If I invest and the government uses my "vote" that something will happen to prevent it from happening, I lose my money, because the event didn't occur.
Do you really think that's likely? I mean, seriously, despite how funny that comment sounds - a large number of people will be using this system. A lot of them are going to have all kinds of weird contracts open at any one time. One or two people moving in a certain direction is not going to set any alarm bells ringing. Its if a whole load of people suddenly move a lot of money in a particular direction that they'll get worried.
If the event actually happens and I have predicted it, I get my phones tapped, a visit from the FBI, probable interrogation and possible imprisionment for having "terrorist ties." What exactly is the incentive for an investor to enter this market?
And a few weeks later the government having run their extensive background checks on you decide that you aren't a terrorist, couldn't possibly have had anything to do with the incident, and that they have no evidence whatsoever to hold you with and release you. You gain your $200,000 or whatever you made. It's also very unlikely that you'll be investigated more than briefly, because a whole load of other people will have made the same contract you did, and most of those'll be innocent too.
You see, this _isn't_ actually about catching terrorists, however appealing that might sound. They may get one or two particularly stupid ones, but I doubt it'll add up to much in that direction. The real value is that it will help iron out the economic costs of terrorism.
Take two situations.
1. You run a travel company which specialises in holidays in Israel. You make a fair sum of money off of this, but you realise that any terrorism in the area puts off large numbers of your clients. Through past experience, you know that a major terrorist attack costs you $15,000 in lost business. However, due to the ratios in the market, you can place a contract which only costs you $10 per month to keep open which will compensate you for this cost in the event that this does happen. It'll give you peace of mind, so you do it - its definitely worthwhile.
Of course, because this is a derivatives market, somebody has to be offering to make those contracts with you, which brings us on to situation 2...
2. You're a defense contractor. Global unrest is generally good for you because when middle eastern countries are more nervous about terrorist threats they tend to buy more up-to-date equipment for their armies. But you don't like the fact that you feel like a celebration every time somebody blows up a car outside an embassy. So you've been looking for years for a way of balancing the good times and the bad, and here it is. There are hundreds of companies who are willing to pay a small price every so often in exchange for a bigger payout whenever there is a terrorist attack. You look at your books, work out what you can offer whenever anything like that happens, and feed your offer into the market... it comes back telling you that it has sold a number of contracts on the terms you suggested. Now you have a steady stream of income rather than one that jumps around all the time.
Good, isn't it?
Of course, it might not work in practice (I'm thinking there might not be enough people in category 2 to fulfill all the contracts that category 1 people would want), but the theory is excellent.
Re:Which to choose for DBs?
on
Reiser4 Benchmarks
·
· Score: 2, Informative
There's little advantage in passing every byte through two seperate journal
1. If you're using a filesystem for databases, then the chances are you're keeping other stuff besides the database on them. Otherwise, you'd use a raw partition.
2. I believe that by default data changes to files aren't journalled in either of the systems that were being compared; only filesystem metadata changes. This doesn't cost much in terms of resources, particularly if the metadata rarely changes (as I suspect it doesn't on oracle data files, which I assume are fixed size allocations that are opened when the server is started and never closed...).
Sadly, the poetry evolved to the point where it attracted a predator's attention, and now it's gone extinct.
This is why I'm against broadcasting our presence to the stars.
NEWSFLASH
INTERGALACTIC NEWS - MILKY WAY, ALPHA QUADRANT
Sentient species from all around the galaxy were stunned today by the news of the total annihilation of all life on Earth.
Earth, a small blue green planet orbiting an uninteresting single star towards the outer edge of the civilised portion of the galaxy, first came to the attention of scientists only a handful of years ago. "They started by broadcasting simple messages between themselves," said Humanologist Urk-Thr'qar of Threep University. "They didn't seem aware of our presence. But then they started sending messages that seemed directed at us. We had to investigate further."
But it was one of the messages transmitted internally within the planet that caused their eventual downfall. A human, it seems, had an interesting and imaginative new idea. Unfortunately, the actual idea has been lost forever due to what happened next.
A message was posted on the intergalactic network news site, slashdot.galaxy. The message read simply: "Hey, check out Earth. They're not as dumb as we thought," and included a set of hyperspatial co-ordinates for Earth.
Within seconds, the planet was surrounded by a cloud of interplanetary space ships. Only hours later, the ships had all departed, taking with them every scrap of life that existed on the planet.
Late arrivals were disappointed to find the planet barren.
"Huh!" said one three-eyed creature from Ph'tarra. "I come all the way over here and the entire planet's slashdotted. You'd have thought they would have had the courtesy to lay on extra bioreplicators to at least attempt to handle the extra specimen collectors they were bound to receive."
One lucky specimen collector remarked "Well, I mean I got myself a specimen, but it was only a small quadroped. I don't think its even house trained. But I'll put up a copy on GeneTorrent if you like."
I think nasm (which I was working on a lot at the time) made it as far as redhat in a 0.3something release. Of course we jumped straight over all the intervening versions to 0.9, cos it looked like we were nearly ready for a 'proper' release.
It might happen one of these decades, too...
Seriously, though, the only real way of getting your software into distributions is to get it used. With nasm, myself and Simon (the primary author) spent many many hours hanging around relevant newsgroups, letting people know that it was available, making sure they knew its advantages over other similar packages, etc. After a few months of that we had maybe a few hundred regular users. That was when we started hearing from redhat & debian (primarily that they didn't like our license and could we change it a little bit... if you want my advice, make sure you're either gpl or lgpl, it saves a lot of hassle in the long run...).
Is there any case law that states that dynamic linking creates a derivative work?
A few weeks back, there was a post about some documents from an old Unix Systems Labs copyright infringement case that people might have considered interesting in relation to the SCO case.
In one of these documents (from the BSD camp) there was a reference to a precedent which (according to that document) meant that merely copying interface information would not cause a work to be a derivative. I don't know how much weight that holds though, or even if the decision was later overturned....
One thing that interests me is your assertion that OLE/DCOM is slower than XML based message passing.
I don't think he actually said that. Looking through the message again, he states that.NET Remoting (something I've never looked at in detail, I assume its somewhat like Java RMI?) is slower than rolling your own XML based interface, and then compares it with OLE (the implication that I infer being that rolling your own system with similar features to OLE is likely to produce something faster than OLE itself).
Dictionaries contain false entries intended to serve as markers and preserve the collection copyright
I can understand maps, but... dictionaries? I don't think this would work. You see, there are quite a lot of people who will, from time to time, just pick up a dictionary and browse. Try to find interesting words that they've never heard of before. So, if a dictionary had (for example) the entry 'Grig (n) the piece of a ringpull on a drink can that gets pushed inside the can upon opening' someone will probably notice this. Then they'll be talking to somebody at some point and an opportunity to use this fabulous word 'grig' will arise. So they will. And the person they're talking to will say 'what's a grig?' and they'll explain. And so on, until maybe a few thousand people are using the word. And maybe one of those looks at another dictionary and finds that it isn't in there, so he calls his friend the lexicographer who happens to work for that dictionary publisher and tells him about the word.
'Are you sure about this?' 'Oh yes. I've looked it up in and its in there, so it must be a real word.' 'Oh, I'll get it put into the next edition then'.
You see? This doesn't work with languages, which aren't actually a real thing that can be studied independently of the works that describe them.
I've been a little concerned about applying the LGPL to Java code for a while. Let me explain why, and perhaps you can address my concerns.
Java linking does not work in the same way as the systems that the LGPL was originally intended for (i.e. systems with static or dynamic linking in a broadly similar sense to those provided by 'ld' or 'ld.so' under Unix).
If I produce a program that uses a Java library and compile it and wrap it up in a jar file, one would usually expect that this jar file could be treated in a similar manner to an executable program in Unix. However, a jar file provides more power to the end user than an executable program - that is it includes not just the executable code from but everything necessary to compile & link new code against any classes that are contained within it.
This means that any Java program could be treated as a 'library facility', and therefore appears to come under the scope of section 7, rather than section 6. And, as far as I have been able to deduce from the phrasing in section 7 (it isn't altogether clear), it seems that in order to distribute a work under section 7 I must permit further redistribution of my 'work based on the library'.
Am I simply misreading section 7, does this not apply for some reason I haven't considered, or am I correct in thinking that the LGPL just doesn't work for Java?
For reference, section 7 reads:
7. You may place library facilities that are a work based on the Library side-by-side in a single library together with other library facilities not covered by this License, and distribute such a combined library, provided that the separate distribution of the work based on the Library and of the other library facilities is otherwise permitted, and provided that you do these two things:
* a) Accompany the combined library with a copy of the same work based on the Library, uncombined with any other library facilities. This must be distributed under the terms of the Sections above.
* b) Give prominent notice with the combined library of the fact that part of it is a work based on the Library, and explaining where to find the accompanying uncombined form of the same work.
However, when you refer at compile time to methods, fields, and constants in another class, the compiler actually reads in that class and does the equivalent of '#include done right'. This makes Java code that directly refers at compile time to LGPL code a derivative work.
Are you sure about this? I know that a lot of library vendors would have you think this way (so they can say "your code dynamically links to our library which means you've copied our code..." even if the libary DLL/.so file isn't distributed with your app) but I'm sure I read a document not too long ago that gave reference to a precedent that states that merely copying interface information (eg names of functions, variables, classes, types of parameters, etc) does not cause a work to become derivative.
If somebody want's to dig it out, I think it was a document referenced from one of the stories on the SCO/Linux fiasco...
So, essentially, all of this is irrelevant. Nobody can stop you linking to a DLL that is already on your system. For these purposes, a java class file counts as a DLL (seeing as all linking is done at runtime in Java - the compiler only does typechecking, not linking).
Currently prosecutors must prove that at least 10 copies (or whatever) of the files you share have been downloaded from you in order for it to be a felony.
With this law they can just point at the files in the right location on your disk, show you've connected to the network, and say 'this person offered to make copies - we can assume that that offer was accepted enough times'.
I'm not sure if its overly concerning. Here in the UK we have a much more serious change being considered in the laws on rape that would mean that in order to prove rape it only has to be shown to be _likely_ that consent was withheld, not proved. That's a much more worrying state of affairs, IMHO.
Disclaimer: IANAL and may have misunderstood things, but I think I know what I'm talking about...
I don't think you own the copyright on your own DNA. In fact, I suspect that right resides with your mother (unless she has signed any contracts which give you specific rights) as it was actions performed by her body which gave rise to it (note that if you were conceived by IVF, the copyright may reside with the clinic unless you were a work for hire, I'm not aware of any precedents in this area). This means that any action in which your DNA is copied is an act prohibited by copyright, and which your mother has the right to control. This would not only include reproduction, but also growth of any form that involves the creation of new cells.
it's only a matter of time before the world leading superpower pressures other nation states to "harmonize" their laws with the US:
CNN Article from 2000 "There are rather insane laws in the U.S. about reverse engineering, and so we sidestepped those by having the work done in Europe under the European Union fair-use laws," said Jeremy Allison
The US can do many things with their economic muscle. But I suspect that the EU are beyond the reach of such intimidation, and the EU right to reverse engineer is an entrenched part of EU law now. I don't see it being changed for a long time to come.
Well, you have to understand, that this isn't really a situation that I myself am in.
OTOH, having been running a small business off the back of Linux servers for the past 6 years, it hasn't exactly all been roses. There's been the impossibility of finding cheap ISDN hardware that worked correctly with Linux. There's been the ADSL modem that kept crashing when run with Linux (although it wouldn't have worked at all with NT4). There have been data corruption issues and random system crashes. There have been failures in our IMAP servers that cause deleted messages to spontaneously reappear days later.
All of these are issues that I wouldn't usually expect to see with NT4.
As it happens, we could cope with them. And many of them are less likely to happen today and only happened because we were running cutting edge (the ISDN hardware we got would only work with an experimental 2.1 series kernel; the ADSL modem needed an unsupported kernel module and patches to pppd to support ppp over ATM). But an average small business probably couldn't cope with this kind of problem, and that is why Linux is not suitable for many people.
It is kind of sad to see Linux kernel series 1.99 go.
If:
1) The more recent Linux kernels weren't better in almost all respects, 2) Linux wasn't open source, and 3) Linux kernels came packaged with various servers and network clients many of which are regularly found to contain hideous security holes
I would agree with you. All 3 of these conditions hold for NT4.
Think of upgrading as feet of snow. When it's snowing outside, you can go out and shovel for hours and keep up with it, or you can wait until it's done, and there's 5 feet out there and try to shovel it. If you do it while it's snowing, you'll spend just as much time doing it, but you won't be as sore the next day (since you never lifted more than 20 lbs.). If you do it all at once after it's stopped snowing, your back is going to really hurt, and you'll still spend several hours out there.
That analogy is completely pants. It doesn't take into account the fact that you can economise on time taken upgrading by skipping version (eg upgrading from NT4 directly to W2K3 or XP).
That could lead to problems installing NT4 on a Pentium 6. Yeah, it'll run fast... if it runs.
I see no reason that it shouldn't. You can still run DOS 3 on a Pentium 4 should you feel like it. Or Windows 3. These all still work. Why should NT4 stop working all of a sudden? Only a very small minority of badly written software stops working due to mostly-backwards-compatible processor upgrades - the only example I can actually think of is that anything compiled with Borland's Pascal compilers in the early nineties stopped working when processor speeds starting exceeding about 200MHz - this was due to delay loop calibration code that overflowed a 16 bit register when run on machines this fast. This kind of problem is thankfully rare in modern software: pretty much everyone learnt their lessons fairly early on in this game.
Besides, you could just underclock your processor to solve this kind of problem.
Killing off NT4 and its old Microsoft LAN Manager "networking" was like killing off the 9x line. It had to be done and it'll hurt now and months later you'll be wondering what exactly the fuss was again
Err... maybe the approx. £1000 that forking out for a Win2K server license will cost. Take it from me, but for a small business, even months later you can be feeling the pain of an unnecessary cost like that.
Also, what exactly is wrong with NT4's networking that is fixed by more recent Windows systems? I mean, OK, XP has a hacked Kerberos system which is kind of useful when working with multiple servers (I don't). What exactly are the other improvements that have been made over what NT4 supports?
Which part of Transparent Web Caching is not covered by or is different to the following claim of the patent?
A method for transferring information via the Internet, comprising the steps of:
intercepting a message from an Internet user directed to a content provider address;
determining whether or not the message is an information request;
sending the message to the Internet without being affected if the message is not an information request;
determining whether or not said information request relates to a content provider address having a corresponding alternative address, said alternative address providing at least part of the information provided at said content provider address; and
directing said information request to said corresponding alternative address, if existing, or sending said information request to the Internet without being affected, if not.
Can somebody explain to me WTF the difference between audio and data CDRs is? In my experience, you just put the disc in the drive, tell your CD recording software which type of disc you want to produce, give it a list of the files to put on there, and away you go. One disc works for either format. Or is there something I'm missing?
Don't mean to nitpick, but Qinetiq isn't really a "firm". They're a government department. Used to be called DERA (Defense Evaluation Research Agency if my memory is correct), and are the UK equivalent of the US DARPA.
If they aren't statically or dynamically linking to the code. Then exactly how are they linking to it? Magically?
Maybe they're not linking to it, but just using services provided by it, eg the system call interface, which doesn't use linking at all, but instead uses a platform-dependendant method that generally requires no linkage to occur. Such methods include raising a software interrupt or using a special system call instruction, depending on the processor in use.
There is a theory, albeit with little if any actual supporting evidence, that suggests that a mutation occurred in the human genome, probably initially arising somewhere in central or west asia, which led to the ability to develop and understand strict codes of behaviour. Without this capability, large scale civilisation would have been impossible, and humanity would have been restricted to living in small tribal groups. The mutation would have slowly spread to the rest of the world by population migration (etc), leaving in its wake a tide of primitive civilisations. History seems to bear out this version of events.
Of course, there is at present no way of proving this theory, and probably little chance of disproving it either. So as a theory, its not particularly remarkable. But it does fit with what little evidence there is (at least according to the source I picked it up from, and what was said there agrees generally with what I know of the development of early civilisations).
I would. Thunking is fun, and a wonderful way of ensuring you keep compatible. Try thunking to another operating system running in a virtual machine, that's cool:-)
Except of course that part of their settlement for abusing their monopoly position was a court order that they open all the information on their protocols to the general public.
Slashdot Mirror
If I invest and the government uses my "vote" that something will happen to prevent it from happening, I lose my money, because the event didn't occur.
Do you really think that's likely? I mean, seriously, despite how funny that comment sounds - a large number of people will be using this system. A lot of them are going to have all kinds of weird contracts open at any one time. One or two people moving in a certain direction is not going to set any alarm bells ringing. Its if a whole load of people suddenly move a lot of money in a particular direction that they'll get worried.
If the event actually happens and I have predicted it, I get my phones tapped, a visit from the FBI, probable interrogation and possible imprisionment for having "terrorist ties." What exactly is the incentive for an investor to enter this market?
And a few weeks later the government having run their extensive background checks on you decide that you aren't a terrorist, couldn't possibly have had anything to do with the incident, and that they have no evidence whatsoever to hold you with and release you. You gain your $200,000 or whatever you made. It's also very unlikely that you'll be investigated more than briefly, because a whole load of other people will have made the same contract you did, and most of those'll be innocent too.
You see, this _isn't_ actually about catching terrorists, however appealing that might sound. They may get one or two particularly stupid ones, but I doubt it'll add up to much in that direction. The real value is that it will help iron out the economic costs of terrorism.
Take two situations.
1. You run a travel company which specialises in holidays in Israel. You make a fair sum of money off of this, but you realise that any terrorism in the area puts off large numbers of your clients. Through past experience, you know that a major terrorist attack costs you $15,000 in lost business. However, due to the ratios in the market, you can place a contract which only costs you $10 per month to keep open which will compensate you for this cost in the event that this does happen. It'll give you peace of mind, so you do it - its definitely worthwhile.
Of course, because this is a derivatives market, somebody has to be offering to make those contracts with you, which brings us on to situation 2...
2. You're a defense contractor. Global unrest is generally good for you because when middle eastern countries are more nervous about terrorist threats they tend to buy more up-to-date equipment for their armies. But you don't like the fact that you feel like a celebration every time somebody blows up a car outside an embassy. So you've been looking for years for a way of balancing the good times and the bad, and here it is. There are hundreds of companies who are willing to pay a small price every so often in exchange for a bigger payout whenever there is a terrorist attack. You look at your books, work out what you can offer whenever anything like that happens, and feed your offer into the market... it comes back telling you that it has sold a number of contracts on the terms you suggested. Now you have a steady stream of income rather than one that jumps around all the time.
Good, isn't it?
Of course, it might not work in practice (I'm thinking there might not be enough people in category 2 to fulfill all the contracts that category 1 people would want), but the theory is excellent.
There's little advantage in passing every byte through two seperate journal
1. If you're using a filesystem for databases, then the chances are you're keeping other stuff besides the database on them. Otherwise, you'd use a raw partition.
2. I believe that by default data changes to files aren't journalled in either of the systems that were being compared; only filesystem metadata changes. This doesn't cost much in terms of resources, particularly if the metadata rarely changes (as I suspect it doesn't on oracle data files, which I assume are fixed size allocations that are opened when the server is started and never closed...).
Sadly, the poetry evolved to the point where it attracted a predator's attention, and now it's gone extinct.
This is why I'm against broadcasting our presence to the stars.
NEWSFLASH
INTERGALACTIC NEWS - MILKY WAY, ALPHA QUADRANT
Sentient species from all around the galaxy were stunned today by the news of the total annihilation of all life on Earth.
Earth, a small blue green planet orbiting an uninteresting single star towards the outer edge of the civilised portion of the galaxy, first came to the attention of scientists only a handful of years ago. "They started by broadcasting simple messages between themselves," said Humanologist Urk-Thr'qar of Threep University. "They didn't seem aware of our presence. But then they started sending messages that seemed directed at us. We had to investigate further."
But it was one of the messages transmitted internally within the planet that caused their eventual downfall. A human, it seems, had an interesting and imaginative new idea. Unfortunately, the actual idea has been lost forever due to what happened next.
A message was posted on the intergalactic network news site, slashdot.galaxy. The message read simply: "Hey, check out Earth. They're not as dumb as we thought," and included a set of hyperspatial co-ordinates for Earth.
Within seconds, the planet was surrounded by a cloud of interplanetary space ships. Only hours later, the ships had all departed, taking with them every scrap of life that existed on the planet.
Late arrivals were disappointed to find the planet barren.
"Huh!" said one three-eyed creature from Ph'tarra. "I come all the way over here and the entire planet's slashdotted. You'd have thought they would have had the courtesy to lay on extra bioreplicators to at least attempt to handle the extra specimen collectors they were bound to receive."
One lucky specimen collector remarked "Well, I mean I got myself a specimen, but it was only a small quadroped. I don't think its even house trained. But I'll put up a copy on GeneTorrent if you like."
I think nasm (which I was working on a lot at the time) made it as far as redhat in a 0.3something release. Of course we jumped straight over all the intervening versions to 0.9, cos it looked like we were nearly ready for a 'proper' release.
It might happen one of these decades, too...
Seriously, though, the only real way of getting your software into distributions is to get it used. With nasm, myself and Simon (the primary author) spent many many hours hanging around relevant newsgroups, letting people know that it was available, making sure they knew its advantages over other similar packages, etc. After a few months of that we had maybe a few hundred regular users. That was when we started hearing from redhat & debian (primarily that they didn't like our license and could we change it a little bit... if you want my advice, make sure you're either gpl or lgpl, it saves a lot of hassle in the long run...).
Is there any case law that states that dynamic linking creates a derivative work?
A few weeks back, there was a post about some documents from an old Unix Systems Labs copyright infringement case that people might have considered interesting in relation to the SCO case.
In one of these documents (from the BSD camp) there was a reference to a precedent which (according to that document) meant that merely copying interface information would not cause a work to be a derivative. I don't know how much weight that holds though, or even if the decision was later overturned....
One thing that interests me is your assertion that OLE/DCOM is slower than XML based message passing.
.NET Remoting (something I've never looked at in detail, I assume its somewhat like Java RMI?) is slower than rolling your own XML based interface, and then compares it with OLE (the implication that I infer being that rolling your own system with similar features to OLE is likely to produce something faster than OLE itself).
I don't think he actually said that. Looking through the message again, he states that
Dictionaries contain false entries intended to serve as markers and preserve the collection copyright
... dictionaries? I don't think this would work. You see, there are quite a lot of people who will, from time to time, just pick up a dictionary and browse. Try to find interesting words that they've never heard of before. So, if a dictionary had (for example) the entry 'Grig (n) the piece of a ringpull on a drink can that gets pushed inside the can upon opening' someone will probably notice this. Then they'll be talking to somebody at some point and an opportunity to use this fabulous word 'grig' will arise. So they will. And the person they're talking to will say 'what's a grig?' and they'll explain. And so on, until maybe a few thousand people are using the word. And maybe one of those looks at another dictionary and finds that it isn't in there, so he calls his friend the lexicographer who happens to work for that dictionary publisher and tells him about the word.
I can understand maps, but
'Are you sure about this?'
'Oh yes. I've looked it up in and its in there, so it must be a real word.'
'Oh, I'll get it put into the next edition then'.
You see? This doesn't work with languages, which aren't actually a real thing that can be studied independently of the works that describe them.
I've been a little concerned about applying the LGPL to Java code for a while. Let me explain why, and perhaps you can address my concerns.
Java linking does not work in the same way as the systems that the LGPL was originally intended for (i.e. systems with static or dynamic linking in a broadly similar sense to those provided by 'ld' or 'ld.so' under Unix).
If I produce a program that uses a Java library and compile it and wrap it up in a jar file, one would usually expect that this jar file could be treated in a similar manner to an executable program in Unix. However, a jar file provides more power to the end user than an executable program - that is it includes not just the executable code from but everything necessary to compile & link new code against any classes that are contained within it.
This means that any Java program could be treated as a 'library facility', and therefore appears to come under the scope of section 7, rather than section 6. And, as far as I have been able to deduce from the phrasing in section 7 (it isn't altogether clear), it seems that in order to distribute a work under section 7 I must permit further redistribution of my 'work based on the library'.
Am I simply misreading section 7, does this not apply for some reason I haven't considered, or am I correct in thinking that the LGPL just doesn't work for Java?
For reference, section 7 reads:
7. You may place library facilities that are a work based on the Library side-by-side in a single library together with other library facilities not covered by this License, and distribute such a combined library, provided that the separate distribution of the work based on the Library and of the other library facilities is otherwise permitted, and provided that you do these two things:
* a) Accompany the combined library with a copy of the same work based on the Library, uncombined with any other library facilities. This must be distributed under the terms of the Sections above.
* b) Give prominent notice with the combined library of the fact that part of it is a work based on the Library, and explaining where to find the accompanying uncombined form of the same work.
However, when you refer at compile time to methods, fields, and constants in another class, the compiler actually reads in that class and does the equivalent of '#include done right'. This makes Java code that directly refers at compile time to LGPL code a derivative work.
Are you sure about this? I know that a lot of library vendors would have you think this way (so they can say "your code dynamically links to our library which means you've copied our code..." even if the libary DLL/.so file isn't distributed with your app) but I'm sure I read a document not too long ago that gave reference to a precedent that states that merely copying interface information (eg names of functions, variables, classes, types of parameters, etc) does not cause a work to become derivative.
If somebody want's to dig it out, I think it was a document referenced from one of the stories on the SCO/Linux fiasco...
So, essentially, all of this is irrelevant. Nobody can stop you linking to a DLL that is already on your system. For these purposes, a java class file counts as a DLL (seeing as all linking is done at runtime in Java - the compiler only does typechecking, not linking).
Currently prosecutors must prove that at least 10 copies (or whatever) of the files you share have been downloaded from you in order for it to be a felony.
With this law they can just point at the files in the right location on your disk, show you've connected to the network, and say 'this person offered to make copies - we can assume that that offer was accepted enough times'.
I'm not sure if its overly concerning. Here in the UK we have a much more serious change being considered in the laws on rape that would mean that in order to prove rape it only has to be shown to be _likely_ that consent was withheld, not proved. That's a much more worrying state of affairs, IMHO.
Disclaimer: IANAL and may have misunderstood things, but I think I know what I'm talking about...
Could somebody tell me if those billion are 10^9 or 10^12? I suspect the latter, although an American source kind-of implies the former...?
And isn't it about time something was done to stop this confusion?
Messages! Luxury!
In my day, we were lucky t'have couple phonemes to string together t'make what sound' like might be one them new fangled words.
And if you try telling that to kids nowadays, will they believe you?
I don't think you own the copyright on your own DNA. In fact, I suspect that right resides with your mother (unless she has signed any contracts which give you specific rights) as it was actions performed by her body which gave rise to it (note that if you were conceived by IVF, the copyright may reside with the clinic unless you were a work for hire, I'm not aware of any precedents in this area). This means that any action in which your DNA is copied is an act prohibited by copyright, and which your mother has the right to control. This would not only include reproduction, but also growth of any form that involves the creation of new cells.
Stand by to be sued...
it's only a matter of time before the world leading superpower pressures other nation states to "harmonize" their laws with the US:
CNN Article from 2000 "There are rather insane laws in the U.S. about reverse engineering, and so we sidestepped those by having the work done in Europe under the European Union fair-use laws," said Jeremy Allison
The US can do many things with their economic muscle. But I suspect that the EU are beyond the reach of such intimidation, and the EU right to reverse engineer is an entrenched part of EU law now. I don't see it being changed for a long time to come.
Well, you have to understand, that this isn't really a situation that I myself am in.
OTOH, having been running a small business off the back of Linux servers for the past 6 years, it hasn't exactly all been roses. There's been the impossibility of finding cheap ISDN hardware that worked correctly with Linux. There's been the ADSL modem that kept crashing when run with Linux (although it wouldn't have worked at all with NT4). There have been data corruption issues and random system crashes. There have been failures in our IMAP servers that cause deleted messages to spontaneously reappear days later.
All of these are issues that I wouldn't usually expect to see with NT4.
As it happens, we could cope with them. And many of them are less likely to happen today and only happened because we were running cutting edge (the ISDN hardware we got would only work with an experimental 2.1 series kernel; the ADSL modem needed an unsupported kernel module and patches to pppd to support ppp over ATM). But an average small business probably couldn't cope with this kind of problem, and that is why Linux is not suitable for many people.
It is kind of sad to see Linux kernel series 1.99 go.
If:
1) The more recent Linux kernels weren't better in almost all respects,
2) Linux wasn't open source, and
3) Linux kernels came packaged with various servers and network clients many of which are regularly found to contain hideous security holes
I would agree with you. All 3 of these conditions hold for NT4.
Think of upgrading as feet of snow. When it's snowing outside, you can go out and shovel for hours and keep up with it, or you can wait until it's done, and there's 5 feet out there and try to shovel it. If you do it while it's snowing, you'll spend just as much time doing it, but you won't be as sore the next day (since you never lifted more than 20 lbs.). If you do it all at once after it's stopped snowing, your back is going to really hurt, and you'll still spend several hours out there.
That analogy is completely pants. It doesn't take into account the fact that you can economise on time taken upgrading by skipping version (eg upgrading from NT4 directly to W2K3 or XP).
That could lead to problems installing NT4 on a Pentium 6. Yeah, it'll run fast... if it runs.
I see no reason that it shouldn't. You can still run DOS 3 on a Pentium 4 should you feel like it. Or Windows 3. These all still work. Why should NT4 stop working all of a sudden? Only a very small minority of badly written software stops working due to mostly-backwards-compatible processor upgrades - the only example I can actually think of is that anything compiled with Borland's Pascal compilers in the early nineties stopped working when processor speeds starting exceeding about 200MHz - this was due to delay loop calibration code that overflowed a 16 bit register when run on machines this fast. This kind of problem is thankfully rare in modern software: pretty much everyone learnt their lessons fairly early on in this game.
Besides, you could just underclock your processor to solve this kind of problem.
Killing off NT4 and its old Microsoft LAN Manager "networking" was like killing off the 9x line. It had to be done and it'll hurt now and months later you'll be wondering what exactly the fuss was again
Err... maybe the approx. £1000 that forking out for a Win2K server license will cost. Take it from me, but for a small business, even months later you can be feeling the pain of an unnecessary cost like that.
Also, what exactly is wrong with NT4's networking that is fixed by more recent Windows systems? I mean, OK, XP has a hacked Kerberos system which is kind of useful when working with multiple servers (I don't). What exactly are the other improvements that have been made over what NT4 supports?
Have you read the patent?
Which part of Transparent Web Caching is not covered by or is different to the following claim of the patent?
A method for transferring information via the Internet, comprising the steps of:
intercepting a message from an Internet user directed to a content provider address;
determining whether or not the message is an information request;
sending the message to the Internet without being affected if the message is not an information request;
determining whether or not said information request relates to a content provider address having a corresponding alternative address, said alternative address providing at least part of the information provided at said content provider address; and
directing said information request to said corresponding alternative address, if existing, or sending said information request to the Internet without being affected, if not.
Can somebody explain to me WTF the difference between audio and data CDRs is? In my experience, you just put the disc in the drive, tell your CD recording software which type of disc you want to produce, give it a list of the files to put on there, and away you go. One disc works for either format. Or is there something I'm missing?
Don't mean to nitpick, but Qinetiq isn't really a "firm". They're a government department. Used to be called DERA (Defense Evaluation Research Agency if my memory is correct), and are the UK equivalent of the US DARPA.
If they aren't statically or dynamically linking to the code. Then exactly how are they linking to it? Magically?
Maybe they're not linking to it, but just using services provided by it, eg the system call interface, which doesn't use linking at all, but instead uses a platform-dependendant method that generally requires no linkage to occur. Such methods include raising a software interrupt or using a special system call instruction, depending on the processor in use.
There is a theory, albeit with little if any actual supporting evidence, that suggests that a mutation occurred in the human genome, probably initially arising somewhere in central or west asia, which led to the ability to develop and understand strict codes of behaviour. Without this capability, large scale civilisation would have been impossible, and humanity would have been restricted to living in small tribal groups. The mutation would have slowly spread to the rest of the world by population migration (etc), leaving in its wake a tide of primitive civilisations. History seems to bear out this version of events.
Of course, there is at present no way of proving this theory, and probably little chance of disproving it either. So as a theory, its not particularly remarkable. But it does fit with what little evidence there is (at least according to the source I picked it up from, and what was said there agrees generally with what I know of the development of early civilisations).
Of course, IANAH (I am not a historian).
Whoda thunk?
:-)
I would. Thunking is fun, and a wonderful way of ensuring you keep compatible. Try thunking to another operating system running in a virtual machine, that's cool
Except of course that part of their settlement for abusing their monopoly position was a court order that they open all the information on their protocols to the general public.