I would say we are in the green. We are not voilating anyone's copyright and we are not circumventing any copy protection scheme.
This has been a big point for me from the beginning. Some people wanted us to distribute patched versions of the original firmware (language fixes, charset fixes etc), but that would have violated Archos' copyright so we never did that.
For those of you looking for a REALLY small system, check out the Nano Module from Nano Systems (a small Swedish start-up).
It's a Pentium 266-class PC in a 5.25"-size package, containing:
National Geode SC2200 CPU/microcontroller (Pentium compatible)
Standard 168-pin 133 MHz SDRAM DIMM
Compact Flash (ships w/ RedHat on a 64MB card)
Dual 10/100 ethernet controllers
USB controller
IDE33 controller (for external disk)
VGA controller
1 PCI slot (w/ 90 degrees angle adapter)
It's powered by a small 12V laptop power supply. No fans. It's definitely not for the 3D game boys in the crowd, but for embedded use or a firewall/mail gateway/proxy/whatever it's close to perfect.
It's so small you can put it in one of the 5.25" bays of your computer while developing. It even has a connector so you can power it from the PC power supply.
Many don't seem to reflect about this, but the one thing that sets GPL and other software libre licenses apart from the likes of Oracle or Microsoft is that GPL grants you additional rights beyond what normal copyright law does. Most other licenses try to limit your rights.
All works are automatically copyrighted. I may give someone a limited right to distribute my work, tied to a number of conditions. Breaking these conditions means you no longer have that right.
The issue then boils down to standard copyright law, under which you are absolutely denied any right to distribute the work beyond Fair Use.
In short: Accept the GPL and enjoy the right to distribute my code, or accept plain copyright law under which you don't have that right. There are no other alternatives.
The idea was to have a time format that spans the entire 20th century (almost, it starts at 1901-12-13 20:45:52) and some of the 21st.
Eventually, The Bearded Men of POSIX decided that negative time_t was a bad idea, yet they kept the variable type signed. Probably they just chickened out in fear of breaking existing systems.:-)
And what about databases that store dates BEFORE January 1 1970 (in unix time format)?
They belong in the 'fubar' category. time_t was never intended for database use, and shouldn't be used for dates before 1970 either. Read the POSIX spec:
2.2.2.77 seconds since the Epoch: A value
to be interpreted as the number of
seconds between a specified time and the
Epoch....
If the year < 1970 or the value is
negative, the relationship is undefined.
You can make anything break given enough foolishness.
Then you have the other music player that costs less and plays every audio format under the sun. Hmm which one do you pick?
Since the uncrippled player gets sued out of existence, there will be no choice. Just wait a year or two and Hilary Rosen's "if we don't approve it, it won't happen" will be the law.
Don't compare it with Divx, compare it with DVD. We've already lost that battle, but does anyone care?
The author controls the disposition of his work. If he wants it released only on Windows, tough. He always have the right not to buy it. You never have the right to tell him how to apply his copyright.
Absolute bullshit. Copyright is not a license to control customers' lives. If I buy a book from you, it is legally mine. I may read it, burn it, bury it or feed it to the dogs. The only thing I'm not allowed is to sell copies of it. Hence the name copyright.
In the portable MP3 world that would cost me, assuming $50 per 64 meg memory card and I can put what, 2 CD in that space?, that would be 25 cards or $1250.
Think again. Current MP3 players store 6 GB (100 CDs @128kbit), cost $350, are about the same size as your MD player (smaller but thicker), double as portable harddisks and last for 8 hours on a charge.
Let's assume a very optimistic 10MB/s
[...] So it would take me about four hours to fill that disk with data. [...] What I'm saying is that we're approaching storage densities where our current data transfer busses simply can't reasonably cope anymore.
Actually, our busses are fine. SCSI runs at 160 MB/s (14 min for 140GB), ATA is at 100 MB/s (23 min for 140GB). You're referring to the storage devices, which currently peak at somewhere around the 20 MB/s mark. Nevertheless, the busses will improve, too. Serial ATA promises 6 gigabits in the not-so-distant future and surely there will be competitors.
ALL medical embedded stuff runs OSE by Enea systems.
Yeah, right. I happen to be working on "medical embedded stuff" right now, for one of the major players in life support systems. And we aren't using OSE anywhere. Not linux either.
Here's a great quote: "it is impossible for user processes to corrupt the OSE kernel."
It's called "marketing", ever heard of that? Don't believe everything you read in glossy brochures.
I really don't see what you guys are getting so excited about. What sets QNX apart from all the other commercial RTOSes out there? It's a commercial endeavour! Yes, you can download a demo disk. But you don't get full source and you're not allowed to use it for any real work. Sounds pretty much like VxWorks or any one of the dozens of other competitors to me. Call them, they'll send you a free demo too!
Read the FAQ. Their opinions are stated in plain view:
Q: Why doesn't QNX provide source to the kernel and other core OS modules?
A: Because QNX developers don't need kernel source to extend the OS.
Anyone who's ever done serious work in embedded systems know the kernel source is absolutely essential for debugging, not only the application but also the kernel. All OSes contain kernel bugs. They're a pain to find and fix without source, and those of us who've been there are not going back lightheartedly. You all know this, that's why we're embracing open source. How come so many of you are now eager to jump back into the dark hole that is proprietary software?
For embedded work, there's
ECOS already. It's Free Software and runs on a dozen different CPUs, with new ports coming all the time. If you want the 3D acceleration, anti-aliases graphics and macromedia player, you're probably not looking for embedded stuff in the first place.
Sure, QNX is fun. Play away. But it isn't the future.
Slightly offtopic, but interesting IMHO. Quoting the brief:
Under the seminal Sony case, as long as a technology is "capable of substantial non-infringing uses," a provider making a technology available cannot be liable for copyright infringement, even where it may have encouraged infringing uses and the technology may in fact have been used for infringing activity.
What's the difference between the Sony (Betamax) case and the DeCSS case? They can both be used to infringe copyright, but they both have substantial non-infringing uses. Even more important:
Under the Sony doctrine,
it is enough to show a single potential non-infringing use of social or commercial importance. Vault v. Quaid, 847 F.2d 255, 266-67 (6th Cir. 1988) (because technology which enabled unfettered copying of copy-protected software could be used to make archival copies of the software, the product was non-infringing; relative proportion of the single lawful use to unlawful copying not even considered);
commercial software should be warrantied, byt free software should be exempt, for who knows what reason
Isn't it rather obvious?
If you pay $20K for a new car, you expect to get a working car. If it doesn't work, you have the right to repair, refund or replacement. It stems from the basic concept that your money should be exchanged for a good of similar worth.
If, on the other hand, I give you a car, you have no right to expect anything from it. It could break down in a mile for all you know, and there's nothing you can do about it short of giving it back or rejecting the gift from the start.
Anything gratis is a gift. Gifts are never warranted. (If they were, who would give toys to kids?;-))
Use emacs python-mode with the tabs meaning a certain number of spaces.
Sadly, Emacs is crippled by the Python syntax. Since there is no way to know where a block ends, emacs can't help much with indentation of Python. You can't use C-i to auto-indent to the correct level, you can't use indent-region to adjust a whole block/file of code to your liking and you can't have it insert a new block around an old block (like wrap a statement in an 'if' condition).
In short, you have to sit there and pad and visually verify each and every line manually. And you people have the balls to claim this is a feature. [shakes head in disbelief]
And don't get me started about the idiosyncrasies in Pythons array indexing...
I don't know whether Delphi suffers from the same malady that Borland C++ suffers from, but Borland C++ defaults to 16-bit integers at the command-line, but 32-bits at the IDE level.
That's a pretty clueless statement coming from the director of a programming competition. Borland C++ in fact ships with two command-line compilers:
bcc.exe - The DOS compiler, with 16-bit integers and pointers suitable for DOS' segmented memory model.
bcc32.exe - The win32 compiler, with 32-bit integers and pointers suitable for Win32's flat memory model.
First rule of programming: Choose the right tool for the job. Apparently, this fellow skipped that class.
Slashdot Mirror
I would say we are in the green. We are not voilating anyone's copyright and we are not circumventing any copy protection scheme.
This has been a big point for me from the beginning. Some people wanted us to distribute patched versions of the original firmware (language fixes, charset fixes etc), but that would have violated Archos' copyright so we never did that.
For those of you looking for a REALLY small system, check out the Nano Module from Nano Systems (a small Swedish start-up).
It's a Pentium 266-class PC in a 5.25"-size package, containing:
It's powered by a small 12V laptop power supply. No fans. It's definitely not for the 3D game boys in the crowd, but for embedded use or a firewall/mail gateway/proxy/whatever it's close to perfect.
It's so small you can put it in one of the 5.25" bays of your computer while developing. It even has a connector so you can power it from the PC power supply.
I bought one a couple of weeks ago, and it's sweet. Here is a picture of the internals.
All works are automatically copyrighted. I may give someone a limited right to distribute my work, tied to a number of conditions. Breaking these conditions means you no longer have that right. The issue then boils down to standard copyright law, under which you are absolutely denied any right to distribute the work beyond Fair Use.
In short: Accept the GPL and enjoy the right to distribute my code, or accept plain copyright law under which you don't have that right. There are no other alternatives.
The Archos players are supported by Linux as of kernel 2.4.8.
See here for more info.
The bottleneck is the RAM, not the bus. Making a fast bus is easy. Making fast RAM is hard.
Eventually, The Bearded Men of POSIX decided that negative time_t was a bad idea, yet they kept the variable type signed. Probably they just chickened out in fear of breaking existing systems. :-)
They belong in the 'fubar' category. time_t was never intended for database use, and shouldn't be used for dates before 1970 either. Read the POSIX spec:
You can make anything break given enough foolishness.
Umm, no. That'd be a 31-bit number, as is the case with time_t which is a SIGNED integer. A 32-bit value can hold 2^32-1 = 4,294,967,295.
Can anyone tell me why this is a Real Problem? The obvious solution is to simply change the compilers from:
typedef long time_t;
to:
typedef unsigned long time_t;
And we can merrily keep using time_t on our 32-bit systems until 2106.
Yeah, fubar systems will break. And yeah, we'll have to change some kernel API parameter types. Cry me a river!
Since the uncrippled player gets sued out of existence, there will be no choice. Just wait a year or two and Hilary Rosen's "if we don't approve it, it won't happen" will be the law.
Don't compare it with Divx, compare it with DVD. We've already lost that battle, but does anyone care?
Absolute bullshit. Copyright is not a license to control customers' lives. If I buy a book from you, it is legally mine. I may read it, burn it, bury it or feed it to the dogs. The only thing I'm not allowed is to sell copies of it. Hence the name copyright.
How about MySQL? Significant enough for you?
There are dozens more, but you didn't even bother looking did you?
Think again. Current MP3 players store 6 GB (100 CDs @128kbit), cost $350, are about the same size as your MD player (smaller but thicker), double as portable harddisks and last for 8 hours on a charge.
Umm, ignore parent article. I've had too little coffee today. :-|
Just for the record, your numbers are only odd because your accuracy is off. Way off.
Add some decimals to your numbers and you'll soon find that "cracks/%" is, naturally, a constant equal to the total sum of cracks!
OS......cracks..%........cracks/%.....norm
nt......4750....57.9268..8200.004143..1.000000
linux...1750....21.3414..8200.024365..1.000002
solaris..700.....8.5365..8200.082000..1.000010
bsd......500.....6.0975..8200.082000..1.000010
others...500.....6.0975..8200.082000..1.000010
TOTAL...8200....99.9997
What was your point again?
Actually, our busses are fine. SCSI runs at 160 MB/s (14 min for 140GB), ATA is at 100 MB/s (23 min for 140GB). You're referring to the storage devices, which currently peak at somewhere around the 20 MB/s mark. Nevertheless, the busses will improve, too. Serial ATA promises 6 gigabits in the not-so-distant future and surely there will be competitors.
Yeah, right. I happen to be working on "medical embedded stuff" right now, for one of the major players in life support systems. And we aren't using OSE anywhere. Not linux either.
It's called "marketing", ever heard of that? Don't believe everything you read in glossy brochures.
Read the FAQ. Their opinions are stated in plain view:
Q: Why doesn't QNX provide source to the kernel and other core OS modules?
A: Because QNX developers don't need kernel source to extend the OS.
Anyone who's ever done serious work in embedded systems know the kernel source is absolutely essential for debugging, not only the application but also the kernel. All OSes contain kernel bugs. They're a pain to find and fix without source, and those of us who've been there are not going back lightheartedly. You all know this, that's why we're embracing open source. How come so many of you are now eager to jump back into the dark hole that is proprietary software?
For embedded work, there's ECOS already. It's Free Software and runs on a dozen different CPUs, with new ports coming all the time. If you want the 3D acceleration, anti-aliases graphics and macromedia player, you're probably not looking for embedded stuff in the first place.
Sure, QNX is fun. Play away. But it isn't the future.
Really? You can't have many phones where you live, then...
Yes, but that doesn't help much if CSS2.0 turns out to use for example 64-bit RC5. We've been brute-forcing that for the last two years...
Only the parrot doesn't stomp its feet. It says "four". Not much room for reading reactions there.
What's the difference between the Sony (Betamax) case and the DeCSS case? They can both be used to infringe copyright, but they both have substantial non-infringing uses. Even more important:
Again, isn't this exactly the case with DeCSS?
Isn't it rather obvious?
If you pay $20K for a new car, you expect to get a working car. If it doesn't work, you have the right to repair, refund or replacement. It stems from the basic concept that your money should be exchanged for a good of similar worth.
If, on the other hand, I give you a car, you have no right to expect anything from it. It could break down in a mile for all you know, and there's nothing you can do about it short of giving it back or rejecting the gift from the start.
Anything gratis is a gift. Gifts are never warranted. (If they were, who would give toys to kids? ;-))
Sadly, Emacs is crippled by the Python syntax. Since there is no way to know where a block ends, emacs can't help much with indentation of Python. You can't use C-i to auto-indent to the correct level, you can't use indent-region to adjust a whole block/file of code to your liking and you can't have it insert a new block around an old block (like wrap a statement in an 'if' condition).
In short, you have to sit there and pad and visually verify each and every line manually. And you people have the balls to claim this is a feature. [shakes head in disbelief]
And don't get me started about the idiosyncrasies in Pythons array indexing...
That's a pretty clueless statement coming from the director of a programming competition. Borland C++ in fact ships with two command-line compilers:
First rule of programming: Choose the right tool for the job. Apparently, this fellow skipped that class.