That "tainted" just gave me an idea... Now, someone correct me if I'm wrong, but
I think the NVidia drivers get away with this because they are two parts: a kernel patch, and a separate copiled kernel module. The kernel patch IS a part of the kernel and so must be publicly released, blah blah blah, or you get toasted by the FSF - but the kernel patch is really just a bunch of hooks that let the binary driver interface with the kernel.
So, the patch is just a publically available interface to the kernel. And the the kernel module is a module (which, if above posts are correct, Linus has blessed as not being subject to the same license as the kernel) which uses only published interfaces (so there's no longer any argument that it's linked directly to the kernel).
The risk here is only that the patch gives away too many implementation details - but there also happens to be no requirements that you not obscure the workings of the interface (unfortunately).
An embedded device may be the one thing that WOULDN'T have to be called GNU/Linux...
Because what GNU programs does an embedded device need? I'm pretty sure you're not loading a compiler onto it, nor a debugger, nor just about anything else...
The only thing I can think of offhand that you would even use is glibc for userland programs. And I have a hunch most embedded devices don't use a whole lot of glibc - and could very easily be changed to another c library. The kernel doesn't use it... and for an embedded device, most of the interesting stuff will be in kernelspace anyway!
Not having written for the kernel myself, I'm not 100% sure, but...
I'd guess that anyone who submits a patch back to Linus is de facto making him a copyright holder of the code in the patch (well, assuming he is also the holder of the original code... but I'm pretty sure he has the copyright for the 0.01 kernel!). IANAL, but I think it's a safe bet that sending a patch (w/o an explicit license agreement) constitutes a gift.
All linux patches then work on the assumption that Linus is a nice guy and will push the patch out in a publicly-licensed release of Linux. And this is so far a completely safe assumption: Linus has not shown any commercial interest in the Linux source code, but has shown extreme interest in making Linux as good a kernel as possible.
Although, for comparison, look at SourceForge: when they went proprietary, they stripped out all changes to their original codebase and went back to a version they owned the copyright for. Very conservative approach, but you'd never find a hole in it...
It's
(2) publish a user's guide and call it a "cheats" book.
The point is, for all the really good games I already figured out everything in the strategy guide (yes, I skimmed it at the store... and didn't learn a thing). Civilization II, anyone? That game also happens to be the last one that came with any sort of legitimate guide-as-a-guide - something like Homeworld has a guide that basically says, "here's how to install, here's what a bunch of keys do (in doublespeak - the descriptions only make sense after having learned to use the key...), and here's some concept art."
Hence, the reason that I've bought more games that are more than a year old than new releases.
And at this point, I'd even suggest digging up the Standard C library implementation of quicksort - I remember coming across something somewhere that mentions that the standard implementation is a hybrid quicksort that is actually more optimal than the theoretical / canonical quicksort. I'm thinking it was either a hashed-quicksort or a multi-quicksort, but I could just be making those up - and if so I'm sure someone will correct me. As joto mentions, any sorting is still at best O(n log n), just a lower constant term, and a greater likelyhood of better-case over worst-case.
The original poster mentions that his algorithm isn't a simple in-place sort - which, based on my best guess, means that he's probably sorting pointers quickly then re-arranging the elements in a single pass. Maybe, maybe not - but still no faster than O(n log n).
But at this point, I'd start worrying more about scalability, memory usage, etc. - one of the nice things about quicksort is that it is so general, it can quickly be adapted to low-memory or multi-processor implementations. Sure the algorithm is faster - but faster does not necessarily mean better. If my implementation guess above is correct, the algorithm would use significant additional memory, making it unsuitable for sorting simple numbers (where the pointers would dwarf the elements themselves) or for use in low-memory (i.e. embedded) applications.
The true test of whether the algorithm is innovative is: 1) have a professor look over it, and then 2) get a (respectable) journal to publish it. If it passes the peer review required for publication, then it's an innovative algorithm. If not, it's one of thousands of alternative implementations of the same idea. Clever, maybe even useful, but not spectacular.
I suggest that the parent poster is completely correct - you just need to re-think your idea of a web site.
At the very least, I'd suggest an ADA-accessible front page that basically says, This is a visual arts web site, if you proceed further you probably won't get anything out of this site. And this can be as simple as a message in an ALT tag.
Your front page needs to be accessible to anyone - including people using Lynx. Why? If I'm connected to a remote server, there's no way I'm going to run a complicated Mozillia-over-X-over-SSH setup just to hit the internet while I'm looking for "deco wallpaper". I'm going to use Lynx! And even if all you do is tell me "you can't get anything out of my site if you can't see this image", you've still done more than 95% of the sites out there - and with 3 minutes of your time. I'd recognize your site doesn't have value for me - and I'd thank you for saving my time.
These technologies:
Java - bloated, unwieldy, and EXTREMELY difficult to set up or upgrade. Rely on it, and you've alienated 99% of web users
Javascript - probably THE easiest way to get viruses into a system. Slow, unwieldy, bug-prone (I get Javascript errors every time I visit Microsoft's site, even under IE), and really only useful for fancy, glitzy stuff whose only purpose would be to cover up an otherwise poorly designed site. Many security advisories simply suggest disabling Javascript - you shouldn't rely on it at all, or EXPECT to be disappointed.
Flash - a useful technology. IF the plugin works. A very big if.
Graphics - Relying on too many graphics makes your site slow to load. And glitches in displaying graphics are one of the biggest browser-compatibility issues around. I don't have your high-speed internet connection, I don't really want to download gratuitous graphics. If I want to see something I'll click on it - don't make me download megabites of junk.
streaming media - Let me repeat: very few internet users have the bandwidth for decent streaming media. And if it's integral to your site's design, you have a poorly designed site.
Please, before you go out and bash someone for complaining about your site, take a second to think about what they want. Of course a blind person or text user doesn't want to poke around your fancy, glitzy visual-arts site. But he certainly doesn't want to become trapped in it, waiting for slow and useless graphics and extensions to load and unable to navigate without SOME text.
A site doesn't need a complete make-over. It just shouldn't be blatently offensive to blind or text users. And if you can't spare the 5 minutes of your time it takes to make a few simple changes to your front page alone, I truly pity you - and hope I never come across your site.
You don't have to bend over backwards, but do not dismiss a legitimate complaint without hearing it out. It's rude, and frankly it's exactly the reason the ADA was passed in the first place.
No, but RAM has pages. 4K on x86, as I recall. You can't pull anything smaller out of the kernel - though you could pack it into another process's address space (but that defeats the purpose of a small executable anyway).
And no, you can't change the page table size, it's hardware-dependent. Most of the other archs seem to have similar or larger pages, too.
Why do I know this? It's "write your own VM" month in my OS class. Next week we get to start swapping out to disk...
It's perfectly legal for me to download the pre-patch kernel - probably from some archive somewhere. And it's perfectly legal for me to download the patched version. But, by running diff and creating my own patch, I've just discovered how to bypass the security. So just running diff is a criminal act?
The only way I see out of this is even worse: it would be illegal to have an unpatched kernel! Since an unpatched kernel could be diffed against a patched kernel and break the DMCA, the unpathced version must be illegal. Then, to enforce this, OS vendors would be able to require (a la the new MS EULAs) that you allow them to arbitrarily update the kernel without telling you.
The scary thing is, this all makes sense. Can anyone see another way out? (besides declaring the DMCA illegal - that one's easy!)
VERY good pints, though I think I might comment/elaborate on a few of them.
Remember that just because a law is "bad" and horribly unbalanced towards lobbists doesn't make it illegal unless there is some specific legal reason the law is unconstitutional.
Exactly right - and the Supreme Court would be correct in maintaining a law passed entirely by lobbyists over the objections of 90% of the American people. Why? Because the Constitution doesn't prevent it. And in the U.S., there is no law higher than the Constitution. Immoral, yes, but illegal, no.
I even think this is necessary - because think of the opposite. How many laws were passed with the support of 90% of the people over the objections of lobbyists? Answer: a whole bunch of important ones. Taft-Hartley (sp?) on labor laws, Sherman Antitrust, Amendment 19 (?) (women voting). If we want the people to overrule lobbyists sometimes, we need to let lobbyists overrule us sometimes - because the people can always re-elect a more favorable Congress.
So far it seems the court is buying [the limited powers] argument. The court seems to be unsure though if it has any power to do anything about it. This is good news to Lessig, because it means the court buys the limited power argument.
It's actually a very good thing the court is unsure if it has power. In fact, the court almost ALWAYS asks if it has power to intervene... you'll notice that since the court rejects upwards of 90% of the cases appealed to them. A big part of those appeals are cases where the court doesn't feel it should issue a ruling.
Another example, the Florida election stuff. That debacle was an example of the Court deciding it DID have power (and deciding very quickly, and without much justification). IMO, a more correct ruling would have been to just say that "how a state runs its elections is its own business" and just refuse the matter entirely. This was basically the minority opinion. I'm not arguing the merits or outcome of the decision (I would have remanded it with instructions for a re-hearing, but I'm not a Supreme), but most of the mess occurred because the Court jumped too quickly into an area it shouldn't have entered.
Perhaps the Court even feels burned because of the Florida voting episode, and is trying to be extraordinarily cautious with its power (and even more so because more eyes are looking for faults!)
The case was also helped by a government bumble. The government argued that there is no constitutional limit on the ability of Congress to extend copyright, thus the extention was legal. This actually helped Eldred because the court did not like this view at all. The court did not support the idea that the constitution limits the powers of Congress, but that Congress gets to set what the limits are. In effect, the government proved Eldred's point themselves.
As much as the Court hates giving itself power, it likes giving someone else power even less. A better argument might have been that there is a limit but Congress should determine it; but instead, the government chose to try to defend the whole thing. It almost feels like they're giving the case away. Thus: I'm watching carefully to see if someone responds to "an erosion of copyright law" by introducing a new, draconian law with overwhelming support. Cyncial, yes, but I'm very cynical when politicians are involved.
Very good summary, my complements to you good sir.
The problem comes in when the CPU does a context switch to a different thread/process or handles an interrupt during the called function. The context switch is responsible for saving all the registers currently in use onto the thread's stack (in main memory). So this extra CPU-internal stack of registers suddenly needs to be saved too... and you've just doubled the size of the stack push.
The other problem is more serious: most programs go many, many function calls deep. If each function pushes it's own registers onto a register stack, the stack will be proportional to the execution depth. You've seen Java error messages - those are easily 10-20 calls deep. OSes can go much, much deeper... and I'm not even getting into recursion, which can go thousands or more levels. Try to keep all these registers on a special register stack would mean you have a 5K or so stack register (and the architects are already fighting over a 20K L1 cache!), AND as mentioned above, this special stack register has to be flushed to main memory every context switch.
Although, I wouldn't be surprised if this idea were actually implemented on much older (i.e. pre-386) type processors that were pretty much single-threaded anyway. It's not a bad idea, it's just that it's too out-of-date for a modern OS.
There most certainly is a way... my VMWare images are on a third partition mounted at C:\VMWare right now. It does require the base filesystem (in this case on C:) to be NTFS, though. (not sure if the mounted one has to be NTFS or not)
But, like all truly useful Windows tools, you gotta dig to find this one.
Go to Control Panel/Administrative Tools/Computer Management. Choose the Disk Management item (left pane), which will bring up a diagram of all your partitions in the lower-right pane. Right-click on any available partition (except your C: partition!) and choose Change Drive Letter and Path. When you try to add a new location, it will allow you to specify a directory as a mount point - a la the *nix mount command.
Plus, you get a little drive icon in the middle of a list of folders - I'll bet very few people have one of those! This way only works on physical disks as far as I know - but it's a great way to make a drive bigger when you run out of space.
My prediction for the future: the next edition of Windows will encourage you to mount your/hom... er... C:/Documents and Settings/ directory from Hailstor... er... your Really Big And Expensive Windows Server!
By posting a link on/., and knowing that the result (particularly from a bunch of suitably angry geeks) is a slashdotting and is enough to crash servers, a lawyer would probably argue that the poster DELIBERATELY caused a DoS attack. And suddenly this e-mail becomes a prelude for a more serious - and documentable - attack.
Likely to happen? I hope not. But chilling to think that it might.
The difference with a photocopier is that it's actually legal - most (as in almost all) of the people who use a photocopier fall under fair use. It's legal to copy a few pages (maybe even most of a chapter) out of any book at your local library - heck, that's why they have photocopiers there in the first place! The reason it's legal is that you aren't copying the entire book - you're keeping a part of it for reference (that one recipe you've been looking for the past three years and know you'll forget if you don't copy it), or for criticism (you're a student who needs to quote a few passages out of War and Peace to demonstrate how long-winded it is), or for information (you copied the map of Texas out of the encyclopedia because you're about to vacation there). Of course you can't legally copy the entire book, but everyone knows you won't because you'd pay more in photocopier fees than you would for the book itself.
Fair use even has a special carve-out for academic uses - that's why a teacher can photocopy a "classroom set" of articles out of Scientific American to assign as reading, and hand them out to students. (There are catches - the teacher is supposed to destroy all copies after they are done, and they ususally don't try... but what student saves their classroom readings after the test anyway?). That's also why you can block-quote someone else's writing - it's legal as long as you give proper credit.
How sure am I of this? IANAL, but I've had to read the US Copyright Office's publications for a research project, and all this is laid out pretty clearly.
I think I've strayed from the point... which is that photocopying is a completely legal "fair use". You just can't make the same case for copying software... it's not like you're making a limited use of a small part of it, when you copy software you copy the whole thing!
Cash. In low bills - like $1 bills. And quarters - lots of them.
Really. The campus probably has an ATM machine, it's easy to get $5s and $20s out of a bank account. But the change machines always seem to be out of money or out of service. I don't want to spend money on something to break even a $5 into bills small enough to go into the pop machine...
Yeah, small bills and quarters. Better than lots of money any day of the week.
That technological lead will last for all of 3 seconds until someone puts out an identical product without DRM.
That's exactly the point - with our highly touted free market system, everyone is free to make a better product. "Better" in one way might be "has more features, but isn't as nice for geeks to use" (i.e. DRM features), but to another consumer "better" might mean "same as the other guys, except w/o the DRM junk". In a free market, both of these products are perfectly valid.
As for someone putting out a non-DRM product 3 seconds later... How about the difference between a game console and a PC? I would argue that neither one is really more technologically advanced than the other (beyond maybe a few-month gap for a product life cycle). But because consoles are a "better" product (for playing a game on a large-screen TV with an interactive controller, and for not crashing as often, and for being cheaper than a PC), they have a strong market. PC games have a market because they are "better" than games at 1) saving, 2) multiplayer, and 3) better controls for complex games (just a few ideas, I'm sure there are more). Just copying a product and stripping DRM out of it doesn't necessarily mean the stripped product can compete with the original.
What I'm suggesting is that the DRM-favoring industries will only be successful if they can produce a product that is innovative enough to be BEYOND what someone else can copy "in 3 seconds". Some industries can produce a more innovative product - you go to a restaurant for well-prepared food and atmosphere, when it's cheaper and you have more freedom to choose when you go to the supermarket.
But, being a Slashdot user, I am obliged to present these two observations (because I'd get flamed for anything else?):
The industries that rely on DRM don't seem to be innovative enough to produce anything useful.
The DRM industries run with such a crippled, non-standard business model that they could never survive in any real business beyond their cute, rights-managed-by-overprotective-laws world.
In other words, I defend the right of companies to produce products with DRM. I just don't think they are competent to be successful with it - I think that, fairly soon, someone else is going to come along, with a better vision of how to run a business in a content-type world, and will drive these DRM-favoring companies straight out of business. Some companies are already successful - look at Red Hat, they are "innovating" the Linux OS with something of value (polish, and support), and therefore have a product to sell. I think the only reason there aren't more Red Hat-type companies in this world is that there hasn't been enough time to work through the trial-and-error process of coming up with a way to do business.
The only way any industry will ever get everyone to accept hardware "rights management" like this is if they make a better product. That's right: they have to make some DVD-player system (or whatever else they feel like) that has MORE features and more useful goodies than the modern computer.
Why are people switching from old videotape to DVD? It's the addons - no degredation of quality over time, extra interviews, a smaller disc, you can play a DVD on your computer, or whatever else it is that Joe Consumer happens to like. Laserdiscs didn't have as nice a "feature set" as DVDs, and we can see how few people actually have laserdisc players...
If the content industry wants people to use rights-management hardware, they have to make the hardware desirable (or pass a law banning everything else...). And if they don't make the hardware SIGNIFICANTLY better than the stuff we all have already - our boxes, our DVD players, etc., then the market is going to drop "rights management" like a hot potato.
And if the content industry actually DOES create a better product, and gets the market, I say more power to them! Then, and only then, are they actually working in a capitalist economy. But I don't see that kind of creativity on the part of the content industry, or whoever else wants "digital rights management."
Actually, I know someone who did a high-school science project on that. He video-caputred the launch scenes, and was able to measure the launch ramp that was in a few frames of the video. And according to his calculations, the bus would actually have made it across the gap. Strange, but true...
The GPL/LGPL basically says you can't change the license on the code to anything non-GPL/LGPL.
The MS license says you can't ever change the license to GPL/LGPL - or, in other words, MS must always have the option to copy/buy/(steal?) the code back.
Really, MS just took the GPL and turned it around on itself. If the ideas behind the GPL are valid, then the ideas behind this license are valid. Clever trick... you ALMOST have to admire their lawyers.
MS has faith that open source can't survive without MS code. Open source has faith that they can survive without MS code. I wonder who's right...
"Thousands of copies of movies and hundreds of thousands of songs and software programs are illegally traded on the Internet every day"
Your senator is right - I don't think anyone can dispute these numbers. Not even Slashdot, unfortunately. However,
Thousands of cars run red lights and hundreds of thousands of cars speed every day.
If the act in the Senate suggests embedding copyright management technology in all digital devices, then for the same reason we should also embed technology in every single car sold in this country that prevents the car from running red lights and speeding. A car manufacturer will tell you it is impossible to create that technology (how does a car know if it's near a red light? What about scooting around the corner to let the firetruck through, or avoid the semi behind you?). And Slashdot will tell you that it is equally impossible to protect copyrights in a reasonable and fair manner on digital devices - there are too many unknowns (some things don't have to be copyrighted, not everyone has access to the copyright protection standards or certifications, etc.).
This is a classic case of enforcing existing laws. We already have laws against speeding; we already have laws against unauthorized copying of copyrighted material. There is no need to mandate that technology change to enforce existing laws - instead, just enforce the existing laws!
Be careful with your analogy - in this case, the ISP is closer to the strip mall that "hosts" the offending store. So saying that the ISP is culpable is closer to saying that the mall management is culpable when the liquor store sells alcohol to a minor.
Why is this analogy better? The consumer is more like the visitor to the liquor store, and the WEB SITE (which your analogy left out) is the liquor store. The ISP is the management of the mall - which leases out the store space our liquor store is using. A mall has as its prevailing use the concentration of lots of stores and clearly doesn't have the means to enforce the laws within those stores (do you really think mall security can arrest you?). So to attack the infringing party, the two culpable parties are really the infringer and the web site at the beginning - not the ISP. The ISP is the mall that brings the two together - and the same mall that I go to for my groceries. Sue the ISP, and you might just sue my grocery store out of existence...
As I understand it, the open source movement basically says, "we can write better, more efficient, less buggy code than the proprietary guys." If, between maybe Wine (with Lindows' help? or not?) and SDL and this DirectX - OpenGL wrapper, someone can create a Linux system that runs games as fast as and as stable as Windows... then OSS will hijack Microsoft's standards.
If I'm a gamer and I can get a Linux level of stability without sacrificing the latest graphics goodies, I'm going to switch to Linux real quick. No more crashes...:)
Every Linux fanactic claims that Linux is faster and more stable than Windows. Here's a chance to prove it. Get the Linux implementation to be better than the Windows "reference" implementation, and Microsoft will lose control of the standards. Why use buggy MS DirectX 17 when you can use stable Linux DirectX 10.2, get practically the same features, better performance, and not have to worry about MS changing the standard half-way through your code?
The distributors pay full value for the merchandise and accept the risk that the software may be damaged or lost. The distributors also accept the risk that they will be unable to resell the product. The distributors then resell the product to other distributors in the secondary market. The secondary market and the ultimate consumer also pay full value for the produce, and accept the risk that the product may be lost or damaged. (page 13 of ruling, emphasis mine)
Does this imply that software sold under the license model has an implied guarantee by the distributor that the product will not be "lost or damaged"? Since it is a license for a service, then the original distributor is responsible for any problems in the distribution - that's why they slap the EULA on the software, to protect against malfunctions, etc. But this implies that the distributor is responsible for guaranteeing the media - if it breaks, they would HAVE to replace it, so long as I could prove I bought it. They would have to replace it for the duration of the licence - in many cases, forever. You scratch a Windows 95 CD? They gotta replace it - at cost, because it's a licensed service and not a sold product.
IANAL - but wouldn't this create some additional problems for distributors, if people actually make use if this?
Small devices are the key? Probably. But I have a caveat: too many gizmos, and too much trouble making them all work together.
Here's what I want:
A Palm to carry around: take notes, (finally) figure out my schedule, kill spare time, and anything else I feel like.
A server or server-space: to run a web site (with CGI scripts I control), to have an internet-accessible repository for all my files, a place where I can mirror important files (the CounterStrike install someone needs every week), and a machine that I can get to from anywhere to do anything I need - ping myself remotely, or something else, from a machine I control.
A flexible, high-horsepower machine: writing letters, writing papers, printing, and the essential games. And, since I know I'll be doing crazy things on this machine, THIS MACHINE CAN'T BE CRITICAL FOR ANY OF MY OTHER NEEDS!
Right now, something like this costs a boatload and takes just as much time to set up. A cheap linux box for the server, a Windows box for my stuff (remember - compatibility), and a Palm for everywhere else? I've just paid for THREE operating systems AND I have to worry about linking THREE separate locations.
Wonder why.NET is going to be successful? One operating system on each of these boxes. ANYONE who can create a standardized system - something where I don't have to worry about what OS is where, which chunks of my data are current on which machine, and do it all securely, with stability, and with a minimum of hassle, will get my $$$. Right now, Microsoft would cost WAY too much for too little security, and Linux isn't stable enough or unified enough to use across all these devices.
Here is the market potential. Integrating a palm device with 1) other palm-type devices (or wall devices, or whatever you feel like) AND 2) with higher-grade computing power (servers, workstations - and something that can run some sort of heavy-duty office). And making the integration seemless enough that adding some new device (the Super Palm XP 2020) to this mix DOESN'T crash things and DOESN'T make me go through five zillion configuration setups. Best done: on a single operating system..NET or Linux, or whoever fills the vaccuum.
If you read the Philips link from the Register, they mention this argument as Philip's defense.
... the protection system is not a protection system as such, but simply a mechanism for stopping the playback of music. This interesting claim allows him to contend that the protection systems are not covered by the Digital Millenium Copyright Act
Even Philips doesn't consider hardware "copyright protection" to relate at all to copyright - just COPYING.
Slashdot Mirror
http://ask.slashdot.org/comments.pl?sid=44192&cid= 4599885
I think the NVidia drivers get away with this because they are two parts: a kernel patch, and a separate copiled kernel module. The kernel patch IS a part of the kernel and so must be publicly released, blah blah blah, or you get toasted by the FSF - but the kernel patch is really just a bunch of hooks that let the binary driver interface with the kernel.
So, the patch is just a publically available interface to the kernel. And the the kernel module is a module (which, if above posts are correct, Linus has blessed as not being subject to the same license as the kernel) which uses only published interfaces (so there's no longer any argument that it's linked directly to the kernel).
The risk here is only that the patch gives away too many implementation details - but there also happens to be no requirements that you not obscure the workings of the interface (unfortunately).
Because what GNU programs does an embedded device need? I'm pretty sure you're not loading a compiler onto it, nor a debugger, nor just about anything else...
The only thing I can think of offhand that you would even use is glibc for userland programs. And I have a hunch most embedded devices don't use a whole lot of glibc - and could very easily be changed to another c library. The kernel doesn't use it... and for an embedded device, most of the interesting stuff will be in kernelspace anyway!
I'd guess that anyone who submits a patch back to Linus is de facto making him a copyright holder of the code in the patch (well, assuming he is also the holder of the original code ... but I'm pretty sure he has the copyright for the 0.01 kernel!). IANAL, but I think it's a safe bet that sending a patch (w/o an explicit license agreement) constitutes a gift.
All linux patches then work on the assumption that Linus is a nice guy and will push the patch out in a publicly-licensed release of Linux. And this is so far a completely safe assumption: Linus has not shown any commercial interest in the Linux source code, but has shown extreme interest in making Linux as good a kernel as possible.
Although, for comparison, look at SourceForge: when they went proprietary, they stripped out all changes to their original codebase and went back to a version they owned the copyright for. Very conservative approach, but you'd never find a hole in it...
It's (2) publish a user's guide and call it a "cheats" book.
The point is, for all the really good games I already figured out everything in the strategy guide (yes, I skimmed it at the store ... and didn't learn a thing). Civilization II, anyone? That game also happens to be the last one that came with any sort of legitimate guide-as-a-guide - something like Homeworld has a guide that basically says, "here's how to install, here's what a bunch of keys do (in doublespeak - the descriptions only make sense after having learned to use the key...), and here's some concept art."
Hence, the reason that I've bought more games that are more than a year old than new releases.
The original poster mentions that his algorithm isn't a simple in-place sort - which, based on my best guess, means that he's probably sorting pointers quickly then re-arranging the elements in a single pass. Maybe, maybe not - but still no faster than O(n log n).
But at this point, I'd start worrying more about scalability, memory usage, etc. - one of the nice things about quicksort is that it is so general, it can quickly be adapted to low-memory or multi-processor implementations. Sure the algorithm is faster - but faster does not necessarily mean better. If my implementation guess above is correct, the algorithm would use significant additional memory, making it unsuitable for sorting simple numbers (where the pointers would dwarf the elements themselves) or for use in low-memory (i.e. embedded) applications.
The true test of whether the algorithm is innovative is: 1) have a professor look over it, and then 2) get a (respectable) journal to publish it. If it passes the peer review required for publication, then it's an innovative algorithm. If not, it's one of thousands of alternative implementations of the same idea. Clever, maybe even useful, but not spectacular.
At the very least, I'd suggest an ADA-accessible front page that basically says, This is a visual arts web site, if you proceed further you probably won't get anything out of this site. And this can be as simple as a message in an ALT tag.
Your front page needs to be accessible to anyone - including people using Lynx. Why? If I'm connected to a remote server, there's no way I'm going to run a complicated Mozillia-over-X-over-SSH setup just to hit the internet while I'm looking for "deco wallpaper". I'm going to use Lynx! And even if all you do is tell me "you can't get anything out of my site if you can't see this image", you've still done more than 95% of the sites out there - and with 3 minutes of your time. I'd recognize your site doesn't have value for me - and I'd thank you for saving my time.
These technologies:
- Java - bloated, unwieldy, and EXTREMELY difficult to set up or upgrade. Rely on it, and you've alienated 99% of web users
- Javascript - probably THE easiest way to get viruses into a system. Slow, unwieldy, bug-prone (I get Javascript errors every time I visit Microsoft's site, even under IE), and really only useful for fancy, glitzy stuff whose only purpose would be to cover up an otherwise poorly designed site. Many security advisories simply suggest disabling Javascript - you shouldn't rely on it at all, or EXPECT to be disappointed.
- Flash - a useful technology. IF the plugin works. A very big if.
- Graphics - Relying on too many graphics makes your site slow to load. And glitches in displaying graphics are one of the biggest browser-compatibility issues around. I don't have your high-speed internet connection, I don't really want to download gratuitous graphics. If I want to see something I'll click on it - don't make me download megabites of junk.
- streaming media - Let me repeat: very few internet users have the bandwidth for decent streaming media. And if it's integral to your site's design, you have a poorly designed site.
Please, before you go out and bash someone for complaining about your site, take a second to think about what they want. Of course a blind person or text user doesn't want to poke around your fancy, glitzy visual-arts site. But he certainly doesn't want to become trapped in it, waiting for slow and useless graphics and extensions to load and unable to navigate without SOME text.A site doesn't need a complete make-over. It just shouldn't be blatently offensive to blind or text users. And if you can't spare the 5 minutes of your time it takes to make a few simple changes to your front page alone, I truly pity you - and hope I never come across your site.
You don't have to bend over backwards, but do not dismiss a legitimate complaint without hearing it out. It's rude, and frankly it's exactly the reason the ADA was passed in the first place.
And no, you can't change the page table size, it's hardware-dependent. Most of the other archs seem to have similar or larger pages, too.
Why do I know this? It's "write your own VM" month in my OS class. Next week we get to start swapping out to disk...
It's perfectly legal for me to download the pre-patch kernel - probably from some archive somewhere. And it's perfectly legal for me to download the patched version. But, by running diff and creating my own patch, I've just discovered how to bypass the security. So just running diff is a criminal act?
The only way I see out of this is even worse: it would be illegal to have an unpatched kernel! Since an unpatched kernel could be diffed against a patched kernel and break the DMCA, the unpathced version must be illegal. Then, to enforce this, OS vendors would be able to require (a la the new MS EULAs) that you allow them to arbitrarily update the kernel without telling you.
The scary thing is, this all makes sense. Can anyone see another way out? (besides declaring the DMCA illegal - that one's easy!)
Remember that just because a law is "bad" and horribly unbalanced towards lobbists doesn't make it illegal unless there is some specific legal reason the law is unconstitutional.
Exactly right - and the Supreme Court would be correct in maintaining a law passed entirely by lobbyists over the objections of 90% of the American people. Why? Because the Constitution doesn't prevent it. And in the U.S., there is no law higher than the Constitution. Immoral, yes, but illegal, no.
I even think this is necessary - because think of the opposite. How many laws were passed with the support of 90% of the people over the objections of lobbyists? Answer: a whole bunch of important ones. Taft-Hartley (sp?) on labor laws, Sherman Antitrust, Amendment 19 (?) (women voting). If we want the people to overrule lobbyists sometimes, we need to let lobbyists overrule us sometimes - because the people can always re-elect a more favorable Congress.
So far it seems the court is buying [the limited powers] argument. The court seems to be unsure though if it has any power to do anything about it. This is good news to Lessig, because it means the court buys the limited power argument.
It's actually a very good thing the court is unsure if it has power. In fact, the court almost ALWAYS asks if it has power to intervene... you'll notice that since the court rejects upwards of 90% of the cases appealed to them. A big part of those appeals are cases where the court doesn't feel it should issue a ruling.
Another example, the Florida election stuff. That debacle was an example of the Court deciding it DID have power (and deciding very quickly, and without much justification). IMO, a more correct ruling would have been to just say that "how a state runs its elections is its own business" and just refuse the matter entirely. This was basically the minority opinion. I'm not arguing the merits or outcome of the decision (I would have remanded it with instructions for a re-hearing, but I'm not a Supreme), but most of the mess occurred because the Court jumped too quickly into an area it shouldn't have entered.
Perhaps the Court even feels burned because of the Florida voting episode, and is trying to be extraordinarily cautious with its power (and even more so because more eyes are looking for faults!)
The case was also helped by a government bumble. The government argued that there is no constitutional limit on the ability of Congress to extend copyright, thus the extention was legal. This actually helped Eldred because the court did not like this view at all. The court did not support the idea that the constitution limits the powers of Congress, but that Congress gets to set what the limits are. In effect, the government proved Eldred's point themselves.
As much as the Court hates giving itself power, it likes giving someone else power even less. A better argument might have been that there is a limit but Congress should determine it; but instead, the government chose to try to defend the whole thing. It almost feels like they're giving the case away. Thus: I'm watching carefully to see if someone responds to "an erosion of copyright law" by introducing a new, draconian law with overwhelming support. Cyncial, yes, but I'm very cynical when politicians are involved.
Very good summary, my complements to you good sir.
The other problem is more serious: most programs go many, many function calls deep. If each function pushes it's own registers onto a register stack, the stack will be proportional to the execution depth. You've seen Java error messages - those are easily 10-20 calls deep. OSes can go much, much deeper ... and I'm not even getting into recursion, which can go thousands or more levels. Try to keep all these registers on a special register stack would mean you have a 5K or so stack register (and the architects are already fighting over a 20K L1 cache!), AND as mentioned above, this special stack register has to be flushed to main memory every context switch.
Although, I wouldn't be surprised if this idea were actually implemented on much older (i.e. pre-386) type processors that were pretty much single-threaded anyway. It's not a bad idea, it's just that it's too out-of-date for a modern OS.
But, like all truly useful Windows tools, you gotta dig to find this one.
Go to Control Panel/Administrative Tools/Computer Management. Choose the Disk Management item (left pane), which will bring up a diagram of all your partitions in the lower-right pane. Right-click on any available partition (except your C: partition!) and choose Change Drive Letter and Path. When you try to add a new location, it will allow you to specify a directory as a mount point - a la the *nix mount command.
Plus, you get a little drive icon in the middle of a list of folders - I'll bet very few people have one of those! This way only works on physical disks as far as I know - but it's a great way to make a drive bigger when you run out of space.
My prediction for the future: the next edition of Windows will encourage you to mount your /hom... er... C:/Documents and Settings/ directory from Hailstor... er... your Really Big And Expensive Windows Server!
By posting a link on /., and knowing that the result (particularly from a bunch of suitably angry geeks) is a slashdotting and is enough to crash servers, a lawyer would probably argue that the poster DELIBERATELY caused a DoS attack. And suddenly this e-mail becomes a prelude for a more serious - and documentable - attack.
Likely to happen? I hope not. But chilling to think that it might.
-Paranoid whenever Lawyers are involved
The difference with a photocopier is that it's actually legal - most (as in almost all) of the people who use a photocopier fall under fair use. It's legal to copy a few pages (maybe even most of a chapter) out of any book at your local library - heck, that's why they have photocopiers there in the first place! The reason it's legal is that you aren't copying the entire book - you're keeping a part of it for reference (that one recipe you've been looking for the past three years and know you'll forget if you don't copy it), or for criticism (you're a student who needs to quote a few passages out of War and Peace to demonstrate how long-winded it is), or for information (you copied the map of Texas out of the encyclopedia because you're about to vacation there). Of course you can't legally copy the entire book, but everyone knows you won't because you'd pay more in photocopier fees than you would for the book itself.
Fair use even has a special carve-out for academic uses - that's why a teacher can photocopy a "classroom set" of articles out of Scientific American to assign as reading, and hand them out to students. (There are catches - the teacher is supposed to destroy all copies after they are done, and they ususally don't try... but what student saves their classroom readings after the test anyway?). That's also why you can block-quote someone else's writing - it's legal as long as you give proper credit.
How sure am I of this? IANAL, but I've had to read the US Copyright Office's publications for a research project, and all this is laid out pretty clearly.
I think I've strayed from the point... which is that photocopying is a completely legal "fair use". You just can't make the same case for copying software... it's not like you're making a limited use of a small part of it, when you copy software you copy the whole thing!
Really. The campus probably has an ATM machine, it's easy to get $5s and $20s out of a bank account. But the change machines always seem to be out of money or out of service. I don't want to spend money on something to break even a $5 into bills small enough to go into the pop machine...
Yeah, small bills and quarters. Better than lots of money any day of the week.
That technological lead will last for all of 3 seconds until someone puts out an identical product without DRM.
That's exactly the point - with our highly touted free market system, everyone is free to make a better product. "Better" in one way might be "has more features, but isn't as nice for geeks to use" (i.e. DRM features), but to another consumer "better" might mean "same as the other guys, except w/o the DRM junk". In a free market, both of these products are perfectly valid.
As for someone putting out a non-DRM product 3 seconds later... How about the difference between a game console and a PC? I would argue that neither one is really more technologically advanced than the other (beyond maybe a few-month gap for a product life cycle). But because consoles are a "better" product (for playing a game on a large-screen TV with an interactive controller, and for not crashing as often, and for being cheaper than a PC), they have a strong market. PC games have a market because they are "better" than games at 1) saving, 2) multiplayer, and 3) better controls for complex games (just a few ideas, I'm sure there are more). Just copying a product and stripping DRM out of it doesn't necessarily mean the stripped product can compete with the original.
What I'm suggesting is that the DRM-favoring industries will only be successful if they can produce a product that is innovative enough to be BEYOND what someone else can copy "in 3 seconds". Some industries can produce a more innovative product - you go to a restaurant for well-prepared food and atmosphere, when it's cheaper and you have more freedom to choose when you go to the supermarket.
But, being a Slashdot user, I am obliged to present these two observations (because I'd get flamed for anything else?):
In other words, I defend the right of companies to produce products with DRM. I just don't think they are competent to be successful with it - I think that, fairly soon, someone else is going to come along, with a better vision of how to run a business in a content-type world, and will drive these DRM-favoring companies straight out of business. Some companies are already successful - look at Red Hat, they are "innovating" the Linux OS with something of value (polish, and support), and therefore have a product to sell. I think the only reason there aren't more Red Hat-type companies in this world is that there hasn't been enough time to work through the trial-and-error process of coming up with a way to do business.
Why are people switching from old videotape to DVD? It's the addons - no degredation of quality over time, extra interviews, a smaller disc, you can play a DVD on your computer, or whatever else it is that Joe Consumer happens to like. Laserdiscs didn't have as nice a "feature set" as DVDs, and we can see how few people actually have laserdisc players...
If the content industry wants people to use rights-management hardware, they have to make the hardware desirable (or pass a law banning everything else...). And if they don't make the hardware SIGNIFICANTLY better than the stuff we all have already - our boxes, our DVD players, etc., then the market is going to drop "rights management" like a hot potato.
And if the content industry actually DOES create a better product, and gets the market, I say more power to them! Then, and only then, are they actually working in a capitalist economy. But I don't see that kind of creativity on the part of the content industry, or whoever else wants "digital rights management."
Actually, I know someone who did a high-school science project on that. He video-caputred the launch scenes, and was able to measure the launch ramp that was in a few frames of the video. And according to his calculations, the bus would actually have made it across the gap. Strange, but true...
The GPL/LGPL basically says you can't change the license on the code to anything non-GPL/LGPL.
The MS license says you can't ever change the license to GPL/LGPL - or, in other words, MS must always have the option to copy/buy/(steal?) the code back.
Really, MS just took the GPL and turned it around on itself. If the ideas behind the GPL are valid, then the ideas behind this license are valid. Clever trick... you ALMOST have to admire their lawyers.
MS has faith that open source can't survive without MS code. Open source has faith that they can survive without MS code. I wonder who's right...
Your senator is right - I don't think anyone can dispute these numbers. Not even Slashdot, unfortunately. However,
Thousands of cars run red lights and hundreds of thousands of cars speed every day.
If the act in the Senate suggests embedding copyright management technology in all digital devices, then for the same reason we should also embed technology in every single car sold in this country that prevents the car from running red lights and speeding. A car manufacturer will tell you it is impossible to create that technology (how does a car know if it's near a red light? What about scooting around the corner to let the firetruck through, or avoid the semi behind you?). And Slashdot will tell you that it is equally impossible to protect copyrights in a reasonable and fair manner on digital devices - there are too many unknowns (some things don't have to be copyrighted, not everyone has access to the copyright protection standards or certifications, etc.).
This is a classic case of enforcing existing laws. We already have laws against speeding; we already have laws against unauthorized copying of copyrighted material. There is no need to mandate that technology change to enforce existing laws - instead, just enforce the existing laws!
Be careful with your analogy - in this case, the ISP is closer to the strip mall that "hosts" the offending store. So saying that the ISP is culpable is closer to saying that the mall management is culpable when the liquor store sells alcohol to a minor.
Why is this analogy better? The consumer is more like the visitor to the liquor store, and the WEB SITE (which your analogy left out) is the liquor store. The ISP is the management of the mall - which leases out the store space our liquor store is using. A mall has as its prevailing use the concentration of lots of stores and clearly doesn't have the means to enforce the laws within those stores (do you really think mall security can arrest you?). So to attack the infringing party, the two culpable parties are really the infringer and the web site at the beginning - not the ISP. The ISP is the mall that brings the two together - and the same mall that I go to for my groceries. Sue the ISP, and you might just sue my grocery store out of existence...
If I'm a gamer and I can get a Linux level of stability without sacrificing the latest graphics goodies, I'm going to switch to Linux real quick. No more crashes... :)
Every Linux fanactic claims that Linux is faster and more stable than Windows. Here's a chance to prove it. Get the Linux implementation to be better than the Windows "reference" implementation, and Microsoft will lose control of the standards. Why use buggy MS DirectX 17 when you can use stable Linux DirectX 10.2, get practically the same features, better performance, and not have to worry about MS changing the standard half-way through your code?
Does this imply that software sold under the license model has an implied guarantee by the distributor that the product will not be "lost or damaged"? Since it is a license for a service, then the original distributor is responsible for any problems in the distribution - that's why they slap the EULA on the software, to protect against malfunctions, etc. But this implies that the distributor is responsible for guaranteeing the media - if it breaks, they would HAVE to replace it, so long as I could prove I bought it. They would have to replace it for the duration of the licence - in many cases, forever. You scratch a Windows 95 CD? They gotta replace it - at cost, because it's a licensed service and not a sold product.
IANAL - but wouldn't this create some additional problems for distributors, if people actually make use if this?
Here's what I want:
Right now, something like this costs a boatload and takes just as much time to set up. A cheap linux box for the server, a Windows box for my stuff (remember - compatibility), and a Palm for everywhere else? I've just paid for THREE operating systems AND I have to worry about linking THREE separate locations.
Wonder why .NET is going to be successful? One operating system on each of these boxes. ANYONE who can create a standardized system - something where I don't have to worry about what OS is where, which chunks of my data are current on which machine, and do it all securely, with stability, and with a minimum of hassle, will get my $$$. Right now, Microsoft would cost WAY too much for too little security, and Linux isn't stable enough or unified enough to use across all these devices.
Here is the market potential. Integrating a palm device with 1) other palm-type devices (or wall devices, or whatever you feel like) AND 2) with higher-grade computing power (servers, workstations - and something that can run some sort of heavy-duty office). And making the integration seemless enough that adding some new device (the Super Palm XP 2020) to this mix DOESN'T crash things and DOESN'T make me go through five zillion configuration setups. Best done: on a single operating system. .NET or Linux, or whoever fills the vaccuum.
Even Philips doesn't consider hardware "copyright protection" to relate at all to copyright - just COPYING.