DeCSS gives people the ability to copy DVDs? What was keeping me from copying them before?
Of course, it was possible to copy them before. But DeCSS enables casual, one-click duplication (DVD->MPG->CDR or DVD->AVI->P2P).
The previous methods each had drawbacks, making them unattractive to random users
Make a 1:1 copy of the disc, encryption and all. You'd need a DVD burner, though, and even today's modern DVD-RW systems can't store the same 13 gigabytes that a factory-pressed disc can contain. And a few years ago, no consumer DVD writing was available. Organized criminals, of course, can rent a Chinese DVD factory for a few nights to make all the copies they need to sell on the sidewalks of Manhattan. DeCSS never stopped professionals.
Use an official hardware DVD player to create an analog video stream, which you then record to VHS or capture onto a PC. This approach requires a fair amount of time and effort for the encoder, as well as a tolerance for quality loss on the part of the recipient. DeCSS never stopped dedicated hobbyists.
Neither of those methods enables the fearsome Napster-scenario, where consumers trade copyrighted data with each other because it's almost harder not to share the files. But with DeCSS present, and 80 gigabyte harddrives standard-issue, it becomes very likely that a fully-automatic DVD redistribution software will spring up.
Although it's clear that shrinkwrap EULAs are legally meaningless (anything included in the box, either on paper, or click-thru in the installer), online EULAs may be more enforcable.
An EULA that came with the original distribution media is only an obstacle to installation, and installation is something you've already bought the right to do. There's no exchange of consideration, and there's no communication between the parties of the contract.
However, if you download additional materials, like the requisite security updates for Microsoft(tm) Windows(r), there may be an additional EULA to click through.
The publisher would have a stronger case to enforce that contract, because you recieved additional goods in exchange for clicking on it. The patches were not promised with the original purchase, so "consideration is exchanged". The claim is even stronger if the click-thru agreement is required before the download occurs.
Of course, if this argument succeeds, and is cemented in precedent, Microsoft(tm) would have innovative ways to boost revenue, if they ever hit a surprise cash-crunch in the far future. They could randomly apply different EULAs to individual patching customers. Most would be normal, but 1% might include promises to pay an additional $299 for the enhancements. The bill will arrive 30 days after the download.
(That scenario sounds crazy- but if EULAs are binding, there's nothing to stop it)
But yet, if you walk into CompUSA and say "Will you sell me Windows XP(tm) for $299?", the clerk will say "Yes". The signage hanging on the shelves and walls reinforces the idea that selling will occur. More importantly, the reciept says "Sold".
Either the vendors of computer software are committing fraud on a gargantuan scale, or you are being sold software.
(Software publishers wish to change this- why they include those EULA that are legally nonbinding, and why they've pushed US states to create laws making EULAs effective. Virginia, so far, has agreed)
However, the reason normal EULAs are meaningless is because no contract terms were presented before money and product were exchanged. So, they have no similarity with the agreement between SCO and IBM. It was presumably conducted with lawyers, signatures, and even handshakes.
As a Linux user (wanting to read a Linux article), I can't click through the Salon day-pass system. It's presented as a Flash(tm) file, version 4, and Macromedia's plugin only supports versions 5 and 6.
Additionally, today's ad is for WindowsXP. Funny that most of the users who can view the commericial are already Windows(r) customers.
There are other copiers whose operation doesn't resemble that of a Xerox machine, and they've never been called "xerox". Mimeographs, for example.
What does the word 'Unix' refer to? Servers? An OS? I never heard anyone with a bit of sense refer to ALL OSes as 'Unix.'
It refers, naturally, to all Unices. A group of operating systems providing extremely similar core interfaces. This includes AT&T and Berkeley products named Unix, as well as Solaris, Irix, HPUX, AIX, Xenix, FreeBSD, Darwin (aka MacOS X), Linux, and Minix. Some of the things on that list are UNIX(tm), others (those without backing from a deep-pocketed corporation) are merely Unix (sometimes written *nix, to emphasis the lack of trademark authority).
However, it is technically (not legally) accurate to describe any of them as "Unix", for software purposes. For example, if a program is known to run on Unix, then any recent version of any of those OSs will have a similar chance to let it work (after a recompile).
On a non-Unix OS, like Microsoft(tm) Windows(r), BeOS, or MacOS 9, the odds of the program functioning without conceptual re-arrangement are drastically lower.
Now I have heard people say 'Windows' when referring to an OS.
That's a new one on me. Could you provide an example? Something like "My Mac's Windows is OSX!" prehaps?
That clause has been toothless forever. Just look at the entire body of copyright law; the majority of the things it protects are emphatically neither science, nor "useful" arts.
And of course, the Eldered case already explained that the "limited Times" portion is also ignored.
3) You don't contact anyone, because you never even created a sample ("prototype") of the flavor. You just sit on your couch until an ad for a similar-sounding flavor comes on, then dig out the patent and start planning how to invest the new windfall.
That 3rd case is really what's going on in the situations mentioned in the article. These people aren't pursuing companies and offering them new ideas- they wait for someone else to really make the invention, and then attack them. Just look at patent 6,339,842, and see if there's any grain of an idea deserving protection. Calling Dennis Fernandez an "inventor" is an insult to the profession.
(However, calling him a "lawyer" is both accurate, and conventionally interpreted as insulting to him)
But the main thing here is that offensive patents help people with little cash the most.
Read the article. It states that getting a patent normally takes from $8000 to $15000. People with "little cash" can't invest that much in lottery tickets.
If an impoverished man genuinely invents something useful, he should be protected. $8000 in fees won't help there.
but if patents are awarded as they should, then even offensive patents benefit everybody
The very concept of "offensive patents", as described in the article, depends on the fact that patents are awarded much too broadly. You can patent something which you have no idea how to make yourself- you just have to gamble that someone will decide to try building a kind of product in the next two decades, so you can slam him for royalties. No technical knowledge is required- just look at Dennis Fernandez. Do you think he knows anything about television, video cameras, or the elaborate data networks and software that would be required to make TV watching a communal experience?
I doubt it. There's a lot of hard engineering work there, which he has no ablilty to pursue. But if someone does succeed at that challenging work, and further survives the marketing risk of actually finding customers for the service, then he'll be ready to sue the real trailblazers as soon as they finally get a revenue stream.
The supposely "original idea" he came up with was painfully obvious even back when Star Trek first aired. Thousands of people have probably thought of the same concept, but they didn't productize it because the technical barriers were too large. And they didn't patent it because they weren't greedy opportunists.
A patent is useless and counterproductive to the good of capitalism if no product comes out of it.
The patents as described in the article are either irrelevant to the creation of products, or an impairment. These patents certainly contribute nothing to creating products, but they impose a "rent" that might stop a commodity from being profitably sold. Offensive patents are a snare you lay to catch someone smarter- it takes more intelligence to create a good invention than to recognize when someone else is working on one.
the chances are much better that a small company will have the capital to file a patent but not ramp up to production
Arguments on the general truth of that statement are possible, but there's one area where it is most clearly wrong: software patents.
With software, there is no production to "ramp up"- the per-unit costs to sell software are neglible compared to any other industry. Yet to acquire software patents requires a stable of $140,000 lawyers to do the filing and searching (and to eventually sit in court defending claims that should've been invalid from the start).
Dead Sea Scrolls, Hammurabi's Laws, the Rosetta Stone
Only one of those can really be called "handwritten", as meant today. And it's the one that survived in worst shape. Pressing sticks into clay can even be called imprinting.
All that article says is that the "S/360 distribution" of Linux is assigned to the FSF. It mentions nothing about the abundant code IBM placed in the general (x86 et al) Linux tree. I don't know if the copyrights have since been assigned, but all the FSF copyright statements are in the system specific, ppc/m68k/arm areas. IBM has many more, primarily in linux/fs/jfs.
Actually, it matters a lot. Moral decline and all that.
The only laws regulating what you can and cannot say about something are consumer protection laws.
That's 100% wrong. More like 400%. There are dozens and dozens of laws specifying what you can say. Libel and slander for starters. Conspiracy. Negligence through inaction.
And, fraud can occur without money changing hands. (Most obviously, when trying for admission to a school on the strength of other's work)
FOSS
Been reading Mitre?
If they feel the need to leave older versions available, they can do so, and simply replace the relevant code for those versions.
Then they're not really the older versions, if they've been changed. Besides, this whole discussion springs from the idea that Linus would be concealing the fact that a certain old version was infringing. Those vendors won't know what to change, if he falsified the Changelogs.
Neither is fair use.
Wrong. Not 400% wrong, but 900% wrong or more. Fair use is absolutely a defense against infringing.
You cannot punish someone -- either by jail-time or fines -- for infringing on something which they had no practical way of finding out they were infringing on.
You can punish them by informing them of the infringement at the time of your choosing, and Cease & Desisting them from operating the computer systems running their day-to-day money-earning business operations. If that were to happen, and some high profile Linux adopter lost 3 days of revenue, the Free Software reputation would be badly tarnished.
In any event, SCO does not have the resources to sue individual users, nor even the many distributors of GNU/Linux.
If their case had some merit, lawyers would volunteer for the chance of a payoff ("contingency"). (Boies, their current lawyer, is very well regarded. But he's being paid, so his service doesn't imply the case is strong)
Also, you independently re-invented or re-implemented whatever patented process is being spoken of, so there's a good argument that you have just as much right to use it as does the person who originally patented it.
That's a sensible argument, but patent law is completely opposed. Even if you invent the process before the patenter, if you don't publish, he can get the patent and then bar you from using your own idea. (This has happened; "prior art" only counts if it's public)
In any event, this entire case is trumped up. SCO has provided no evidence to the public;
Yes. If they'd simply provide the filenames of some part of Linux they alledge to be infringing, they'd gain back a smidgen of credibility.
(They won't do that, because narrowing it down to a filename would enable the legions of interested parties to scour Changelogs and guess the name of the sinister inserter. Then the work to purge the bulk of Linux could begin. And SCO wants to keep Linux using-companies confused, so the threat of a sudden C&D shutdown hangs over them)
In my example, the corporation was distributing, releasing, and publishing the derivative works. By "making changes", I meant "distribute a derivative work".
The GPL only applies to distribution.
The GPL applies to copying. You must agree to the GPL when making a copy (modified or not), regardless of whether you distribute it. See the 3rd word of GPL section 3, and the 4th word of GPL section 4.
Lying is sometimes legal, but usually wrong. Lying regarding authorship of a work (essentially denying that previous versions of Linux were in fact infringing) may be legally dangerous.
As a free, open project, they're free to call their updates whatever they like.
Not necessarily. No one is allowed to evade the laws of his nation, of course. See this example of laws tampering with the Linux changelog.
It will not affect distributors of GNU/Linux
It will effect distributors, because if they leave old versions available, they could be liable for infringement.
and no obligation to spend money to find out.
Ignorance of a copyright is not a defense for infringing. (Demonstrating you were ignorant, and not willfully infringing, will reduce damages awarded, but is not a full protection. It will stop you from getting jail time, though!)
Several areas of US Intellectual Property law obligate the public to make unfeasible efforts to determine they are uninfringing- most severly in the area of patents.
but doesn't this mean SCO have released their IP under GPL?
That has been argued. But courts won't hold someone responsible for a contract made in response to fraudulent information from the other party.
Besides, the GPL says that "you must apply GPL when you make changes". SCO alleges that they didn't make changes to put their code in Linux- they say it was already there.
Extreme example: Suppose a company (Sun) hosts a free software archive on the web. A hacker breaks into their network, and copies proprietary code into a GPL package on their public server. When someone downloads that, Sun has distributed a GPL program containing their own changes. Have their really put their entire codebase under GPL? Of course not, they didn't know it was in there, and weren't making a willful decision. Same as with SCO: "We didn't know it was there"
It's absolutely not copyright GNU (GNU is software, and can't hold copyrights). It's also not copyright FSF (the GNU organization), because they have their own non-Linux kernel to work on.
Instead, the Linux kernel is copyright the many individuals who contribute to it. Most prominently, Linus Torvalds.
and give it a changelog entry of something like "re-write portions to improve compiler optimizations
That would be a lie. It's dishonest, and potentially some (international) lawyer could cast it as illegal fraud.
Besides, it wouldn't work. The Linux kernel is a high-profile, widely scrutinized codebase. A message like that would inspire some GCC maintainer to check on exactly how the changes effected optimization- and 10 minutes of work would reveal that they do nothing of the sort.
A brief comment like "Update documentation and re-factor" would have a better chance of going undiscovered (especially if it turns out that many of SCO's "identical code" is in fact just comments)
Redhat might migrate to the new clean version of the tree and be forced to discontinue support for the older versions.
They already don't support older versions (unless prehaps a customer has negotiated a special contract).
For example, RedHat 6.2 came out in 2000, and is unsupported. You cannot install it on a system with a Pentium4 chip, even though it would take a minor patch to the installer to allow this.
Redhat abandons old versions faster than Microsoft does. (Of course, since they allow users to upgrade for $0, they could argue that it's not as bad in their case)
Anyhow, if a "new, legal Linux" is created, Redhat could just back-port those changes into their old products. Then they could continue supporting their old releases by saying "Before we can help you with your problem, sir, you should take these patches and upgrade to version 6.2.1 or 7.3.1"
it must be possible to relink the product with later libraries
This can also be accomplished by giving customers source code to your program. When you do that, the source code can still be proprietary. Just because the customer got one copy of the source from you, doesn't mean it has to be GLPed and redistributable.
Under LGPL section 6, there are other ways to comply with the "ability to re-link to new versions of LGPL libraries". Many corporations would prefer option 6c, which only requires them to supply object files if requested within 3 years (after all, few customers will have the technical knowhow to actually wish to make that request)
Ummm, no, the protocol overhead I was talking about was the packet headers for each TCP packet.
The header is 56 bytes, packets are 1500. (Usually. Weird networks can choose smaller) That's < 1% overhead. Assuming you send data larger than a packet, which is usually the case if you bother to measure bandwidth.
1 start bit, 8 data bits, 1 stop bit possibly?
Issues like stop bits are already taken care of by TCP/IP. Or rather, TCP/IP requires the network it runs on top of to provide any needed stop bits. Ethernet, for example, uses 1 stop bit per data bit. That's right, 50% of the bandwidth goes to measuring the length of the data being sent. (That amount is deducted before they label 10Mbps on the card). Understandably, if the underlying network has so much control structure, higher level protocols shouldn't need to add much more.
In any case, dividing by 10 is probably the right way to measure bytes per second on his connection. I was talking about the additional overhead that is there for the tcp protocol itself.
Lets talk about that. If you work the numbers from the TCP protocol, the 20-25% quantity shows up very naturally.
The protocol changes the amount of bandwidth it attempts to use over time (because it's impossible for the sending computer to know the bandwidth all along the way to the remote box). Assuming there is no other cause of packet loss besides exceeding available bandwidth, then TCP will increase the bandwidth it uses linearly until hitting the max. Then it halves it, and starts increasing again.
So at steady state, TCP is repeatedly dipping back to 50% utilization, climbing up to 100%, and starting over. Which means that on average, it will be sending data at 75% of the total of the network. That 25% is left behind was intentional, so that other users have room to start up new data streams without (immediately) colliding with you.
Of course, 75% utilization is absolute best case. If there are any other users of the network, or even random electron oddities, TCP will think the bandwidth is less than it really is.
were given their old land back as much as possible.
That would mean there'd be no room for any more Jews to live in Jerusalem today than did in 1947. (You may recall a vast Muslim majority from around 800-1945) Obivously, the Jews ruling Israel would never let that happen. The only reason they settled in Israel, as opposed to say Baja California, was to possess the Holy City.
But anyway, lets suppose that the the Palestinians were allowed full rights as citizens of Israel, and that they decided to live in their ghetto slums of their own free will.
Why, if they're allowed to be citizens, can't they vote for the Prime Minister of the nation where they were born and live?
(Answer: because their vote, in concert with guilt-ridden liberal Jews, would be enough to split Palestine into a separate nation. Thus, they're only allowed to vote after changing their religion)
All recent-model Zaures are either bulky or non-expandable. Mostly bulky.
Sliding an 802.11b card into the CF slot of a Zaurus only adds a little to the overall size (if you choose the right card), but the Zaurus itself is already large. And if you want a longer range connection, you either need a different card with a bigger antenna, or you'll have to plug it into a cellphone.
For sale in the US are the 5500 and 5600, approximately the same size. Each has is as wide as a regular PocketPC device (like the HP Ipaq for example), and has the same size screen. But to allow for a pull-out hardware keyboard, it's 3 centimeters taller. That's a small increase which can make a big difference fitting something into your pocket. (And of course, PocketPCs are bigger than Palm/Handspring/Sony PDAs with Palm OS)
The current article covers even bigger systems that are only sold in Japan. I don't think you'd want one of those. However, in Japan there's also a Zaurus model (newer than the 5500, older than the other recent models) which has no hardware keyboard at all. It looks the same size as an Ipaq, and like an Ipaq, there is no CompactFlash slot for an ethernet or 802.11B card. There is a SecureDigital slot, and SecureDigital bluetooth adapters are on the market, but I haven't heard of linux drivers for them.
I, for one, will be glad to see the M16 go. While it performs well under 'ideal' conditions, its performance under real-world conditions (inadequate lubrication, dust and sand in the magazines, etc...) leaves much to be desired.
But can you expect the new XM29 ("OICW") weapon to be any better? It might accelerate the M16's "textbook engineering" tendencies, and be even more complicated and damage-prone.
The thing has got two barrels and one trigger. Computer chips inside will decide which round to ignite when you pull the trigger. Do you have to replace the batteries when changing a clip? And the recent prototypes have been 18 lbs; hard to carry spare parts, but easy to drop your weapon on a rock the first time you dive for cover.
Fortunately, less than half of infantrymen will be issued XM29s, so there will be someone to provide covering fire when they run back to the supply unit.
Slashdot Mirror
Of course, it was possible to copy them before. But DeCSS enables casual, one-click duplication (DVD->MPG->CDR or DVD->AVI->P2P).
The previous methods each had drawbacks, making them unattractive to random users
- Make a 1:1 copy of the disc, encryption and all. You'd need a DVD burner, though, and even today's modern DVD-RW systems can't store the same 13 gigabytes that a factory-pressed disc can contain. And a few years ago, no consumer DVD writing was available. Organized criminals, of course, can rent a Chinese DVD factory for a few nights to make all the copies they need to sell on the sidewalks of Manhattan. DeCSS never stopped professionals.
- Use an official hardware DVD player to create an analog video stream, which you then record to VHS or capture onto a PC. This approach requires a fair amount of time and effort for the encoder, as well as a tolerance for quality loss on the part of the recipient. DeCSS never stopped dedicated hobbyists.
Neither of those methods enables the fearsome Napster-scenario, where consumers trade copyrighted data with each other because it's almost harder not to share the files. But with DeCSS present, and 80 gigabyte harddrives standard-issue, it becomes very likely that a fully-automatic DVD redistribution software will spring up.Although it's clear that shrinkwrap EULAs are legally meaningless (anything included in the box, either on paper, or click-thru in the installer), online EULAs may be more enforcable.
An EULA that came with the original distribution media is only an obstacle to installation, and installation is something you've already bought the right to do. There's no exchange of consideration, and there's no communication between the parties of the contract.
However, if you download additional materials, like the requisite security updates for Microsoft(tm) Windows(r), there may be an additional EULA to click through.
The publisher would have a stronger case to enforce that contract, because you recieved additional goods in exchange for clicking on it. The patches were not promised with the original purchase, so "consideration is exchanged". The claim is even stronger if the click-thru agreement is required before the download occurs.
Of course, if this argument succeeds, and is cemented in precedent, Microsoft(tm) would have innovative ways to boost revenue, if they ever hit a surprise cash-crunch in the far future. They could randomly apply different EULAs to individual patching customers. Most would be normal, but 1% might include promises to pay an additional $299 for the enhancements. The bill will arrive 30 days after the download.
(That scenario sounds crazy- but if EULAs are binding, there's nothing to stop it)
But yet, if you walk into CompUSA and say "Will you sell me Windows XP(tm) for $299?", the clerk will say "Yes". The signage hanging on the shelves and walls reinforces the idea that selling will occur. More importantly, the reciept says "Sold".
Either the vendors of computer software are committing fraud on a gargantuan scale, or you are being sold software.
(Software publishers wish to change this- why they include those EULA that are legally nonbinding, and why they've pushed US states to create laws making EULAs effective. Virginia, so far, has agreed)
However, the reason normal EULAs are meaningless is because no contract terms were presented before money and product were exchanged. So, they have no similarity with the agreement between SCO and IBM. It was presumably conducted with lawyers, signatures, and even handshakes.
As a Linux user (wanting to read a Linux article), I can't click through the Salon day-pass system. It's presented as a Flash(tm) file, version 4, and Macromedia's plugin only supports versions 5 and 6.
Additionally, today's ad is for WindowsXP. Funny that most of the users who can view the commericial are already Windows(r) customers.
the name become a generic phrase for ALL copiers
There are other copiers whose operation doesn't resemble that of a Xerox machine, and they've never been called "xerox". Mimeographs, for example.
What does the word 'Unix' refer to? Servers? An OS? I never heard anyone with a bit of sense refer to ALL OSes as 'Unix.'
It refers, naturally, to all Unices. A group of operating systems providing extremely similar core interfaces. This includes AT&T and Berkeley products named Unix, as well as Solaris, Irix, HPUX, AIX, Xenix, FreeBSD, Darwin (aka MacOS X), Linux, and Minix. Some of the things on that list are UNIX(tm), others (those without backing from a deep-pocketed corporation) are merely Unix (sometimes written *nix, to emphasis the lack of trademark authority).
However, it is technically (not legally) accurate to describe any of them as "Unix", for software purposes. For example, if a program is known to run on Unix, then any recent version of any of those OSs will have a similar chance to let it work (after a recompile).
On a non-Unix OS, like Microsoft(tm) Windows(r), BeOS, or MacOS 9, the odds of the program functioning without conceptual re-arrangement are drastically lower.
Now I have heard people say 'Windows' when referring to an OS.
That's a new one on me. Could you provide an example? Something like "My Mac's Windows is OSX!" prehaps?
That clause has been toothless forever. Just look at the entire body of copyright law; the majority of the things it protects are emphatically neither science, nor "useful" arts.
And of course, the Eldered case already explained that the "limited Times" portion is also ignored.
3) You don't contact anyone, because you never even created a sample ("prototype") of the flavor. You just sit on your couch until an ad for a similar-sounding flavor comes on, then dig out the patent and start planning how to invest the new windfall.
That 3rd case is really what's going on in the situations mentioned in the article. These people aren't pursuing companies and offering them new ideas- they wait for someone else to really make the invention, and then attack them. Just look at patent 6,339,842, and see if there's any grain of an idea deserving protection. Calling Dennis Fernandez an "inventor" is an insult to the profession.
(However, calling him a "lawyer" is both accurate, and conventionally interpreted as insulting to him)
But the main thing here is that offensive patents help people with little cash the most.
Read the article. It states that getting a patent normally takes from $8000 to $15000. People with "little cash" can't invest that much in lottery tickets.
If an impoverished man genuinely invents something useful, he should be protected. $8000 in fees won't help there.
but if patents are awarded as they should, then even offensive patents benefit everybody
The very concept of "offensive patents", as described in the article, depends on the fact that patents are awarded much too broadly. You can patent something which you have no idea how to make yourself- you just have to gamble that someone will decide to try building a kind of product in the next two decades, so you can slam him for royalties. No technical knowledge is required- just look at Dennis Fernandez. Do you think he knows anything about television, video cameras, or the elaborate data networks and software that would be required to make TV watching a communal experience?
I doubt it. There's a lot of hard engineering work there, which he has no ablilty to pursue. But if someone does succeed at that challenging work, and further survives the marketing risk of actually finding customers for the service, then he'll be ready to sue the real trailblazers as soon as they finally get a revenue stream.
The supposely "original idea" he came up with was painfully obvious even back when Star Trek first aired. Thousands of people have probably thought of the same concept, but they didn't productize it because the technical barriers were too large. And they didn't patent it because they weren't greedy opportunists.
A patent is useless and counterproductive to the good of capitalism if no product comes out of it.
The patents as described in the article are either irrelevant to the creation of products, or an impairment. These patents certainly contribute nothing to creating products, but they impose a "rent" that might stop a commodity from being profitably sold. Offensive patents are a snare you lay to catch someone smarter- it takes more intelligence to create a good invention than to recognize when someone else is working on one.
the chances are much better that a small company will have the capital to file a patent but not ramp up to production
Arguments on the general truth of that statement are possible, but there's one area where it is most clearly wrong: software patents.
With software, there is no production to "ramp up"- the per-unit costs to sell software are neglible compared to any other industry. Yet to acquire software patents requires a stable of $140,000 lawyers to do the filing and searching (and to eventually sit in court defending claims that should've been invalid from the start).
Dead Sea Scrolls, Hammurabi's Laws, the Rosetta Stone
Only one of those can really be called "handwritten", as meant today. And it's the one that survived in worst shape. Pressing sticks into clay can even be called imprinting.
All that article says is that the "S/360 distribution" of Linux is assigned to the FSF. It mentions nothing about the abundant code IBM placed in the general (x86 et al) Linux tree. I don't know if the copyrights have since been assigned, but all the FSF copyright statements are in the system specific, ppc/m68k/arm areas. IBM has many more, primarily in linux/fs/jfs.
And which clause is that, exactly? (The words "privacy" and "private" are not part of the GPL)
When you use fair-use in court, you plead guilty to copyright infringement, but say that your infringement was covered under fair-use;
... ... is not an infringement"
No... As I already told you.
Check the law, if you don't believe me. It says "the fair use of a copyrighted work
Whether something is wrong or not is irrelevant.
Actually, it matters a lot. Moral decline and all that.
The only laws regulating what you can and cannot say about something are consumer protection laws.
That's 100% wrong. More like 400%. There are dozens and dozens of laws specifying what you can say. Libel and slander for starters. Conspiracy. Negligence through inaction.
And, fraud can occur without money changing hands. (Most obviously, when trying for admission to a school on the strength of other's work)
FOSS
Been reading Mitre?
If they feel the need to leave older versions available, they can do so, and simply replace the relevant code for those versions.
Then they're not really the older versions, if they've been changed. Besides, this whole discussion springs from the idea that Linus would be concealing the fact that a certain old version was infringing. Those vendors won't know what to change, if he falsified the Changelogs.
Neither is fair use.
Wrong. Not 400% wrong, but 900% wrong or more. Fair use is absolutely a defense against infringing.
You cannot punish someone -- either by jail-time or fines -- for infringing on something which they had no practical way of finding out they were infringing on.
You can punish them by informing them of the infringement at the time of your choosing, and Cease & Desisting them from operating the computer systems running their day-to-day money-earning business operations. If that were to happen, and some high profile Linux adopter lost 3 days of revenue, the Free Software reputation would be badly tarnished.
In any event, SCO does not have the resources to sue individual users, nor even the many distributors of GNU/Linux.
If their case had some merit, lawyers would volunteer for the chance of a payoff ("contingency"). (Boies, their current lawyer, is very well regarded. But he's being paid, so his service doesn't imply the case is strong)
Also, you independently re-invented or re-implemented whatever patented process is being spoken of, so there's a good argument that you have just as much right to use it as does the person who originally patented it.
That's a sensible argument, but patent law is completely opposed. Even if you invent the process before the patenter, if you don't publish, he can get the patent and then bar you from using your own idea. (This has happened; "prior art" only counts if it's public)
In any event, this entire case is trumped up. SCO has provided no evidence to the public;
Yes. If they'd simply provide the filenames of some part of Linux they alledge to be infringing, they'd gain back a smidgen of credibility.
(They won't do that, because narrowing it down to a filename would enable the legions of interested parties to scour Changelogs and guess the name of the sinister inserter. Then the work to purge the bulk of Linux could begin. And SCO wants to keep Linux using-companies confused, so the threat of a sudden C&D shutdown hangs over them)
In my example, the corporation was distributing, releasing, and publishing the derivative works. By "making changes", I meant "distribute a derivative work".
The GPL only applies to distribution.
The GPL applies to copying. You must agree to the GPL when making a copy (modified or not), regardless of whether you distribute it. See the 3rd word of GPL section 3, and the 4th word of GPL section 4.
Lying is sometimes legal, but usually wrong. Lying regarding authorship of a work (essentially denying that previous versions of Linux were in fact infringing) may be legally dangerous.
As a free, open project, they're free to call their updates whatever they like.
Not necessarily. No one is allowed to evade the laws of his nation, of course. See this example of laws tampering with the Linux changelog.
It will not affect distributors of GNU/Linux
It will effect distributors, because if they leave old versions available, they could be liable for infringement.
and no obligation to spend money to find out.
Ignorance of a copyright is not a defense for infringing. (Demonstrating you were ignorant, and not willfully infringing, will reduce damages awarded, but is not a full protection. It will stop you from getting jail time, though!)
Several areas of US Intellectual Property law obligate the public to make unfeasible efforts to determine they are uninfringing- most severly in the area of patents.
but doesn't this mean SCO have released their IP under GPL?
That has been argued. But courts won't hold someone responsible for a contract made in response to fraudulent information from the other party.
Besides, the GPL says that "you must apply GPL when you make changes". SCO alleges that they didn't make changes to put their code in Linux- they say it was already there.
Extreme example: Suppose a company (Sun) hosts a free software archive on the web. A hacker breaks into their network, and copies proprietary code into a GPL package on their public server. When someone downloads that, Sun has distributed a GPL program containing their own changes. Have their really put their entire codebase under GPL? Of course not, they didn't know it was in there, and weren't making a willful decision. Same as with SCO: "We didn't know it was there"
It's absolutely not copyright GNU (GNU is software, and can't hold copyrights). It's also not copyright FSF (the GNU organization), because they have their own non-Linux kernel to work on.
Instead, the Linux kernel is copyright the many individuals who contribute to it. Most prominently, Linus Torvalds.
and give it a changelog entry of something like "re-write portions to improve compiler optimizations
That would be a lie. It's dishonest, and potentially some (international) lawyer could cast it as illegal fraud.
Besides, it wouldn't work. The Linux kernel is a high-profile, widely scrutinized codebase. A message like that would inspire some GCC maintainer to check on exactly how the changes effected optimization- and 10 minutes of work would reveal that they do nothing of the sort.
A brief comment like "Update documentation and re-factor" would have a better chance of going undiscovered (especially if it turns out that many of SCO's "identical code" is in fact just comments)
Redhat might migrate to the new clean version of the tree and be forced to discontinue support for the older versions.
They already don't support older versions (unless prehaps a customer has negotiated a special contract).
For example, RedHat 6.2 came out in 2000, and is unsupported. You cannot install it on a system with a Pentium4 chip, even though it would take a minor patch to the installer to allow this.
Redhat abandons old versions faster than Microsoft does. (Of course, since they allow users to upgrade for $0, they could argue that it's not as bad in their case)
Anyhow, if a "new, legal Linux" is created, Redhat could just back-port those changes into their old products. Then they could continue supporting their old releases by saying "Before we can help you with your problem, sir, you should take these patches and upgrade to version 6.2.1 or 7.3.1"
Crime >>>> stealing, fraud, infringement, murder, kidnapping
it must be possible to relink the product with later libraries
This can also be accomplished by giving customers source code to your program. When you do that, the source code can still be proprietary. Just because the customer got one copy of the source from you, doesn't mean it has to be GLPed and redistributable.
Under LGPL section 6, there are other ways to comply with the "ability to re-link to new versions of LGPL libraries". Many corporations would prefer option 6c, which only requires them to supply object files if requested within 3 years (after all, few customers will have the technical knowhow to actually wish to make that request)
Ummm, no, the protocol overhead I was talking about was the packet headers for each TCP packet.
The header is 56 bytes, packets are 1500. (Usually. Weird networks can choose smaller) That's < 1% overhead. Assuming you send data larger than a packet, which is usually the case if you bother to measure bandwidth.
1 start bit, 8 data bits, 1 stop bit possibly?
Issues like stop bits are already taken care of by TCP/IP. Or rather, TCP/IP requires the network it runs on top of to provide any needed stop bits. Ethernet, for example, uses 1 stop bit per data bit. That's right, 50% of the bandwidth goes to measuring the length of the data being sent. (That amount is deducted before they label 10Mbps on the card). Understandably, if the underlying network has so much control structure, higher level protocols shouldn't need to add much more.
In any case, dividing by 10 is probably the right way to measure bytes per second on his connection. I was talking about the additional overhead that is there for the tcp protocol itself.
Lets talk about that. If you work the numbers from the TCP protocol, the 20-25% quantity shows up very naturally.
The protocol changes the amount of bandwidth it attempts to use over time (because it's impossible for the sending computer to know the bandwidth all along the way to the remote box). Assuming there is no other cause of packet loss besides exceeding available bandwidth, then TCP will increase the bandwidth it uses linearly until hitting the max. Then it halves it, and starts increasing again.
So at steady state, TCP is repeatedly dipping back to 50% utilization, climbing up to 100%, and starting over. Which means that on average, it will be sending data at 75% of the total of the network. That 25% is left behind was intentional, so that other users have room to start up new data streams without (immediately) colliding with you.
Of course, 75% utilization is absolute best case. If there are any other users of the network, or even random electron oddities, TCP will think the bandwidth is less than it really is.
were given their old land back as much as possible.
That would mean there'd be no room for any more Jews to live in Jerusalem today than did in 1947. (You may recall a vast Muslim majority from around 800-1945) Obivously, the Jews ruling Israel would never let that happen. The only reason they settled in Israel, as opposed to say Baja California, was to possess the Holy City.
But anyway, lets suppose that the the Palestinians were allowed full rights as citizens of Israel, and that they decided to live in their ghetto slums of their own free will.
Why, if they're allowed to be citizens, can't they vote for the Prime Minister of the nation where they were born and live?
(Answer: because their vote, in concert with guilt-ridden liberal Jews, would be enough to split Palestine into a separate nation. Thus, they're only allowed to vote after changing their religion)
All recent-model Zaures are either bulky or non-expandable. Mostly bulky.
Sliding an 802.11b card into the CF slot of a Zaurus only adds a little to the overall size (if you choose the right card), but the Zaurus itself is already large. And if you want a longer range connection, you either need a different card with a bigger antenna, or you'll have to plug it into a cellphone.
For sale in the US are the 5500 and 5600, approximately the same size. Each has is as wide as a regular PocketPC device (like the HP Ipaq for example), and has the same size screen. But to allow for a pull-out hardware keyboard, it's 3 centimeters taller. That's a small increase which can make a big difference fitting something into your pocket. (And of course, PocketPCs are bigger than Palm/Handspring/Sony PDAs with Palm OS)
The current article covers even bigger systems that are only sold in Japan. I don't think you'd want one of those. However, in Japan there's also a Zaurus model (newer than the 5500, older than the other recent models) which has no hardware keyboard at all. It looks the same size as an Ipaq, and like an Ipaq, there is no CompactFlash slot for an ethernet or 802.11B card. There is a SecureDigital slot, and SecureDigital bluetooth adapters are on the market, but I haven't heard of linux drivers for them.
I, for one, will be glad to see the M16 go. While it performs well under 'ideal' conditions, its performance under real-world conditions (inadequate lubrication, dust and sand in the magazines, etc...) leaves much to be desired.
But can you expect the new XM29 ("OICW") weapon to be any better? It might accelerate the M16's "textbook engineering" tendencies, and be even more complicated and damage-prone.
The thing has got two barrels and one trigger. Computer chips inside will decide which round to ignite when you pull the trigger. Do you have to replace the batteries when changing a clip? And the recent prototypes have been 18 lbs; hard to carry spare parts, but easy to drop your weapon on a rock the first time you dive for cover.
Fortunately, less than half of infantrymen will be issued XM29s, so there will be someone to provide covering fire when they run back to the supply unit.