A computer is not alive, and a program is not an infectious disease. "Computer virus" is a wonderful metaphor, but if we get carried away with it, we lose sight of the difference between computer-virus hosts and real-life virus hosts.
Consider a live, naturally-evolved plant. It can't be too picky about the kind of soil it's planted in, the temperature of its surroundings, the amount of sunlight it gets each day, the amount of rain that falls on it, the chemical content of the rain, etc. It has to convert whatever resources it has available into forms of matter and energy that it can use for growth, reproduction, and defense against predators.
Since it's competing with other plants for survival, it has to make engineering trade-offs. For example, a plant may secrete a bad-tasting chemical, protecting it from predators -- but the raw materials for that chemical may restrict the plant to certain soil chemistries, or the extra energy necessary to produce it may restrict it to environments with a certain minimal level of sunlight.
By contrast, our computers are designed to "live" in a tightly controlled environment. If a CPU is running a binary program, and comes across an instruction that makes no sense, it doesn't have to try "digesting" the program into a more sensible form. If a network router gets a packet with a syntactically incorrect header, it doesn't have to send the packet any farther.
Security-related protocols add to the level of control; they are filters between sensitive and untrusted parts of a computer system. Since our computers "live" off of electricity, not programs, they can be as picky as we want them to be about what programs they execute and what permissions they execute under. Since our operating systems are designed by (occasionally) intelligent humans, not evolved by natural selection, humans can design better operating systems, in which security against untrusted code is a fundamental part of the architecture.
In fiscal years '97 and '98, Red Hat actually made a profit. Even in FY'99 (according to their S-1), they had a net loss of about $91K on a total revenue of about $10,013K. (Their fiscal year ends on February 28.)
By comparison, Amazon.com, the poster child for the high-flying net.stock, has lost money every year since its inception. In 1999, according to their most recent 10-K, they had net sales of about $610M and a net loss of about $125M. Every year before that, their losses-to-sales ratio has been even higher.
Asides: (1) You can look up this kind of data at www.freedgar.com; they get their information from the electronic filings at the US Securities and Exchange Commission, the government agency that regulates the stock market. (2) Since the Bryar column now has a link to Slashdot, and since people are claiming that the author's mail server is slashdotted, I'll just respond here. (3) High-tech stocks in general seem too overvalued for my taste, so I'm not planning to buy any Red Hat stock.
If you want to optimize Perl for M$ machines, the best way to do it is through M$, simply because they make the OS and they keep the source code private, so their developers will know how to milk the OS before anyone else does. To say that M$ shouldn't be funding Perl development on Windows implies that someone else should be funding it. Who? Borland? Cygnus?
Furthermore, this deal testifies to the strength of Perl in the marketplace. I'm sure Gates would be much happier if he could ignore Perl and focus corporate resources on marketing Visual Basic.
And if M$ feels that having a good Perl implementation will lure people from the Unix world to NT, they might actually put some competent people to work on the project. M$ Word was a decent word processor... back when they were competing with WordPerfect.
We need, essentially, documentation that can be approached in many ways, with assistance for newbies of differing levels that doesn't get in the way of the main document itself. Isn't this what hypertext is good for?
We also need documentation that is accessible to people who don't have any OS running on their machines yet, or who haven't configured X properly yet, or who don't know why the shell is saying "man: command not found".
Where does the extra money go? I suspect it goes to the publisher.
A lot of it goes into the distribution system.
In the US (I don't know about other places), if you buy a book and don't like it, you can return it to the bookstore for a refund. The bookstore then returns the book to the publisher, and the publisher gives the store credit for another purchase from that publisher. If the book is a mass-market paperback, the bookstore doesn't bother returning the whole book; it just returns the torn-off book cover.
This is an accountant's nightmare. The publisher can't predict how much profit a book has made until the returns bounce back. (On average, though, about half of all books sold are returned, so when you buy a book, you're really paying for the printing of two books.)
And as long as the publisher has books sitting in the warehouse, they're taxed at their current wholesale price, even if the publisher knows that a certain proportion of these copies are destined for Buck-A-Book.
And the people who own publishers don't simply want their companies to make money; they want them to make as much money as other businesses. A book that is likely to become gradually more popular over time, or a book that would have a small but devoted following, is therefore not a good investment for the publisher.
And most titles don't make money for their publishers; the profits from Stephen King and Robert Jordan novels cover the losses that their publishers suffer from promising but unsuccessful works by others.
Christopher Browne,[*] in this article, points out (I'm paraphrasing):
Setting up tax-exempt charities to fund free software development is good for the donors' tax bill, but if the charity is paying programmers to write the software, the programmers' salaries are taxable.
On the other hand, if you give a cash gift to your favorite free-software programmer, you can't deduct the gift on your taxes, but the programmer doesn't have to declare it as income, either.
Back on the first hand, a gift, by definition, has no strings attached, so if you give a gift to your favorite free-software programmer, and that programmer takes the money and retires to Bermuda, you have no recourse.
In the USA, libel law has a few important features:
If the court is convinced that the defendant told the truth, the plaintiff loses.
If a "public figure" brings a libel suit, it's not enough to prove that the defendant made a false and damaging statment; the plaintiff must also prove that the defendant either acted with malice or had negligent disregard for whether or not the statement was true.
It's difficlut, if not impossible, for a plaintiff to win when the allegedly libelous work is a parody or a political commentary.
(IANAL... just a former college-newspaper reporter.)
I know that some other countries' libel laws make it easier for plaintiffs to win, and I know that UK law is not as strict about free speech as the US Constitution is... but I don't know details. Can any Brits or legal eagles on/. help us out here?
Perl seems to my mind to actually help readability in many ways. the fact that you can tell at a glance that $array is a scalar variable and @variable is an array is (again depite the name) and array is easy.
Except that $array could also be a reference to an array.
I've spent a bit of time beating my head against complex data structures in Perl, and while I'm sure I'll get the hang of these things eventually, I've had to make liberal use of Data::Dumper to prevent the syntax from biting me.
Of course, since Unicode support is coming to Perl, and all of the Zapf Dingbats are in Unicode, we can just extend the syntax to...
If the number of shootings were to increase to a figure, such as 100, we would see more minority representation.
What do you mean, "if"? School systems in big cities have been dealing with violent students for years.
The media just treat it with a completely different spin. When a poor black teenager commits murder, it's considered a sign of the "problems of the underclass". Reporters consider crime so typical of poor black teenagers, that it's hardly even news.
On the other hand, when a middle-class white teenager commits murder, the murderer is considered deviant from, not typical of, the white middle-class community: "how could this happen here?"
I'm sort of afraid to ask this, but....why do these things seem to be centred around the US?
Maybe it has something to do with American suburbia. The culture of the stereotypical US suburb depends on:
cheap land outside cities
a tax system that favors homeowners
local government control of police departments, fire departments, schools, and zoning
gas taxes that are, by First World standards, extremely low
generous government subsidies for road-building
Because of these conditions, the following things developed over the past forty years or so:
Middle-class people can easily move out of central cities, buy houses in the suburbs, and rely on their cars to take them to work, shopping, etc.
Zoning laws prevent developers from building townhouses or apartment buildings in many suburban areas, so people below a certain economic level can't afford to move there.
Since local governments depend primarily on local property taxes for funding, wealthier suburbs can attract people with their well-funded public schools, well-maintained streets, and so on. Meanwhile, some central cities found their tax base leaking away, so they had trouble funding adequate school systems and police departments. This encouraged more middle-class urban residents to move to the suburbs.
When the US legislature and courts outlawed segregation in the 1960s, many whites moved from the cities to the suburbs, so that their children would go to an all-white suburban school rather than a racially integrated urban school.
These trends are beginning to reverse, but I think the average American suburb is more bland, er, homogeneous (ethnically and economically) than a large city in the US or in Europe. In a large, crowded city, a slightly "abnormal" kid can find friends and hangouts that match his or her interests and quirks. In the suburbs, it's much harder, especially for a kid who's too young to drive, so these kids are more likely to be at the mercy of their classmates.
If I remember correctly, MIT has only given out one honorary degree in its history, and that was to Winston Churchill.
Salman Rushdie was made an honorary visiting professor; as a columnist for the school paper quipped, this made Rushdie MIT's most inaccessible faculty member.
Re:I would much rather call it Linux
on
GNU Inside?
·
· Score: 1
Imagine a system where you can choose any kernel you want (BSD,Linux,IRIX,BeOS, or even NT) use any shell you want (bash, commad.com, KDE or what ever) use any networking layer you want (TCP/IP, ATM...). This is real choice not having to force ppl into a all or nothing situation...
Since the shell and networking layers have to communicate with the kernel, and selecting the right kernel API is an important part of OS design, writing such a system is not practical.
You might be interested in the OSKit, a modular collection of operating-system libraries. The OSKit was developed so that if you want to have an experimental OS with a few novel features, you (or your graduate students) don't have to write a whole OS from the ground up; just download the OSKit, rewrite a single module, compile it, and benchmark away.
don't call GPL "viral", call it "tit-for-tat"
on
BSD vs GPL
·
· Score: 5
A while ago, someone on/. posted a pointer to a Web page by David Rysdam -- I unfortunately didn't save the link, but the title was "Open Source as ESS". For those of you who are familiar with the iterated Prisoner's Dilemma (see Axelrod's The Evolution of Cooperation if you're not), I can summarize the article as follows:
Computer programs competing with each other are like players in an iterated Prisoner's Dilemma game.
Letting another program copy your code is like playing "Cooperate" in the game.
Refusing to let another program copy your code is like playing "Defect".
Closed-source programs are playing the "Always defect" strategy.
Programs with a BSD license are playing the "Always cooperate" strategy.
Programs with a GPL license are playing the "Tit for tat" strategy.
In simulations of the iterated Prisoner's Dilemma, tit-for-tat players usually do better in the long run than players who always cooperate or those who always defect; therefore, GPL programs are likely to displace BSD programs in the long run.
I have some reservations about Rysdam's model, but I would love to see more analyses of this type.
imagine Joe Blow bought the book and signed the contract saying he wouldn't distribute the book. Then say without his permission I photocopy the book. He didn't distribute the book and I now have a copy. I never signed any contract so the author can't legally do anything to punish/stop me.
The publisher could include a clause in the contract holding Joe Blow responsible for the cost of any illicit copies that could be traced to him.
Imagine what distributing The Phantom Menace would be like under such a legal regime. A pirate edition could cost studios immense $$$, and studios wouldn't be able to sue the pirates directly. So every single person with access to the reels of film would have to accept liability for illicit copies. If you didn't have much money, the studio wouldn't want you to have access to the reels at all, unless someone else (your employer or a bonding agency) were willing to accept liability on your behalf. Of course, this means that moderate-income people with access to the reels (e.g., projectionists and couriers) would be screened and personality-tested and watched up the wazoo, lest they expose their employers to massive lawsuits. Every reel, screenplay copy, etc., would have to be watermarked, so that illicit copies could be traced.
How much more would ticket prices cost, to cover all this security?
Compared with garden-variety contract law, copyrights restrict people's freedom to make money off of intellectual property.
For example: let's say I record a song, sell all my rights to it for $5,000, and then watch in dismay as my studio rakes in millions from it. Under section 203 of the Copyright Act, I can recover my copyright from the studio after thirty-five years. That's not "Internet time," but if my music is selling after all those years, it's better than nothing.
Another example: decades ago, many books and records were sold on the condition that the buyer not resell them. This way, publishers hoped, anyone who wanted a copy of a popular book or record would be forced to buy it new. These restrictions were thrown out by the courts.
There are many other such loopholes in copyright law; I hope one of the legal eagles here can provide more examples. Software comes with these absurd shrink-wrap licenses because commercial software companies want more control over their product than a plain copyright allows.
What would happen if conditional-sale contracts replaced copyright, and if US commercial law followed libertarian principles of sanctity of contract? We would see books and CDs decorated with legal restrictions, as publishers try to wring every last dime possible out of their audiences.
So if Roederick Long wants to object to copyright on libertarian grounds, his objection is totally backwards. He should be complaining that copyright law is unfairly permissive, not that it is unfairly restrictive.
One of the key issues in the court's decision, as someone else here pointed out, is that ITAR is exercising prior restraint on free speech.
If you use your Web site to pass top-secret military plans to the Russians, or libel somebody, or violate copyrights, then you can be sued or prosecuted. The First Amendment doesn't give you a blanket immunity for this kind of thing.
However, the government can't require you to pass all of your Web pages through a government censor to make sure that they contain nothing classified, libelous, or copyright-violating.
That's prior restraint. If a law involves prior restraint on expression, the courts tend to look on it with great suspicion.
It's hard enough for me to deal with a computer that follows a program in a perfectly logical and predictable way... even when I'm the one who wrote the program.
Why would I want a machine that's unpredictable by design?
Sigh, this is going to ramble, but actually Disney was in a bit of trouble a year or so ago when the copyright or whatever on Mickey Mouse expired and he would suddenly be in the public domain, and anyone could make profits on him. I forget how they solved the issue,
They passed the "Sonny Bono Copyright Term Extension Act of 1998", extending the term of copyright protection by another 20 years.
For more information and links (from a law professor opposed to the extension), follow this link.
(This is not the first time the Meta tags have come up in court. Playboy seems to have gone apeshit about the terms Playboy and playmate.)
And indeed, the court's decision says...
In Welles, the case most on point, Playboy sought to enjoin former Playmate of the Year Terri Welles ("Welles") from using "Playmate" or "Playboy" on her web site featuring photographs of herself. See 7 F. Supp. 2d at 1100. Welles's web site advertised the fact that she was a former Playmate of the Year, but minimized the use of Playboy's marks; it also contained numerous disclaimers stating that her site was neither endorsed by nor affiliated with Playboy. The district court found that Welles was using "Playboy" and "Playmate" not as trademarks, but rather as descriptive terms fairly and accurately describing her web page, and that her use of "Playboy" and "Playmate" in her web site's metatags was a permissible, good faith attempt to index the content of her web site. It accordingly concluded that her use was permissible under the trademark laws. See [7 F. Supp. 2d] at 1103-04.
He doesn't say, "Windows is ported to more hardware platforms in the marketplace than Linux is."
If someone tried to pin him down on the other, er, dubious statement,
[Linux] lacks an extraordinary number of features that you find in Windows, including transaction features, Web features, security features
he could wriggle out of it, either by defining "Linux" narrowly to exclude other free software available for it, or by finding some narrow set of "features" (e.g., Microsoft-sponsored protocols) that Linux doesn't have.
As Jakob Nielsen said, advertising doesn't work on the Web. He notes that click-through rates on banner ads were about one percent when he wrote his column in October 1997, and a year later, according to NetRatings, it was half that.
Some of the movers and shakers in the Internet world are working on systems to meter Internet usage on a per-user basis, just like telephone calls are metered. Once such systems are in place, the extra bandwidth that banner ads take up will cost the users money as well as download time (unless someone sets up an advertiser-supported "toll-free" site, but then the ads will really have to pull in the bucks). Sites like slashdot could still recover the cost of providing service, however, by setting up the Internet analog of "900 numbers": collecting micropayments through the Internet billing system.
Along the way Stearns debunks the mistaken idea of widespread panics among the European populace; at that time, the arrival of the year 1000 wasn't viewed as a momentous event, and most people used different calendars in which that year wasn't particularly special.
Richard Landes, a history professor at BU, debunks this debunking in his essay "Apocalyptic Expectations around the Year 1000". Landes is the co-founder and director of the Center for Millenial Studies at Boston University. They have a lot of interesting stuff on their Web site, including an invitation to become an "associate archivist" of Y2K.
Once you take into account the cost of buying the refrigeration equipment and the increase in your electric bill from running it, would it be cheaper to just buy a faster processor and run it at room temperature?
Yeah, I know, for some people it wouldn't be as much fun, but....
The important question... should surely be whether their message is heard and remembered, not whether the names of the people delivering the message are.
I think that for every grass-roots political movement, a few people become seen as people who speak for the movement as a whole; the general public treats every one of their statements as a message from the movement as a whole, even if not every statement would be endorsed by the grass roots.
Slashdot Mirror
Consider a live, naturally-evolved plant. It can't be too picky about the kind of soil it's planted in, the temperature of its surroundings, the amount of sunlight it gets each day, the amount of rain that falls on it, the chemical content of the rain, etc. It has to convert whatever resources it has available into forms of matter and energy that it can use for growth, reproduction, and defense against predators.
Since it's competing with other plants for survival, it has to make engineering trade-offs. For example, a plant may secrete a bad-tasting chemical, protecting it from predators -- but the raw materials for that chemical may restrict the plant to certain soil chemistries, or the extra energy necessary to produce it may restrict it to environments with a certain minimal level of sunlight.
By contrast, our computers are designed to "live" in a tightly controlled environment. If a CPU is running a binary program, and comes across an instruction that makes no sense, it doesn't have to try "digesting" the program into a more sensible form. If a network router gets a packet with a syntactically incorrect header, it doesn't have to send the packet any farther.
Security-related protocols add to the level of control; they are filters between sensitive and untrusted parts of a computer system. Since our computers "live" off of electricity, not programs, they can be as picky as we want them to be about what programs they execute and what permissions they execute under. Since our operating systems are designed by (occasionally) intelligent humans, not evolved by natural selection, humans can design better operating systems, in which security against untrusted code is a fundamental part of the architecture.
By comparison, Amazon.com, the poster child for the high-flying net.stock, has lost money every year since its inception. In 1999, according to their most recent 10-K, they had net sales of about $610M and a net loss of about $125M. Every year before that, their losses-to-sales ratio has been even higher.
Asides: (1) You can look up this kind of data at www.freedgar.com; they get their information from the electronic filings at the US Securities and Exchange Commission, the government agency that regulates the stock market. (2) Since the Bryar column now has a link to Slashdot, and since people are claiming that the author's mail server is slashdotted, I'll just respond here. (3) High-tech stocks in general seem too overvalued for my taste, so I'm not planning to buy any Red Hat stock.
Furthermore, this deal testifies to the strength of Perl in the marketplace. I'm sure Gates would be much happier if he could ignore Perl and focus corporate resources on marketing Visual Basic.
And if M$ feels that having a good Perl implementation will lure people from the Unix world to NT, they might actually put some competent people to work on the project. M$ Word was a decent word processor ... back when they were competing with WordPerfect.
In the US (I don't know about other places), if you buy a book and don't like it, you can return it to the bookstore for a refund. The bookstore then returns the book to the publisher, and the publisher gives the store credit for another purchase from that publisher. If the book is a mass-market paperback, the bookstore doesn't bother returning the whole book; it just returns the torn-off book cover.
This is an accountant's nightmare. The publisher can't predict how much profit a book has made until the returns bounce back. (On average, though, about half of all books sold are returned, so when you buy a book, you're really paying for the printing of two books.)
And as long as the publisher has books sitting in the warehouse, they're taxed at their current wholesale price, even if the publisher knows that a certain proportion of these copies are destined for Buck-A-Book.
And the people who own publishers don't simply want their companies to make money; they want them to make as much money as other businesses. A book that is likely to become gradually more popular over time, or a book that would have a small but devoted following, is therefore not a good investment for the publisher.
And most titles don't make money for their publishers; the profits from Stephen King and Robert Jordan novels cover the losses that their publishers suffer from promising but unsuccessful works by others.
Setting up tax-exempt charities to fund free software development is good for the donors' tax bill, but if the charity is paying programmers to write the software, the programmers' salaries are taxable.
On the other hand, if you give a cash gift to your favorite free-software programmer, you can't deduct the gift on your taxes, but the programmer doesn't have to declare it as income, either.
Back on the first hand, a gift, by definition, has no strings attached, so if you give a gift to your favorite free-software programmer, and that programmer takes the money and retires to Bermuda, you have no recourse.
- If the court is convinced that the defendant told the truth, the plaintiff loses.
- If a "public figure" brings a libel suit, it's not enough to prove that the defendant made a false and damaging statment; the plaintiff must also prove that the defendant either acted with malice or had negligent disregard for whether or not the statement was true.
- It's difficlut, if not impossible, for a plaintiff to win when the allegedly libelous work is a parody or a political commentary.
(IANALI know that some other countries' libel laws make it easier for plaintiffs to win, and I know that UK law is not as strict about free speech as the US Constitution is ... but I don't know details. Can any Brits or legal eagles on /. help us out here?
I've spent a bit of time beating my head against complex data structures in Perl, and while I'm sure I'll get the hang of these things eventually, I've had to make liberal use of Data::Dumper to prevent the syntax from biting me.
Of course, since Unicode support is coming to Perl, and all of the Zapf Dingbats are in Unicode, we can just extend the syntax to...
The media just treat it with a completely different spin. When a poor black teenager commits murder, it's considered a sign of the "problems of the underclass". Reporters consider crime so typical of poor black teenagers, that it's hardly even news.
On the other hand, when a middle-class white teenager commits murder, the murderer is considered deviant from, not typical of, the white middle-class community: "how could this happen here?"
- cheap land outside cities
- a tax system that favors homeowners
- local government control of police departments, fire departments, schools, and zoning
- gas taxes that are, by First World standards, extremely low
- generous government subsidies for road-building
Because of these conditions, the following things developed over the past forty years or so:- Middle-class people can easily move out of central cities, buy houses in the suburbs, and rely on their cars to take them to work, shopping, etc.
- Zoning laws prevent developers from building townhouses or apartment buildings in many suburban areas, so people below a certain economic level can't afford to move there.
- Since local governments depend primarily on local property taxes for funding, wealthier suburbs can attract people with their well-funded public schools, well-maintained streets, and so on. Meanwhile, some central cities found their tax base leaking away, so they had trouble funding adequate school systems and police departments. This encouraged more middle-class urban residents to move to the suburbs.
- When the US legislature and courts outlawed segregation in the 1960s, many whites moved from the cities to the suburbs, so that their children would go to an all-white suburban school rather than a racially integrated urban school.
These trends are beginning to reverse, but I think the average American suburb is more bland, er, homogeneous (ethnically and economically) than a large city in the US or in Europe. In a large, crowded city, a slightly "abnormal" kid can find friends and hangouts that match his or her interests and quirks. In the suburbs, it's much harder, especially for a kid who's too young to drive, so these kids are more likely to be at the mercy of their classmates.Salman Rushdie was made an honorary visiting professor; as a columnist for the school paper quipped, this made Rushdie MIT's most inaccessible faculty member.
You might be interested in the OSKit, a modular collection of operating-system libraries. The OSKit was developed so that if you want to have an experimental OS with a few novel features, you (or your graduate students) don't have to write a whole OS from the ground up; just download the OSKit, rewrite a single module, compile it, and benchmark away.
- Computer programs competing with each other are like players in an iterated Prisoner's Dilemma game.
- Letting another program copy your code is like playing "Cooperate" in the game.
- Refusing to let another program copy your code is like playing "Defect".
- Closed-source programs are playing the "Always defect" strategy.
- Programs with a BSD license are playing the "Always cooperate" strategy.
- Programs with a GPL license are playing the "Tit for tat" strategy.
- In simulations of the iterated Prisoner's Dilemma, tit-for-tat players usually do better in the long run than players who always cooperate or those who always defect; therefore, GPL programs are likely to displace BSD programs in the long run.
I have some reservations about Rysdam's model, but I would love to see more analyses of this type.Imagine what distributing The Phantom Menace would be like under such a legal regime. A pirate edition could cost studios immense $$$, and studios wouldn't be able to sue the pirates directly. So every single person with access to the reels of film would have to accept liability for illicit copies. If you didn't have much money, the studio wouldn't want you to have access to the reels at all, unless someone else (your employer or a bonding agency) were willing to accept liability on your behalf. Of course, this means that moderate-income people with access to the reels (e.g., projectionists and couriers) would be screened and personality-tested and watched up the wazoo, lest they expose their employers to massive lawsuits. Every reel, screenplay copy, etc., would have to be watermarked, so that illicit copies could be traced.
How much more would ticket prices cost, to cover all this security?
For example: let's say I record a song, sell all my rights to it for $5,000, and then watch in dismay as my studio rakes in millions from it. Under section 203 of the Copyright Act, I can recover my copyright from the studio after thirty-five years. That's not "Internet time," but if my music is selling after all those years, it's better than nothing.
Another example: decades ago, many books and records were sold on the condition that the buyer not resell them. This way, publishers hoped, anyone who wanted a copy of a popular book or record would be forced to buy it new. These restrictions were thrown out by the courts.
There are many other such loopholes in copyright law; I hope one of the legal eagles here can provide more examples. Software comes with these absurd shrink-wrap licenses because commercial software companies want more control over their product than a plain copyright allows.
What would happen if conditional-sale contracts replaced copyright, and if US commercial law followed libertarian principles of sanctity of contract? We would see books and CDs decorated with legal restrictions, as publishers try to wring every last dime possible out of their audiences.
So if Roederick Long wants to object to copyright on libertarian grounds, his objection is totally backwards. He should be complaining that copyright law is unfairly permissive, not that it is unfairly restrictive.
If you use your Web site to pass top-secret military plans to the Russians, or libel somebody, or violate copyrights, then you can be sued or prosecuted. The First Amendment doesn't give you a blanket immunity for this kind of thing.
However, the government can't require you to pass all of your Web pages through a government censor to make sure that they contain nothing classified, libelous, or copyright-violating.
That's prior restraint. If a law involves prior restraint on expression, the courts tend to look on it with great suspicion.
Why would I want a machine that's unpredictable by design?
For more information and links (from a law professor opposed to the extension), follow this link.
If someone tried to pin him down on the other, er, dubious statement,
he could wriggle out of it, either by defining "Linux" narrowly to exclude other free software available for it, or by finding some narrow set of "features" (e.g., Microsoft-sponsored protocols) that Linux doesn't have.Yeah, I know, for some people it wouldn't be as much fun, but....
What about TeX?