Java programs aren't cross-platform. Otherwise I could download a single package to run on whatever OS I have.
Java is a platform, or at least Sun calls it that. Following that reasoning leads to absurdities like claiming Super Mario Brothers or StarCraft are cross-platform, since they work in emulators. You could say I'm being pedantic, that "running on a platform on platform A" is practically equivalent to "running on platform A", but that's not so.
To the point, suggesting the use of Swing imposes serious limitations on the project- mainly that it has to be written in Java and run on a JVM. The performance penalties are appreciable. (It is possible to use Swing to display a non-Java program, but ugly and impractical).
The other GUI toolkits that got good recommendations generally have multiplelanguage bindings for them
PersonalJava is sold on devices with 64 megabytes of RAM, or even more.
It's true that nukes need an electrical trigger to explode correctly (otherwise it's just a "dirty bomb"). But the electronics are fairly simple, and must've been designed to function in EMP environments.
Remember that the surest way to send an EMP pulse is to detonate a nuke! Therefore any warheads meant for nuculear combat had better be EMP-hardened, or else your "world-shattering war" will fizzle out after one or two blasts.
People have already been killed by directed EMP weapons!
Raytheon was experimenting with those devices- apparently it wasn't as "directed" as they wanted, and their lab safety procedures were inadequate. A research employee was eventually killed.
It wasn't a "zap zap, you're dead" situation- the exposure was accumulated over a long time, and even after the cellular damage was medically detected he lingered for many months.
(I wish I could find a link for this. Read it in newspaper back when it first occured.)
Unix GUI code has been written since 1984 without needing threads.
Unless the logic of my algorithms is more simply represented as threads, I shouldn't use them. If I want to toss a pretty GUI ontop of a large and ancient existing codebase that is resolutely non-threadsafe, I shouldn't have to use threads. If my development tools (linux and gdb) don't support threads easily, it'll be aggravating to use them.
Suppose I had a thread? What would it do? Wait for stuff to come from the network, and stuff it in a buffer for the main thread to read once it's ready. Well, the operating system is already doing that exact thing for me. I shouldn't need to redundantly duplicate that same behavior in my own code.
Fltk's add_timeout doesn't obey it's own documentation. It's wrong. No one to blame but that library.
Don't you think you should background your socket code into another thread?
I don't think so. Threading could fix this particular problem, but then I'd have to use threading. That would add a lot of complexity, and overhead as slower, re-entrant versions of all library functions are pulled in. The code would also be less portable, since pthread functions aren't really compatible across different Unices.
The style I'm using now is the (originally) recommended way to write Unix network code, as per Stevens. That is, a select() call to check for network input (if any exists). It works fine with programs that use tty, ncurses, Qt, Motif, or Fltk without calling certain blocking function. Fltk plainly continues to exectute during the blocking function, as user-input continues to work. But it's neglecting its promise to keep the timeouts working.
I've tried out Fltk. It worked alright for simple tasks, but I couldn't figure out how to make more complex programs.
For instance, I wanted my program to use TCP/IP (regular bsd calls: socket, accept, recv). I used Fl::add_timeout and Fl::repeat_timeout to check the incoming buffer every few seconds (no need for high data rate).
That worked great- unless I made some call like fl_file_chooser, which blocks until it returns. Then the timeout function won't happen until the dialog box is done- a fact completely ommitted from the documention.
I suppose I could dig into the source code (it is LGPL after all) and try to fix it, but things like this don't inspire confidence.
If you have to lug around a huge backpack of support gear
Give him a break! He's a lone student, trying to produce a useful prototype of the HCI loop. The proposal isn't for a consumer level product.
If the demonstration is successful, then a PDA manufacturer could look into engineering the hardware down into a single handheld device, but first they've got to see the concept in action.
Imagine a bus rounding a corner and the text compensating by scrolling.
That could be a beneficial effect! Many people have difficulty reading small text on a moving vehicle, because the page constantly bounces around.
Possibly, this system could act as an "image stabilizer" for the text- causing the text to follow a smoother path than your actual bouncing hand.
Of course, whether or not this can be helpful depends on many factors- Does the screen have 10 millisecond updates? Does your head bounce more or less than your hand? (If they're in sync already, then you're fine.)
And how well does the inertial tracking system distinguish the gross movements of the bus from your localized jittering? (You wouldn't want to leave the POV behind you at the station where you boarded!)
You raise some interesting points, but five years could be a very short time indeed.
Five could be too little, but 80% of the time, if a work is going to earn money, it'll have done it by 5 years. (I've heard films do 40% of their earnings in 18 months of worldwide screenings, then 50% more in video rental, cable, and broadcast TV across the next 4 years). 15-25 years should be more than enough for 98% of the potential profit to come in. (And its just enough time for kids to grow into adults)
(I think the main motive for media companies to desire 50+ year copyright terms isn't to collect more profits on old works, but to reduce the free competition to their new works)
What I can't do is sell, publish, or distribute these stories.
With the pervasiveness of the internet (it's growth has only started), a major part of our communications with our friends and family takes place over it. Interpersonal communication can now be interpreted as "publication". You send someone an email/recorded story- he'll share with his friends, etc etc. Word gets around. And then lawyers drop by.
It hasn't happened much yet, but it may. We've already seen 11 year-olds C&D for their Pokemon websites...
I would be cool if after the last movie goes onto DVD if Lucas would open source Star Wars, but I think it should be his choice.
In a way, Star Wars (and Star Trek, and Buffy Vampire Slayer, and other huge franchises) are already a little like Open Source, in that dozens of authors have contributed stories that became part of the "canon". The copyright holder still holds the position of gatekeeper, but we've seen that they'll allow quite low quality work to get the stamp of approval, as long as it promises a little money. (Author Whedon has admitted that he can't bear to read the Buffy books he's allowed to be published)
Thing is, a close look a the term "Operating System" shows that it is maddeningly poorly defined.
I mean, there's almost no semantic content provided. "A system that is operating". Could be almost anything, as long as it runs.
Computer Science hasn't been around long enough for these "traditional meanings" to acquire much historical force.
There are two definitions we could use for "Computer Operating System": "the thing that runs ontop of the electronic hardware" or "the thing that runs beneath the software programs"
In the past, computers were weaker and could hardly contain a single operating system and one or two programs, so those two definitions were essentially equivalent. Today, we have more freedom to run convoluted arrangements of software, so those definitions have diverged. The question then arises: "Which definition is more useful?"
I prefer the latter ("under the software"), although the bulk of "OS research" in the past 30 years has favored the former. The software-view is from the perspective of the user of an OS, not it's creator, and thus has more utility for more people.
Do they provide process scheduling ?
The marketplace has seen many successful products labelled as Operating Systems which provided no process scheduling. Yet I can think of a few modern application programs which do provide it.
Does Emacs or Mozilla control low-level I/O or memory management ?
Both emacs and mozilla can run programs on top of them (Written in either elisp or javascript). They provide memory management to those programs. And they control their low-level I/O. You might be biased and think that "low-level" necessarily means "on the metal", but that doesn't have to be the case. It might just be "relatively low", from the point-of-view of the application program.
Look at it this way: Is MS-DOS an "Operating System"? Most people agree it is- why, it's even there in the name. Take a program like "Leisure Suit Larry"- it was written for MS-DOS, and needs an Operating System to work.
But I can run DOSemu on top of Linux, and then run Larry ontop of that. You can't claim Larry is running without an OS, so DOSemu must be viewed as providing it's OS (either by itself, or in combination with Linux, however you prefer). From it's perspective, DOSemu is the operating system. Everything below that (Linux) is the "hardware".
Then suppose I also run Emacs on Linux, and then run Tetris on Emacs. The Tetris/Emacs/Linux block diagram is exactly the same as for Larray/DOSemu/Linux, so the argument for calling Emacs (or any sufficiently expandable software) an Operating System is quite strong.
(And besides! You don't have real "hardware" at all! Those wires and chips are all emulated, running on top of The Matrix)
I don't think we need current stuff, but it is unhealthy for a culture to have so many of central, identifying stories be the property of a single, centralized owner. It is, frankly, undemocratic.
I agree that current works need copyright protection for at least a few years, but I think that terms should expire soon enough for adults to create derivative works from stories they've heard as children.
Since ancient times, children would listen to stories told by their elders around a campfire, and hold those imaginary heros in their hearts and minds as they made their way in life. Eventually they grow up and become the elders, and tell new stories of the cultural role-models and villians- based on the earlier tales, but colored by their own life experiences.
This is part of how a culture can grow and mature over the ages. Every listener will be a teller.
But, in the modern world, we sent our children to sit around a film screen and watch Star Wars. The wiles and wisdom of Han Solo and Ben Kenobi formed their subconcious guides. When they grow up and are parents themselves, can they give their children a modified version of their own favorite tales?
No. Only Lucas can do that, and the impotent parents cannot contribute even a little to the dreams of their children (no more than choosing between Star Wars and Harry Potter).
If a helpless child is infused with corporate propatainment in his formative years, we at least owe him the right to someday share the stories they inspired in him, instead of stepping back onto the media industry's treadmill.
"You're buying a movie ticket, and selling your dreams"
But the more translation layers there are between the source code and the hardware, the slower the code will run.
That's not really true, although it happens that way in many cases.
Looking at the example of Java vs C++ specifically, the slowness is actually (sometimes) due to the heavier semantics of the Java language. For instance, take a complier (like gcc/gcj) which can turn both C++ or Java code into a finished native executable.
Program some time-consuming math alg (Fibonacci, Newtonian integration?) in both, and measure the runtimes. Java will be slower by almost a constant factor.
Why?
Maybe skill of the compiler-authors (The GCC team just does't know how to do java well yet?)
If we accept this point, that some compilers are simply better than others, than we can imagine a situation where compiling C++ into Ada is beneficial: You recieve a powerful NSA-surplus machine which only runs Ada, and you'd like to run Apache and Mozilla on it. Rather than re-writing those in Ada, it may be faster to use a "cxx2ada" preprocessor on the application source code.
(That's not likely to happen, but with the new "hardware JVM" chips coming out, you may find a computer which can only run Java programs, and compiling C++ into Java might be a viable intermediate build-step)
Or maybe because the semantic's of Java's arithmetic operators require it to check for math exceptions (overflow, underflow, zero-divide) after every operation?
You'd have to dig deeply into both compilers to find why.
OTOH, the reverse assertion, that source code which is further separated from the hardware will perform faster, has sometimes been demonstrated to come true. (Although most modern compilers are still hobbled by legacy application-building techniques, and can rarely achieve such awesome successes)
That document also states that Congress can make no law which takes effect before it was passed (ex post facto). Plainly, extending copyrights on an already published work does this (just as surely as extending jailtime for an already-convicted felon would)
Further, the Constitution forbids the government from seizing private property without fairly compensating the owners. I had been in possesion of some "futures" on copyrights for a variety of works published in the 1920s, such that I would today be able to publish my own editions.
Congress has taken away that intellectual property from me- not only can I no longer publish my own versions, but I must pay money if I wish to read "Detective Comics #1". They've deprived me of something of measurable dollar value, yet I have recieved no compensation.
Where can I send my bill for 1/284,000,000th of the total worth of all publications from 1921-1941?
Life expectancy went up dramatically in the previous century
The original copyright term was 28 years- below normal lifespan, even in 1500. Copyright today is 90 years, above expected lifespan for children born in 2003.
If you want a law to be based on how long someone will live, then how about basing it on the lifespan of the people in question? No need to keep passing laws whenever there's an incremental medical advance. (However, copyrights should not be based on the lifespans of the author, or any of his descendants. That unfairly discriminates against authors who are old, sick, or childless, independent of their writing skills)
so it's only natural to extend copyright term in accordance
What else might be natural? Digital computer technology has greatly accelerated the speed that works can be authored, published, marketed, and sold- you need less time to make a fair profit today than you did in the past. So less copyright term seems natural.
Also, the US population has increased more than 10-fold in two centuries. Copyright was once a bargain between 1 author and 20 million readers, who agreed not to reproduce his work for 28 years each, or 560 million years collectively. To account for the same net deprivation with today's population, the copyright term should be only 1.9 years. That would balance the author's profit-opportunity across history. Seems natural to me!
Eldred will have access to the material he wants in a few more years
Scientists have determined that "few" means fewer than 7, even in the most extreme circumstances. 20 years at least it will take, and given the pattern established in 1976 and 1998, the actual time will be much much more.
go to Congress and get the extension laws repealed
Oddly, they can't. Or at least that's what the legislators will claim if you bug them about it.
You see, it's unconstitutional to change the length of existing copyright (either to lengthen, or shorten, as Stevens noted in his dissent, page 5). To do so is both ex post facto law and seizing private property without compensation.
The Court let them get away with it this time (and in 1976), because that was the pro-corporate thing to do.
So even if you got a whole new body of congressmen to repeal the law, the MPAA could appeal it to the supreme court and get it tossed out by arguing the same case Lessig just presented! And of course, if it's a corporate attorney before them, we can be sure they'll all nod along like good little puppies.
which are real property and cannot fall into the public domain.
Real property (real estate, or physical goods) can and has been seized by the government for the good of the nation. Claiming eminient domain is a traditional congressional power.
However, they always have to compensate the property-holder the fair market value of the goods taken. Lets see, I was due to recieve unlimited reproduction & derivation rights for Detective Comics #1 by now... I wonder if Congress has mailed my check yet?
"Allow"? I was under the impression that they snuck in.
They sneak in initially, but once you get past the INS-fortified towns along the border, average American citizens don't treat their presence as a crime at all. (More like a *wink-wink* opportunity to avoid the minimum wage and income tax laws when getting some lawncare done)
I had expected the events of Sept. 11, 2001 to make the US public less accepting of illegal immigrants in their midst, but they still don't really care (if the visitors don't look arabic)
If your snowplow driver lets slip that his visa's expired, do you sneak to a phone and dial 911? No, most Americans would find that rude and petty. And the local police won't be really enthusiastic about investigating either. (Unless they had an INS task force visiting their department at the time. In big cities this can be a full-time situation)
Even when it is enforced, illegal immigration is treated differently from other things which are "illegal". Most illegalities lead to a guaranteed jury trial, then a fine or incarceration. Catch someone red-handed at infiltrating the US borders, and he's dumped on the streets of Mexico that same night.
I'm not very in-touch with American opinions, but occasionally I glean something from mass media. I was really astounded last year upon reading a sob-story on the front page of a major newspaper, detailing the plight of illegal-immigrant teens. Apparently these poor people (brought to America at age 8-12, say) were leading successful high school careers, but then were unable to collect college scholarships!
Fight back, Palm; Apple beat Xerox, so you can too.
s/beat/paid/g
Microsoft, on the other hand, beat Apple on the same score (demonstrating to Apple that they hadn't really needed to pay Xerox in the beginning). Amusingly, the FSF and related groups was supportive of Microsoft in those instances
The interesting problem with the Fitaly layout is that it is also patented! (number 5487616)
Fitaly corporation seems willing to license it for low rates, so Palm could afford it if they wanted to go that way.
However, if Fitaly (or Xerox Unistroke, or any other patented input method) becomes popular/standardized, then Free Software programs will be unable to clone the UIs of the next generation of Palmtop software. (In the past it was shown illegal to copyright a UI, but patents still seem to apply)
I can already find one Open Source project that's using the Fitaly layout, in violation of the patent. I wonder if fitaly.com will send them a C&D as soon as they notice, or if they'll wait for the input method to become popular on Linux PDAs before yanking it away.
Or make sure you know what kind of better hardware you could buy, if needed.
When developing a system you should try to overload it so you can recognize what a failure state looks like. This may give your engineers valuable insight.
What is the resource that gets exhausted first? What is the system's behavior when it is completely overloaded? Does it just stop functioning, or does it lose data? Or maybe generate bad data?
These things could be nice to know, and may suggest quick improvements so that, if 6 years later the customer puts in 20 times as much usage as was originally budgeted, the failure isn't completely embarrassing.
Obviously, in the case of "Joe Blow" hooking up a Windows 2000 box to the Internet on his home DSL connection, enabling various server services, and not keeping up with the updates, there is no such accountability. If his machine is compromised, and then used to attack other machines as part of a DDoS attack,
If his machine can be compromised for DDoS, it can also be infiltrated to record his credit card numbers and publicize them. (Or make random purchases in his name, with the goods not necessarily coming to either him or the hacker).
If hackers were basically unpunished by the law, then people who install zombie-able machines will soon face the impromptue penalties of getting robbed blind.
It is the hackers who create the need for security,
If there were zero hackers alive today, and if the concept of "computer intrusion" were just a theoretical problem with no examples from precedent, there'd still be a need for security.
People won't act like it, but they do need it.
2 years ago the concensus was that knives were fine for airline passengers. That was wrong, they needed to be banned, but the public in general didn't care.
Slashdot Mirror
Java programs aren't cross-platform. Otherwise I could download a single package to run on whatever OS I have.
Java is a platform, or at least Sun calls it that. Following that reasoning leads to absurdities like claiming Super Mario Brothers or StarCraft are cross-platform, since they work in emulators. You could say I'm being pedantic, that "running on a platform on platform A" is practically equivalent to "running on platform A", but that's not so.
To the point, suggesting the use of Swing imposes serious limitations on the project- mainly that it has to be written in Java and run on a JVM. The performance penalties are appreciable. (It is possible to use Swing to display a non-Java program, but ugly and impractical).
The other GUI toolkits that got good recommendations generally have multiple language bindings for them
PersonalJava is sold on devices with 64 megabytes of RAM, or even more.
Read: nuclear bombs.
It's true that nukes need an electrical trigger to explode correctly (otherwise it's just a "dirty bomb"). But the electronics are fairly simple, and must've been designed to function in EMP environments.
Remember that the surest way to send an EMP pulse is to detonate a nuke! Therefore any warheads meant for nuculear combat had better be EMP-hardened, or else your "world-shattering war" will fizzle out after one or two blasts.
(Or near an airport)
People have already been killed by directed EMP weapons!
Raytheon was experimenting with those devices- apparently it wasn't as "directed" as they wanted, and their lab safety procedures were inadequate. A research employee was eventually killed.
It wasn't a "zap zap, you're dead" situation- the exposure was accumulated over a long time, and even after the cellular damage was medically detected he lingered for many months.
(I wish I could find a link for this. Read it in newspaper back when it first occured.)
Alas, the networking code is very deep in a library I didn't write. Convincing it to reveal it's fds would take violent agressiveness.
Unix GUI code has been written since 1984 without needing threads.
Unless the logic of my algorithms is more simply represented as threads, I shouldn't use them. If I want to toss a pretty GUI ontop of a large and ancient existing codebase that is resolutely non-threadsafe, I shouldn't have to use threads.
If my development tools (linux and gdb) don't support threads easily, it'll be aggravating to use them.
Suppose I had a thread? What would it do? Wait for stuff to come from the network, and stuff it in a buffer for the main thread to read once it's ready. Well, the operating system is already doing that exact thing for me. I shouldn't need to redundantly duplicate that same behavior in my own code.
It's hardly a failing of FLTK really...?
Fltk's add_timeout doesn't obey it's own documentation. It's wrong. No one to blame but that library.
Don't you think you should background your socket code into another thread?
I don't think so. Threading could fix this particular problem, but then I'd have to use threading. That would add a lot of complexity, and overhead as slower, re-entrant versions of all library functions are pulled in. The code would also be less portable, since pthread functions aren't really compatible across different Unices.
The style I'm using now is the (originally) recommended way to write Unix network code, as per Stevens. That is, a select() call to check for network input (if any exists). It works fine with programs that use tty, ncurses, Qt, Motif, or Fltk without calling certain blocking function. Fltk plainly continues to exectute during the blocking function, as user-input continues to work. But it's neglecting its promise to keep the timeouts working.
Swing works on a single platform: Java. And not all Java, either.
I've tried out Fltk. It worked alright for simple tasks, but I couldn't figure out how to make more complex programs.
For instance, I wanted my program to use TCP/IP (regular bsd calls: socket, accept, recv). I used Fl::add_timeout and Fl::repeat_timeout to check the incoming buffer every few seconds (no need for high data rate).
That worked great- unless I made some call like fl_file_chooser, which blocks until it returns. Then the timeout function won't happen until the dialog box is done- a fact completely ommitted from the documention.
I suppose I could dig into the source code (it is LGPL after all) and try to fix it, but things like this don't inspire confidence.
If you have to lug around a huge backpack of support gear
Give him a break! He's a lone student, trying to produce a useful prototype of the HCI loop. The proposal isn't for a consumer level product.
If the demonstration is successful, then a PDA manufacturer could look into engineering the hardware down into a single handheld device, but first they've got to see the concept in action.
That's related, but different. This new device (hopefully) won't have to tilt.
In fact, the system is exactly the same as a VR viewer headset, but hand-carried instead of glued to your face.
Just like wearing eyephones and position-trackers to let you view CG people, this will use an LCD screen and position-trackers to view CG documents.
Imagine a bus rounding a corner and the text compensating by scrolling.
That could be a beneficial effect! Many people have difficulty reading small text on a moving vehicle, because the page constantly bounces around.
Possibly, this system could act as an "image stabilizer" for the text- causing the text to follow a smoother path than your actual bouncing hand.
Of course, whether or not this can be helpful depends on many factors- Does the screen have 10 millisecond updates? Does your head bounce more or less than your hand? (If they're in sync already, then you're fine.)
And how well does the inertial tracking system distinguish the gross movements of the bus from your localized jittering? (You wouldn't want to leave the POV behind you at the station where you boarded!)
You raise some interesting points, but five years could be a very short time indeed.
Five could be too little, but 80% of the time, if a work is going to earn money, it'll have done it by 5 years. (I've heard films do 40% of their earnings in 18 months of worldwide screenings, then 50% more in video rental, cable, and broadcast TV across the next 4 years). 15-25 years should be more than enough for 98% of the potential profit to come in. (And its just enough time for kids to grow into adults)
(I think the main motive for media companies to desire 50+ year copyright terms isn't to collect more profits on old works, but to reduce the free competition to their new works)
What I can't do is sell, publish, or distribute these stories.
With the pervasiveness of the internet (it's growth has only started), a major part of our communications with our friends and family takes place over it. Interpersonal communication can now be interpreted as "publication". You send someone an email/recorded story- he'll share with his friends, etc etc. Word gets around. And then lawyers drop by.
It hasn't happened much yet, but it may. We've already seen 11 year-olds C&D for their Pokemon websites...
I would be cool if after the last movie goes onto DVD if Lucas would open source Star Wars, but I think it should be his choice.
In a way, Star Wars (and Star Trek, and Buffy Vampire Slayer, and other huge franchises) are already a little like Open Source, in that dozens of authors have contributed stories that became part of the "canon". The copyright holder still holds the position of gatekeeper, but we've seen that they'll allow quite low quality work to get the stamp of approval, as long as it promises a little money. (Author Whedon has admitted that he can't bear to read the Buffy books he's allowed to be published)
Thing is, a close look a the term "Operating System" shows that it is maddeningly poorly defined.
I mean, there's almost no semantic content provided. "A system that is operating". Could be almost anything, as long as it runs.
Computer Science hasn't been around long enough for these "traditional meanings" to acquire much historical force.
There are two definitions we could use for "Computer Operating System": "the thing that runs ontop of the electronic hardware" or "the thing that runs beneath the software programs"
In the past, computers were weaker and could hardly contain a single operating system and one or two programs, so those two definitions were essentially equivalent. Today, we have more freedom to run convoluted arrangements of software, so those definitions have diverged. The question then arises: "Which definition is more useful?"
I prefer the latter ("under the software"), although the bulk of "OS research" in the past 30 years has favored the former. The software-view is from the perspective of the user of an OS, not it's creator, and thus has more utility for more people.
Do they provide process scheduling ?
The marketplace has seen many successful products labelled as Operating Systems which provided no process scheduling. Yet I can think of a few modern application programs which do provide it.
Does Emacs or Mozilla control low-level I/O or memory management ?
Both emacs and mozilla can run programs on top of them (Written in either elisp or javascript). They provide memory management to those programs. And they control their low-level I/O. You might be biased and think that "low-level" necessarily means "on the metal", but that doesn't have to be the case. It might just be "relatively low", from the point-of-view of the application program.
Look at it this way: Is MS-DOS an "Operating System"? Most people agree it is- why, it's even there in the name. Take a program like "Leisure Suit Larry"- it was written for MS-DOS, and needs an Operating System to work.
But I can run DOSemu on top of Linux, and then run Larry ontop of that. You can't claim Larry is running without an OS, so DOSemu must be viewed as providing it's OS (either by itself, or in combination with Linux, however you prefer).
From it's perspective, DOSemu is the operating system. Everything below that (Linux) is the "hardware".
Then suppose I also run Emacs on Linux, and then run Tetris on Emacs. The Tetris/Emacs/Linux block diagram is exactly the same as for Larray/DOSemu/Linux, so the argument for calling Emacs (or any sufficiently expandable software) an Operating System is quite strong.
(And besides! You don't have real "hardware" at all! Those wires and chips are all emulated, running on top of The Matrix)
I don't think we need current stuff, but it is unhealthy for a culture to have so many of central, identifying stories be the property of a single, centralized owner. It is, frankly, undemocratic.
I agree that current works need copyright protection for at least a few years, but I think that terms should expire soon enough for adults to create derivative works from stories they've heard as children.
Since ancient times, children would listen to stories told by their elders around a campfire, and hold those imaginary heros in their hearts and minds as they made their way in life. Eventually they grow up and become the elders, and tell new stories of the cultural role-models and villians- based on the earlier tales, but colored by their own life experiences.
This is part of how a culture can grow and mature over the ages. Every listener will be a teller.
But, in the modern world, we sent our children to sit around a film screen and watch Star Wars. The wiles and wisdom of Han Solo and Ben Kenobi formed their subconcious guides. When they grow up and are parents themselves, can they give their children a modified version of their own favorite tales?
No. Only Lucas can do that, and the impotent parents cannot contribute even a little to the dreams of their children (no more than choosing between Star Wars and Harry Potter).
If a helpless child is infused with corporate propatainment in his formative years, we at least owe him the right to someday share the stories they inspired in him, instead of stepping back onto the media industry's treadmill.
"You're buying a movie ticket, and selling your dreams"
That's not really true, although it happens that way in many cases.
Looking at the example of Java vs C++ specifically, the slowness is actually (sometimes) due to the heavier semantics of the Java language. For instance, take a complier (like gcc/gcj) which can turn both C++ or Java code into a finished native executable.
Program some time-consuming math alg (Fibonacci, Newtonian integration?) in both, and measure the runtimes. Java will be slower by almost a constant factor.
Why?
If we accept this point, that some compilers are simply better than others, than we can imagine a situation where compiling C++ into Ada is beneficial: You recieve a powerful NSA-surplus machine which only runs Ada, and you'd like to run Apache and Mozilla on it. Rather than re-writing those in Ada, it may be faster to use a "cxx2ada" preprocessor on the application source code.
(That's not likely to happen, but with the new "hardware JVM" chips coming out, you may find a computer which can only run Java programs, and compiling C++ into Java might be a viable intermediate build-step)
You'd have to dig deeply into both compilers to find why.
OTOH, the reverse assertion, that source code which is further separated from the hardware will perform faster, has sometimes been demonstrated to come true. (Although most modern compilers are still hobbled by legacy application-building techniques, and can rarely achieve such awesome successes)
That document also states that Congress can make no law which takes effect before it was passed ( ex post facto ). Plainly, extending copyrights on an already published work does this (just as surely as extending jailtime for an already-convicted felon would)
Further, the Constitution forbids the government from seizing private property without fairly compensating the owners. I had been in possesion of some "futures" on copyrights for a variety of works published in the 1920s, such that I would today be able to publish my own editions.
Congress has taken away that intellectual property from me- not only can I no longer publish my own versions, but I must pay money if I wish to read "Detective Comics #1". They've deprived me of something of measurable dollar value, yet I have recieved no compensation.
Where can I send my bill for 1/284,000,000th of the total worth of all publications from 1921-1941?
Life expectancy went up dramatically in the previous century
The original copyright term was 28 years- below normal lifespan, even in 1500. Copyright today is 90 years, above expected lifespan for children born in 2003.
If you want a law to be based on how long someone will live, then how about basing it on the lifespan of the people in question? No need to keep passing laws whenever there's an incremental medical advance. (However, copyrights should not be based on the lifespans of the author, or any of his descendants. That unfairly discriminates against authors who are old, sick, or childless, independent of their writing skills)
so it's only natural to extend copyright term in accordance
What else might be natural? Digital computer technology has greatly accelerated the speed that works can be authored, published, marketed, and sold- you need less time to make a fair profit today than you did in the past. So less copyright term seems natural.
Also, the US population has increased more than 10-fold in two centuries. Copyright was once a bargain between 1 author and 20 million readers, who agreed not to reproduce his work for 28 years each, or 560 million years collectively. To account for the same net deprivation with today's population, the copyright term should be only 1.9 years. That would balance the author's profit-opportunity across history. Seems natural to me!
Eldred will have access to the material he wants in a few more years
Scientists have determined that "few" means fewer than 7, even in the most extreme circumstances. 20 years at least it will take, and given the pattern established in 1976 and 1998, the actual time will be much much more.
go to Congress and get the extension laws repealed
Oddly, they can't. Or at least that's what the legislators will claim if you bug them about it.
You see, it's unconstitutional to change the length of existing copyright (either to lengthen, or shorten, as Stevens noted in his dissent, page 5). To do so is both ex post facto law and seizing private property without compensation.
The Court let them get away with it this time (and in 1976), because that was the pro-corporate thing to do.
So even if you got a whole new body of congressmen to repeal the law, the MPAA could appeal it to the supreme court and get it tossed out by arguing the same case Lessig just presented! And of course, if it's a corporate attorney before them, we can be sure they'll all nod along like good little puppies.
which are real property and cannot fall into the public domain.
Real property (real estate, or physical goods) can and has been seized by the government for the good of the nation. Claiming eminient domain is a traditional congressional power.
However, they always have to compensate the property-holder the fair market value of the goods taken. Lets see, I was due to recieve unlimited reproduction & derivation rights for Detective Comics #1 by now... I wonder if Congress has mailed my check yet?
"Allow"? I was under the impression that they snuck in.
They sneak in initially, but once you get past the INS-fortified towns along the border, average American citizens don't treat their presence as a crime at all. (More like a *wink-wink* opportunity to avoid the minimum wage and income tax laws when getting some lawncare done)
I had expected the events of Sept. 11, 2001 to make the US public less accepting of illegal immigrants in their midst, but they still don't really care (if the visitors don't look arabic)
If your snowplow driver lets slip that his visa's expired, do you sneak to a phone and dial 911? No, most Americans would find that rude and petty. And the local police won't be really enthusiastic about investigating either. (Unless they had an INS task force visiting their department at the time. In big cities this can be a full-time situation)
Even when it is enforced, illegal immigration is treated differently from other things which are "illegal". Most illegalities lead to a guaranteed jury trial, then a fine or incarceration. Catch someone red-handed at infiltrating the US borders, and he's dumped on the streets of Mexico that same night.
I'm not very in-touch with American opinions, but occasionally I glean something from mass media. I was really astounded last year upon reading a sob-story on the front page of a major newspaper, detailing the plight of illegal-immigrant teens. Apparently these poor people (brought to America at age 8-12, say) were leading successful high school careers, but then were unable to collect college scholarships!
Fight back, Palm; Apple beat Xerox, so you can too.
s/beat/paid/g
Microsoft, on the other hand, beat Apple on the same score (demonstrating to Apple that they hadn't really needed to pay Xerox in the beginning). Amusingly, the FSF and related groups was supportive of Microsoft in those instances
The interesting problem with the Fitaly layout is that it is also patented! (number 5487616)
Fitaly corporation seems willing to license it for low rates, so Palm could afford it if they wanted to go that way.
However, if Fitaly (or Xerox Unistroke, or any other patented input method) becomes popular/standardized, then Free Software programs will be unable to clone the UIs of the next generation of Palmtop software. (In the past it was shown illegal to copyright a UI, but patents still seem to apply)
I can already find one Open Source project that's using the Fitaly layout, in violation of the patent. I wonder if fitaly.com will send them a C&D as soon as they notice, or if they'll wait for the input method to become popular on Linux PDAs before yanking it away.
Or make sure you know what kind of better hardware you could buy, if needed.
When developing a system you should try to overload it so you can recognize what a failure state looks like. This may give your engineers valuable insight.
What is the resource that gets exhausted first? What is the system's behavior when it is completely overloaded? Does it just stop functioning, or does it lose data? Or maybe generate bad data?
These things could be nice to know, and may suggest quick improvements so that, if 6 years later the customer puts in 20 times as much usage as was originally budgeted, the failure isn't completely embarrassing.
Obviously, in the case of "Joe Blow" hooking up a Windows 2000 box to the Internet on his home DSL connection, enabling various server services, and not keeping up with the updates, there is no such accountability. If his machine is compromised, and then used to attack other machines as part of a DDoS attack,
If his machine can be compromised for DDoS, it can also be infiltrated to record his credit card numbers and publicize them. (Or make random purchases in his name, with the goods not necessarily coming to either him or the hacker).
If hackers were basically unpunished by the law, then people who install zombie-able machines will soon face the impromptue penalties of getting robbed blind.
It is the hackers who create the need for security,
If there were zero hackers alive today, and if the concept of "computer intrusion" were just a theoretical problem with no examples from precedent, there'd still be a need for security.
People won't act like it, but they do need it.
2 years ago the concensus was that knives were fine for airline passengers. That was wrong, they needed to be banned, but the public in general didn't care.