suit (USL settled because the judge was threatening to rule that their copyrights were invalid) so there's no need for any copyright notice unless it's forced by some other legal reason, like the settlement. And that's assuming that the files actually are copied and haven't been re-implemented from scratch using the POSIX standards.
If you try opening the directory that you get from yum.conf in a browser, it takes you to the directory that I described. I'm not sure why they've done things that way, but I found that I got better results if I rewrote my yum.conf to use the re-directed directory. That, at least, always seems to exist.
I also checked, and you can add the Development directory to your yum.conf and download the latest, greatest, least stable stuff onto your computer. I added:
to my yum.conf and it gave me the option of updating to 2.6.0.
The reason that they were fatal is that their competitor wasn't playing by the rules. If Netscape had decided to give their product away to buy marketshare, Microsoft would simply have kept up their practice of giving theirs away free. Netscape needed to start charging for their products some time, because that was their only source of revenue. It's not as though they could keep giving their products away for long enough to run Microsoft through its cash reserves. Whenever they did start charging, MS would still be there and able to undercut them with a free competing product, and all of their expensively purchased market share would evaporate. The only question was when Netscape was going to start charging, not what would happen when they did.
Anti-trust law forbids what Microsoft did specifically to give companies like Netscape a chance. You can't have a free market without competetion, and you can't have competetion if a monopolist can destroy competitors whenever they try to enter the market. Netscape lost because they could only win with prompt anti-trust enforcement, and the government stepped in too late to save them.
My mistake. It's not under updates/testing. It's under development. Also, FWIW, I found that I just got a 404 if I tried using the addresses in the original yum.conf. I got much better results when I updated mine to include (for updates-testing as an example):
baseurl=http://download.fedora.redhat.com/pub/ fedora/linux/core/updates/testing/$releasever/$bas earch
I'm not 100% certain that the equivalent directory for development (which would be /pub/fedora/linux/core/development/$basearch) is a valid yum channel, but it ought to work if you want to keep on the bleeding edge of things. That looks to be roughly the Fedora equivalent of Rawhide, with newer goodies. It does include kernel-2.6.0-0.1.14.i686.rpm.
It's already available on the testing channel for Fedora. If you have the updates-testing in your yum configuration, you can update to 2.6 with a yum update.
That's probably true, but there are two obvious limits. One is that you need some knowledge base just to know what things are out there that you might be interested in studying. You'll never become interested in the surgeon who perfected heart transplants if you never learn that it's possible to transplant organs from one person to another. You won't become interested in Chuck Yeager if you don't have the idea that supersonic flight didn't exsist at one time (or you don't know that it exists today).
The other is that there are some really important facts that everyone needs to know, whether they're boring or not. Everyone needs to know enough about the law to avoid getting thrown in jail, for instance. It's also important to know how the government works, where one's society came from (socially and politically), enough of a common culture to have the same references as everyone else, enough about science to make informed choices, etc. That stuff may or may not be boring, but forcing people to learn it is a critical function of schools.
Price. Pricewatch lists the cheapest 2GB USB memory key as going for $514 and the cheapest 2GB microdrive as going for $195. In 4 GB sizes I'd expect the microdrive to have an even bigger advantage, but there's no listing for 4 GB USB memory keys, probably because they're too expensive for anyone to think about them.
Actually, IBM (now Toshiba) microdrives are distinctly cheaper than solid-state media of the same storage capacity- at least in the 2 and 4 GB sizes that they're selling these days. A quick check at Pricewatch shows 2 GB microdrives (in Compact Flash packaging) selling for as little as $195, while the cheapest solid state 2 GB CF is $430. 4 GB sizes are not listed at Pricewatch, but the price advantage for Microdrives is likely to be even greater there. When I checked at Ritz Camera, the 4 GB microdrive was only about 50% more than the 2GB, while the 4 GB solid state CF was more than double the price of 2 GB.
And Linux has a solution specifically designed to deal with this problem. There's a JFFS (journaling flash file system) that's designed specifically for use on flash memory devices. I'm not 100% certain, but I'm pretty sure that the designers took the problem of flash wearing out with frequent re-writes into account when designing the filesystem.
There is the tiny problem of the price. The 2 GB is going for $700 and the 4 GB for $1600. The equivalent microdrives are $350 and $500. I don't know about you, but I can think of plenty of things to do with the $350 to $1100 I'd save by going with the hard drive over flash memory.
It's already available, at least for more expensive cameras. You can get an IBM microdrive in a Compact Flash Type II form factor, which is a bit thicker but otherwise dimensionally compatible with the regular Compact Flash cards. Less expensive cameras aren't designed to accept both Type I and Type II, but many of the high-end ones- including all of the Digital SLRs, AFAIK- are. The extra capacity is obviously really useful when dealing with a 6+ megapixel camera that may want to save pictures in raw (i.e. not compressed) format. The availablility of hard drive storage is one of the key things that keeps Compact Flash relevant; it's bigger and clunkier than other card types, but at the very high end it can hold way more than any of the others.
Linus is literally in the minority in this case in that he is a member of Finland's small Swedish-speaking population. He apparently grew up speaking Swedish at home, so I assume that he was fluent in both Swedish and Finnish before learning English.
At least Marcello has the excuse that he's not a native English speaker. I guess you could say the same thing about Bush, though; he speaks Texan. It's interesting, though, that Linus is so much more articulate than Darl McBride even though English must be at least his third language (after Swedish and Finnish).
Somebody else pointed out an additional dimension, which is that IBM feels that its integrity has been impugned. A key part of SCO's suit is that IBM has failed to live up to its contractual obligations. That's a big deal in business, because nobody will want to deal with you if they don't believe that you'll hold up your side of the bargain. IBM clearly has a very strong business reason to want their reputation cleared, which is a driving factor behind the Lanham Act violation countersuit. I sincerely doubt that they would accept any deal that didn't clear up that issue unequivocally.
I'm not sure if that's true, or at least if it's true in the sense that you mean it. Part of the defensive value of IBM's suit is from making an example of SCO. IBM wants to prove to anyone else who might be thinking of launching a nuisance suit that it's a really, really bad idea. If the result is just a draw, or even a small win in IBM's favor (like SCO agreeing to pay legal fees) then there isn't sufficient example. If, OTOH, the final result is smoking craters where SCO, Canopy, et. al. used to be, it sends the much stronger message that suing IBM is a good way to lose everything.
I think that this is really the big point. ISPs have surprisingly good customer loyalty when you consider how little difference there is between one dialup service and another. I guess that a big part of it is just the hassle of switching email addresses. In any case, AOL knows that many of those people who sign up for one year of service will continue to get service from AOL for the forseeable future, bringing in a lot more revenue than the subsidy on the computer. It may be an especially good idea if Microsoft is succeeding in attracting newbies away from AOL by making it easy to sign up for MSN from Windows.
I think that you're remembering incorrectly. ISTR that the thing that really killed the MSN rebate business was that it had legal problems in some jurisdictions. In particular, California law made is such that people could sign up, drop the MSN service immediately, and not have to return the $400. When customers started to abuse this in droves (i.e. not long after somebody figured it out), Microsoft decided to can the idea.
What's wrong with apt or portage (or yum for that matter) other than their UI? They do an excellent job of the essential function of a package-management system, which is resolving dependencies and installing or uninstalling the packages that need to be added or removed. The biggest issues for a good UI would be presenting a reasonable list of possible packages to be installed (which actually is a biggie) and configuring the packages when you're done installing them (which is also a biggie). (I actually think that portage will always have problems with the "Grandma test" because of compile times, but I'll set that aside for the time being.)
As a practical matter, I think that the issue of presenting a list of packages (at least in "Grandma mode") is solvable by using a coarse granularity of packages. If the options are "Basic Desktop", "Business Programs", "Software Development", "Mail Server", etc. rather than individual programs then it should be possible to list them all in a reasonably sized UI. Configuration is harder, but Bruce has already mentioned that it's likely to be one of the big thrusts of the development process. A combination of sane defaults (so that most users don't have to change anything) and maybe auto-running the appropriate GUI based configuration utility on installation might help to solve this kind of problem.
I think that most users don't even think at that level. Most people think at the level of basic function rather than specific programs. If they're setting up a desktop, they want to get a basic desktop environment, email, web browser, word processor, spreadsheet, presentation program, utilities, etc. without having to say so. So the system should just have a "Standard Desktop" option that will install all that stuff without the user needing to fiddle over each and every program in the group. You could have a fallback system that would let you install package-by-package if you needed, but I doubt that most users would use it much.
It seems to me that an obvious example of this would be your idea of having more automated configuration. You suggested having a sort of configuration metapackage that had the primary package as a dependency and then did the work of configuring the primary package after it was installed. Debian might or might not want such a thing as part of their system, so it might be necessary to maintain it separately from the main Debian tree.
I'd assume that some of the ease of installation could be handled by a similar use of metapackages. Each function that a system might fill would have its own metapackage that didn't do anything except for having a set of dependencies that would cause the system to install appropriate functionality. A "desktop" package would have a basic desktop environment, productivity software, and the like, while a "Windows Server" package would include SAMBA plus configuration tools. This would work well with the idea of minimizing the number of different tools for a given job; if you provide only one MTA then you can make a lot more assumptions about setting up a mail server than if you had several choices.
You could try reading the page:
I'm guessing that web browsers probably have some limits on the size of images that they're capable of showing. I'd guess that it's probably easier to program an image viewer including hard limits on the size of the images. After all, how many people are going to try viewing images bigger than a few thousand pixels wide in their web browsers?
It may be true that there are many selfish reasons to pursue a project. I'll admit that most of my own programming is done to solve my own particular problems. But Linus and RMS could very easily have decided to keep their programs to themselves, or to sell what they had made to the highest bidder. Many other people have done just that, and some of them have gone on to make lots of money by doing so. What made Linus and (especially) RMS different is that they decided to share their work with others. The decision to share may not have been motivated by anything so high as bettering humanity, but it was a fundamentally altruistic one.