Sorry, I set my comment size to an very large value, I don't have to worry about the size, copyright infringment isn't that simple. I've got a chronic issue with spelling your/you're, there/their/they're correct. I didn't bother to re-read it, I'm assuming I've just done it again. I try and double check everything before I post to ensure that I get that correct. If slashdot had an edit my post feature I'd fix it for you.
Just because they can't win doesn't make it copyright infringement right. That isn't a justification for being a copyright infringer. Notice I avoid using the term theft, it's not.
I think the RIAA should change their business model. It's why I pointed out that the RIAA doesn't have the right to a business model. It's why I hope that young musicians figure this out. The funny part is most of the musicians I know, actively avoid downloading anything that they don't have a right to (ie: isn't directly from the band). One guy when he realized just how bad it would be if he ever made it, and it was done to his music, that tracked down every band he could and purchased the CD's, precisely because he wanted to be on the up and up. He deleted everything else he couldn't didn't personally own.
Not precisely. The economic value of the song is diminished. That is what was taken away. Even if you wouldn't have purchased it, the economic value is diminised, precisely because a copy was made. So supply is increased, holding demand constant, drops the value. Simple economics. Your simple illegal copy diminishing their value, makes you liable for a lawsuite. Simple legal analysis. Now if your are a clever person, you'll respond that CD's are price fixed. Which is probably true. As a consumer do you know how to make the price move? Stop buying music. That doesnt' involve becoming a copyright infringer to get a copy anyways.
The reason this hasn't been a problem in the past, was that being a copyright infringer wasn't free. Now it is. Doing it to scale, and not getting caught was difficult. The internet and technological advancements have made that possible, and why it's a problem now.
You might not have bought the song either way, but your getting the enjoyment of said song. That has some value to you. If it didn't, you wouldn't have downloaded the song. Or at the very least you wouldn't have put it up on display for others to download. So clearly the song has value to the people whom are putting it up for download, and it has value to some of the people downloading it.
Those songs aren't naturally occuring objects. They don't just grow on trees.
Do you understand the Lockean princepals that are the rational and foundation of our current system of gov't? Know why people can claim they own land? Go read John Locke's "Second Treaties on Government". Very good book, you'll learn a lot about ethics and princepals in it. A lot about the justification for gov't, ownership, and property rights.
They can claim ownership by working the land, and improving it. You can claim ownership because you've put forth the effort to change and create the land from it's natural state. That's how one claims ownership of such things. I think that the individual artist, and the corporations behind them have put in the work to create the objects, and thus have a moral right to their ownership of the music.
Personally, I think that P2P networks should be left alone. They are fine constructions that have legitimate purposes. I think that if the RIAA is going to go after individuals who are copyright infringers, that's great. I think the people they went after recently who created search methods is wrong. I think those people should have been left alone. I think the RIAA should just crucify several copyright infringers in court, and keep doing it, until people realize the risks. It should continue doing so, one after another. It's illegal, it's copyright infringement, it's against the law. That's all there is to it.
Copyright is a *WONDERFUL* thing. It's what makes the GPL tick. It's what makes being a writer, and a programmer a viable proposition. It's what makes so very many occupations work.
Copyright intentionally makes scarcity of non-scarce objects, for the specific purpose of creating economic value. Did you catch that? It was an intentional construction, put forth by the founding fathers, who clearly thought about the matter at length. I think fair use is a good thing. I think making backup copies is a good thing. I think copyright is a good thing (not as currently implemented in the US or internationally by the Berne Convention).
Now, I think that copyright is a screwed up deal. I think they are entirely too long, and that they are fundamentally broken in that respect. However, you should respect copyrights. If you don't, there are innumerable things you enjoy which will disappear, precisely because they are what creates the economic incentive to create. That's why the founding fathers created copyright.
If you've got a problem with the Music business, stop consuming their product, support a different product. The music business doesn't have a right to a business model, so they can't just randomly sue people
It's a real patent, given out by the real patent office. What part of the definition of a patent doesn't the one click patent have?
Clearly it's novel, if absolutely no one did it, while such technology existed for a long time, and no one did it. It most definitely has utility, it is useful.
Now go all the way to the bottem of that, and we might see what the problem is...Looks like the patent office, has a problem with web-based technology. If you can't find documentation where you take an idea from documents A, B, and C, and get the idea in the patent. What if the problem is, it's the god damn web, and it's all oral tradition?
You might have a case for non-obvious, but clearly a judge didn't agree with Barnes and Noble on the point. Now it might be obvious to most people now, but was it obvious in 1996-7 when they applied for that patent. Not sure, I didn't know enough about the web at the time to have been considered an expert in the area. At the time, I didn't know what a cookie was.
I think the Internet is too much of a land grab at the moment, and that too many simple basic ideas right now are coming out, which make things tricker on the internet.
Again, I'm still standing up being a devil's Advocate. I really, really wish, that it was just over.
I think that concept of saying: "Put it on my tab", existed in the real world for a long time, and that this is merely an application of such an idea to the internet. I think that the referral business relationship where by Amazon pays you a piece of the pie if the referral comes from your website, and the associated technology is stupid. I think that saying, well I'm going to automate my process using a computer, shouldn't be patentable, however, doing that falls well within the definition of what is patentable.
I'm trying to say what Lessig says so much better: Read it here
Did Amazon patent "A single mouse click to check out", or did they patent storing information on the customers computer, including all of the items they wish to check out, and their CC information, so that in a single mouse click all the needed information is transmitted so they can be checked out at once. So if you stored it server side, you'd be in the clear.
Remember a patent is a supposed to be specific invention. If you can have a check out in a different way then they patented, you should win the patent case. Hell, patent it, so you can tell Amazon, that clearly the patent office thinks it's unique.
People did actually expend an immense amount of effort to find prior art. I think you'll actually have to give that one to Amazon. At least they didn't patent selling shit over the web. So my guess is that, they really did concieve of a single click patent. I believe that most trivial coding idioms will have prior art. They aren't idioms until they are in common usage.
So, what would it take to implement a single click check out not using their patented method? I'm not really sure, I'd have to go read the patent closely. If somebody wants to piss them off, just extend all of their patents trivially, so if they ever try and do anything more then what they've patented they have to come see you about it. It'd be lots of fun.
FYI: I'm playing devils advocate here mostly. I think the Single click patent is stupid. I think that they patented something obvious, not something novel. However, patents of this nature have existed for a long time in other spheres, it's just that people are starting to see them in computers.
You are correct, typo/thinko on my part. The first use of copyright in my original post was meant to be patent. As far as an example of a copyrighted version of all of those algorithms, any one written in the US is copyrighted. I'll bet that the Mozilla Code has a copyright on a GIF decompression algorithm. I'll bet the guys who wrote the blade encoder/decoder is copyrighted. I forget the guys name, but his initials are PK who wrote PKZIP has a copyright on a PKZIP compression/decompression algorithm.
They don't have a copyright on the algorithm, They have copyright on the specific version they implemented. I'm guessing that patents (the actual patent document) are considered part of the public domain, so the original description of the algorithm is probably not copyrighted at all.
Sorry.
I know copyright is completely different then patenting, both of which are different then trademarks.
I'm an idiot for typing the wrong word on slashdot... I'm not unique or special....
People copyright algorithms all the time. MP3, the original PKZIP, GIF, all patent encumbered algorithms. Sorenson Code, patent encumbered. Now it's probably impossible to patent:
for( i = 0 ; i < sizeof( array ) / sizeof( array[0] ) ; ++i ) initialize(array + 1);
Or a trivial coding idiom. I'll bet there are lots of places I've duplicated simple code from various other places, just because there are only so many ways to implement a merge sort. There are only so many ways to implement a quick sort. There are only a handful of ways to implement binary trees. It's not shocking that mine, and somebody else's I've never met, who have never seen each others code, have nearly identical code.
I won't be shocked to find out that, this whole deal is that SCO copied a BSD driver, and then Linux did, and now they have absolutely identical code in them.... Uhhh, but BSD owns the copyright SCO, how are you suing for copyright infringement. Oh well...
Christ, if they had conduit, I could pull 200 cable runs personally in less then two weeks of full time work. The cable and ends aren't that expensive. Quote out $5,000 in parts, $1,000 in new wiring tools, and $5K for my shiney new Cat5 spectrometer (the tool that measures the attenuation on the copper, I might have the wrong name) so you can be absolutely sure somebody made a killing on two weeks work. The $5K on parts is probably too high. Buying them Female Lucent Cat5 ends 1 at a time, is about $6 * 200 * 2. Assuming they made all of the runs the maximum length ((200 * 300)/1000 * 60), that's roughly: $6,000, that's when I'm not buying in bulk, and negociating a discount. That's just me off the street buying parts as an individual.
Okay, now I might have to get a several hubs. Lets assume I got HP Pro-Curve 80 port capable switch and filled them all the way to the gills, I'd need 3 for 200 drops, and they'd cost no more then $8K fully loaded if I remember correctly. Again that's MSRP, not a price I'd use if I did this for a living. I'd need probably another $3K in hardware (screws, face plates, Communication racks, and what not). Total cost for install, for a completely network that is ready to rock and roll:
$23K of which, $6K are one time costs to get into the business. So for a stock install, the total cost is $17K + Plus 2man weeks. Assuming they couldn't negocaite a decent price. It'd take me 2 weeks to do all the work personally, and I'm not very speedy at it. I'd bid it at no more then $30K-$35K if I was doing it personally. In two weeks, that's roughly $13-$18K in my pocket. Not great for a contractor, but nothing to cry about either. What'd they do that I didn't price out, where'd I miss the price too badly. I intentionally over-estimated everything on price. I could do the job for parts in a lot less then $6K is my guess. Nobody runs that many 300 foot runs. If I'd gotten cheaper female jacks in the cabling closet, it's have been better. Oh, and my prices are about 3-4 years old. I believe everything is cheaper now.
Heck, I'll bet I could run fiber to 200 rooms, and put copper to fiber transceivers in on both ends, or fiber on one end, and used fiber capable hubs they quoted, and still make a decent profit. The transceivers might be a little price, last time I read about it, I thought they ran, $45 a piece, but still 45 * 200 = $9K. So $80K is pretty expensive, assuming there really was conduit installed.
Hmmm.... Several things. First off, I'd really like to see the documentation that shows FTP sends the network address because it saved code. I've always assumed it was because in the RFC, there are control channels and data channels, and they are differet. You can use FTP on machine A to transfer files from machine B to machine C directly. So if you have FTP access to B and C you can directly move files, without having to log into B or C. That's why you send the IP and port over the wire according to the RFC. That might be a feature left over from what you're talking about, but I'll bet that FTP is so old, nobody knew what best practices would be in designing a network protocol. They designed FTP to be everything to everybody....
Second, quality of service is trival to handle now out of band from TCP/IP. QoS is also a rats nest of stuff that will drive up costs. Differentiation of traffic is a neato way to have extra costs foisted on you by the upstream ISP. Multi-homing can be done today by anybody who knows about TCP/IP. Now getting your own AS number and having the BGP protocol broadcast that two ISP's can get to your network is a problem.
IPv6 might have a magic solution for this, I haven't read up on it enough yet. As far as I know they just added 2 octects to the routeable IP space (while they added another 6 octects to the local space). So essentially, you only have to route to the first 6 bytes of the IP, after that, it's the local link/router's job to actually deliver the packets from there.
No the GPL doesn't allow you to distribute a binary program which is linked against something you can't provide the source for. You can as far as I know, provide source for a program for which you must have a proprietary software to link properly. As long as you compile the software yourself, your in the clear. I believe Sun did this with their toolkit to port Linux drivers to x86 Solaris. That's why they could link GPL'ed Linux source into the completely proprietary Solaris code, because the user did the compiling and linking. If Sun had distributed binaries, they'd have been in violation of the GPL.
As long as you do the compiling yourself, I don't believe there is anything you can't do with GPL'ed source. The GPL only comes into play when you start to distributed the source to other seperate legal entities.
Yeah, the only problem is go out to any hacker site, and look for details on how to break out of a chroot jail. Hell just put in "break out chroot" as search terms into google, and 5000+ pages come up. While chroot's nice, it's not the almighty security tool most people believe it is. (I'm not sure that having a VM that runs your software would be). UML might be better, or possibly another VM layer. The problem is, you need your VM to be completely audited for all interactions to make sure nobody can break thru to a layer above it, as soon as they can, game over. I'd rather seem UML get more support then chroot. I think chroot is of very limited utility, where UML seems to have a lot more applications, both interms of security (honeypots and as a jail for specific services) and virtual servers/services. However, chroot is more production ready then UML is, so there you have it.
Curious, this reminds me of the story about Cray computers. Seymour Cray put in a very, very fast circuit to do additions I believe (specifically to add 1). The circuit also gave the wrong answer if the input was one specific value, he could have fixed it, but it would have been a longer delay, and well being right in all but one case was acceptable to him. Well eventually people reported this as a bug, but he claimed it was a feature. It was such a well known bug, that everyone coded around it. They put the check in, and put the special case code in to handle it. Turns out this took much, much longer to do then if Cray had just put in a correct circut.
I suppose if it's documented to only work in certain cases, that's acceptable, however, the the code that calls it without checking for the input is then broken, and buggy. It should be fixed. If it can't be checked before calling the functionality, then the functionality better work for all inputs. That's good software. Stuff that just assumes that unsafe input will never, ever be put in, is a bug. A security hole. It's not reusable code. Reusable code, checks inputs. Reusable code fails gracefully. Reusable code, returns error codes indicating invalid inputs. Reusable code doesn't have security flaws in it.
Distributing code that won't handle all input cases for use in a public distributed computing project for the sake of speed is irresponsible, and stupid. Now, I'm a lot more likely to just never run one of the distributed projects then to risk security flaws if they are willing to sacrifice security for their speed. Security should be the winning factor in all concerns when writting software. When trading security for speed, is an option don't take it. Security or ease of use, take security. Security or correctness, re-write the software using a new protocol, or new algorithm, but still take security and document the correctness flaw. Right now I only run them on machines that don't have any valuable information on them, but I'd prefer they not be used in a DDos, so it'll probably get stripped off all my machines.
No, it's not useless. The structure of their driver, is faster, and more effecient then ATI's. It's widely held that ATI hardware is superior to NVidia hardware. Yet the NVidia still posts faster frame rates in a number of games, specifically because the driver is faster/more effecient, just better. Even when ATI was "cheating" (giving up quality to get frame rate, for the specific purpose of posting a better FPS, and not disclosing the fact publically) on Quake, NVidia was still eating their lunch. Even when ATI had better hardware, that could do more, faster, they still got beat by NVidia.
Now, NVidia might do lots of nasty things, but they still write the best damn 3D video driver on the planet (for consumer level hardware under Windows). Now it might only work because they structure their hardware so it works more effeiciently with software, so they can lower costs, or because they have some Algorithm's expert doing stuff nobody else has every thought of, or because they figured out how to get a few key loops to fit in less then 3 lines of P4/AMD L1-Cache. NVidia's driver is in fact a competitive edge. Giving up that structure to the public domain is bad for NVidia stock holders. It's bad for NVidia, but good for the public. Having all drivers of that kind of quality available to study for new drivers would be good for the public, it would improve quality, and speed of video hardware, and lower the cost. It'd be great for the public.
It'd be great if Coke gave up the formula for Coca-Cola too, but I don't expect them to do it, and I'd surely hope they don't for the sake of their stock holders. I wish NVidia would give up the secrets. I wish they did open source the driver. I wish for a lot of things, that aren't going to happen.
I don't think NVidia would fool around with a legal document. Some one could actually use it in a court of law, as a legally published document. That'd be relatively stupid, the legal department will have somebody's head on a platter for doing it. About the only thing I can see is that the documented last change is Mar 27th. I suppose if the license could be for a revision that doesn't exist. So I'm guess that it is real, if it's an April 1 day joke, it's a very, very good one. I don't have the time to try installing the stupid thing to see if it blows up badly, like the Engineers had a field day building a broken driver to make me feel the fool. However, the dates are all older (and they display very funny, look at the 103 as the year, and the negative hours).
Not really. I've double majored in Mathematics and Computer Science, and have 15 hours of upper level physics, and I'm a code monkey, I've got over 95 hours in Math, Physics, Engineering, and Computer Science related courses. I didn't pass the state certification process to be given the title of Engineer in the state I work in (Nebraska in this case). If I'd done that, I would have earned the right to be called an Engineer. Oh, and my business cards refer to me an Engineer, technically that is my title at work, but I've got right to the title. The guy in charge made it up, and I can't refuse it.
There can't be an Engineer in software. As Alan Cox has written on several occasions, software engineering is roughly in the alchemay stages in relation to state of the art chemical engineering.
In a real engineering, process there are things to check, verification, and known facts which can be double checked against the standing design. The design checked against the implementation.
Building a bridge is a much easier process then say writting a secure OS UNIX clone. It's easier in the sense, that it's been done so many repeatable times since the Romans. The understanding of the underlying structures of a bridge are well know. All intereactions between the people on the bridge, and the bridge are known (hard to calculate them all, but reasonable well known). Recently (in the last 50 years), we learned a new trick about bridges. That the cross sectional area, and wind can interact with forces that will tear down a bridge. Okay, that as far as I know was the last major mistake that was a complete unknown in bridge building. I believe it happened sometime in the 50s. You know right after the first couple of computers we're built. We don't know anything writting software in comparison to bridge building.
Building a Secure UNIX or UNIX clone has yet to be done once. The understanding of all the various layers of software spans when writting an GUI application all the way down to the quantum effects that happen in silicon, I believe they claim 12-15 orders of magnitude of understanding. That's a lot, possibly more then any other intellectual endeavor ever undertaken.
About the only places that can stand up and say, they follow an Engineering process are places that are SEI certified Level 5. They have a repeatable, measurable process by which they do things. That starts to sound like Engineering. Real Engineering is very hard, very tedious, and very boring. Most code jockeys I know, couldn't do it. It'd drive them nuts. You've got absolutely no right to be called an Engineer on an off the shelf software application. If you write software for a company that foists on you a horrible, gut renching process, of checks, double checks, that involves throwing away everything that doesn't meet their excating standards, now you talking about it. Where every single index is documented to not be possible to overflow the array, that's Engineering. I've never seen anybody ever do it, but if it we're Engineering, it would have to be done. Even after all that, you still have to be working in an area where what your doing is extremely well understood, and has been done lots of times before. Where all interactions between every module is well researched, and well understood. Where all the compenents have well defined qualities, that react in a statistically predictable manner in all situations.
Have you considered what solar flares would do to your software? Do you have a fault tolerant memory storage scheme to account for the bit flips that will occur because of it? Do you use an operating system, and do you understand all of interactions between the various parts? Have you documented why everything works the way it does? Do you document every single change with a full risk analysis of why it's being done, and how it only improves the reliability of the system. People in Engineering do that. People who worry about stuff like that, are Engineers. It's not that Engineer's do
This is either the second or third time it's been posted, as a monthly statistic. It's just odd. I find a lot of the BSD stories interesting. When OpenBSD posts a new 3.X version that had the GCC patches for helping to detect and stop stack overflow attacks, or when the big packet filter fiasco happened, those are interesting stories that involved something there was to discuss. When NetBSD finally gets ported to the handheld abacus that Sir Issac Newton used (I made that up). When NetBSD has a new stable release. This is basically saying NetBSD is being maintained... I don't find that shocking, or news worthy in any way shape of form. It was good a story first time or two. If they posted about the package of the month (especially, if 3 of them in 3 months, instead of two).
Look at that posted by timothy, submitted by Dan with a link to bsdforums.
Like I said, I didn't mind the first or second time it was posted, but it seems silly to post them on a monthly basis. I suppose I can start submitting the number of LOC a base Linux kernel has.
There are lots of interesting stories about BSD, this isn't one of them.
I'd rather see them post interesting things about BSD, I'm sure they are out there, this isn't one of them. Statistical updates about NetBSD that consist of a single number seem dull to me. If they had a list of interesting packages, rather then a list of 64 packages of which there are 3 or 4 sets of 6 inter-related packages, so the number isn't particularly meaningful anyways...
Hmmm, if I was really interested in the right NetBSD announcements I'd subscribe to the appropriate list. I don't think this is very, nerdy, and it surely isn't anything that matters.
While I don't mind the occasional, hey look at this new cool software package, this seems completely out of place on Slashdot. Monthly postings of the number of packages just seems incredibly bizarre. It's not like you are posting a new revision, a new feature, a new major release, major changes. In short, this isn't news worthy, probably not even to people who use NetBSD on a daily basis. If there we're some compelling packages it would be better. It's not like they just finished porting KDE or Gnome, or MS Office to it. That would qualify as slashdot worthy.
I thought the first time it was done last month, it was weird but okay for slashdot, however, I can't see the use in posting it every month. What is there to discuss? It is merely a statistic, there isn't anything compelling in the story to discuss, hence my meta-discussion of if it is worthy of discussion.
Okay, so what precisely are the constraints again? If it has to have a Windows Desktop, you will have to pay Microsoft for a bunch of licenses to be legal. Period. You can skimp on hardware, but that is it.
Now, if you are trying to avoid paying for the ICA client, but you are willing to pony up the money to Microsoft, then there is a client named rdesktop that does the Microsoft remote desktop protocol (RDP). It was reverse Engineered from scratch, and supposedly is reasonable stable. So now, you can run this on Linux desktops, but you still have to pay Microsoft a bunch of money for the apps (just because they are all running on one server, doesn't get you out of paying them for as many concurrent users as there could be).
Now, if you have to have Microsoft Applications, but not a Microsoft desktop, you might want to see the guys who develop the Crossover stuff. Now you can run a lot of Windows Apps on a Linux box that has a Wine processes running remote. The product is called Crossover Office Server Edition
I don't follow the legality of this, so get a real good lawyer before you try it out. looks like CodeWeavers is saying, you get to pay Microsoft a bunch of money.
This is probably roughly the same quality, but now your talking about using X for your network transport. Which is a little awkward for remote users, as they will have to run an X server. Cygwin ships with one for Windows desktops.
Now, if all you want is a bunch of desktops you can run remote from a linux server. Get a bunch of machines that can act like X-Terminals. A bunch of old cheap PC's with a good NIC will do the job, as long as the NIC will do PXE, or netbooting of some flavor. Go get PXES from sourceforge and run it. It will net boot, and run rdesktop, a Citrix ICA client, or run as an X Terminal for you. It is very good, and runs pretty well. This is what the city of Largo, FL does. They claim it's great, grand, glorious and best of all, dirt cheap.
I don't understand your requirements. They appear to be directly contradictory. We have to have cool stuff from Microsoft, but we can't afford to pay for it. My guess is the cost of the Citrix Clients isn't nearly as bad as the cost of all of the copies of MS apps you sound like you want to run. Anyways, these are some pretty decent ways to get remote desktops. However, with Microsoft, you don't really get a break on the pricing that way, it does simplify administration of the desktop, and makes replacing broken hardware much easier.
Hmmm, I've never seen the dash in a "dead-on". Dictionary.com lists it as a legal word, but technically it is listed as informal usage.
However your usage of the word "your" in that sentence is completely wrong. The word "your" implies ownership. In fact, it is nothing more then the possessive form of the word "you". Does he own anything in that sentence? He doesn't own Ali G, the dengerate faction, or the verb representing. So clearly the possesive form is absolutely wrong in that sentence. Did you mean the contraction you're? No because that wouldn't make sense in that sentence either, and it would be a gross error on the order of using "a" instead of "an" after editing a sentence. Did you just construct a piss poor English sentence after being a grammer Nazi? Yes, it would appear you did. The correct wording would have been:
"I was dead-on about you representing the..."
If you'd like, I can explain in excruciating detail about why that usage is better then the sentence you worte. I'd be a prick for doing so, but I will happily donate my time to an English lesson for you. Perhaps you would post under a named account, so I can go nit pick every post in your history.
I reword sentences in my posts often, and periodically that means I have mistakenly not corrected the corresponding articles, tenses of the words I should use. If English wasn't so sensitive to such things I'd be much happier, to me that are irrelavent for the most part. I'm attempting to convey information, not get it published in a book, or try and prove my superiour skill in the written form of formal English. This is informal language used on a public website. The situation doesn't dictate to me, that I should take the time to proof read it in detail, when my concept is clear. If it had been a clear cut mistake that was repeated several times in the post you would have a point. However, I've seen published books with more glaring mistakes that that. Those people were actually trying to be grammatically correct in their usage of the English language. I however wasn't.
Feel free to correct all the subtle mistakes I have made in this post. I'm not very careful about checking my posts for grammatical errors, so I won't be surprised to find out I made several.
Sure, my guess is that the Enterprise line (AS, ES, WS) are seen as the supportable versions that they sell to major customers so they can get a stable platform for several years.
The regular ISO's they will see as something nice to do for the community, and have it be the test bed for new features, and the "beta/gamma" release of the upcoming Enterprise series. Then the new enterprise releases will have lots of software that has been tested and released on all kinds of hardware and they will have a very good chance of making a very, very stable release for the enterprise lineup.
Because they have the stable release, I believe you'll see fewer, and fewer X.1 and X.2 releases, and you'll see a lot more.0 releases. Somewhat because it will be enticement to have people buy the Enterprise line, and somewhat to keep the "beta/gamma" testing on the bleeding edge. It's a pretty clever scheme all in all. If they can pull it off, and keep the bugs in the.0 releases down, and put out a.1 to solve big problems in.0, I'll happily use the standard ISO on my desk, and happily pay the money for the Enterprise lineup for my servers.
My guess is that they are following the same rule they always have. If it has a upgrade to a major library (glibc), major kernel internal changes, or a new compiler which isn't backwards compat with the older RedHat version, they bump the major number instead of the minor. It's normally some sort of major binary compatibility upgrade.
The next edition of RedHat I believe is supposed to include the new kernel threads stuff, with the glibc that supports it (hence re-implementing pthreads), it has a new compiler, and the new glibc. So probably the applications aren't binary compat with 8.0, so this is now 9.0. The price you pay for upgrading. It's not like the upgrade path doesn't work, and it's not like upgrading past these things will be vastly superior on Gentoo.
They are pushing out new big things, if you want to stay current, then upgrade to it. What's the big deal about the major version number? I really don't see why your panties are in a bunch with RedHat. Gentoo will do most of the same crapola to your machine that Redhat does when you upgrade, it just won't have a major version number change. Big whoop.
While I agree with you in princepal. Filtering spam after it's sent, means that your ISP's are processing more e-mail, need more bandwidth, and need more disk space to spool e-mail. That cost will be passed onto you the consumer. So paying the 3 cents means that the person generating the cost is paying, rather then having it amortized over the internet community.
I knew several people who used to say deleting spam isn't that hard, 5 years ago. After having the same e-mail address for 7-8 years, they can't empty out e-mail often enough to get rid of the spam on machines with 10Mbyte quotas for e-mail. At last count, they got 3-4 hundred a day.
Filtering spam leads to the possiblity of false positives, which is bad. The extra bonus is that, you don't have to pay 3 cents to send e-mail, but your mail will get rejected if the receiver expects 3 cents to read it.
I believe that in the end, the best solution is one that involves keyrings, and verifiable e-mails. I'll happily receive e-mail from anyone who attaches a PGP digest. I can build up a ring of people whom I trust to identify spams, and spam e-mails. If they register that key as a spam sender, my mail reader will filter it for me automatically, and my server will reject the e-mail when it sees the sees the key is that of a known spammer.
The cost of generating new keys, and the distributed process of identifing known spammers will solve the problem pretty easily. The hard part is being the first guy to transition to it. It'd be even harder then transitioning to IPv6. Which means it'll never happen.
I'll second this. I've got 3 or 4 machines that have this hardware in them, they run no problem. I don't do NT. I'm doubting very highly, they completely screw up Windows NT, but run just fine under Linux. I only have Linux, and they run like champs. I get them in the pre-assembled white box place that has cheap PC's locally. It's the only network card they have pre-installed. They work just fine. The poster might have one that's all screwed up, but that line of hardware runs fine.
Kirby
Re:Someone should start a BSD C/C++ compiler proje
on
RMS Turns 50
·
· Score: 1
Ummm, as I recall the BSD guys did have their own compiler (clearly before 1984 they did as GCC didn't exist then). I don't believe their compiler was the AT&T compiler until the day they switched to the GCC compiler. I thought they had their own until they eventually adopted the GCC compiler, and they dropped it years ago in favor of GCC, precisely because GCC was a much better compiler.
Pulling the plug on a machine, makes it really hard to get the bits off the disk. Destroying a filesystem takes a really long time ( I mean making it unreadable to a determined person who has access to the drive).
The point of encrypting a drive is so that if the drive mechanism itself is stolen your safe. There is some saftey in having data encrypted on a life filesystem, but the saftey of the files is limited unless you encrypt each file/directory/partition with a different key. If they all use the same key, I'll just compromise the OS and read it if the filesystem is mounted. Otherwise, I just compromise the OS and wait around for you to access the encrypted filesystem, and steal it after you put the key in.
Encrypting the filesystem is also handy for drive disposal. If you always write to the encrypted drive, then you can just give the drive to anyone you want and they can't get any data off it. Which means that hospitals or other places with sensitive could feel a lot safer that medical data isn't being given away every time they auction old computers off, or everytime a laptop gets lost.
Swap has a lot of stuff on it. Lets say I know that your using ssh-agent, this stores the decrypted private keys in memory so it's never written on the disk (the private key is normally stored encrypted via a password on disk). So if I pull the plug on the computer, it was in RAM so it's gone now I have to know your password. However, if I can load up enough stuff into memory to force that to be written to swap, now if I pull the plug, I just steal the swap disk (or boot into single user mode and copy the swap partition). Granted this is predicated on me having physical access to the machine, but if swap isn't secured, I've got a legitimate attack on your machine to get me enough information that I can be you, and all I had was a regular user account (or access to an open console), and physical access.
Slashdot Mirror
Just because they can't win doesn't make it copyright infringement right. That isn't a justification for being a copyright infringer. Notice I avoid using the term theft, it's not.
I think the RIAA should change their business model. It's why I pointed out that the RIAA doesn't have the right to a business model. It's why I hope that young musicians figure this out. The funny part is most of the musicians I know, actively avoid downloading anything that they don't have a right to (ie: isn't directly from the band). One guy when he realized just how bad it would be if he ever made it, and it was done to his music, that tracked down every band he could and purchased the CD's, precisely because he wanted to be on the up and up. He deleted everything else he couldn't didn't personally own.
Kirby
The reason this hasn't been a problem in the past, was that being a copyright infringer wasn't free. Now it is. Doing it to scale, and not getting caught was difficult. The internet and technological advancements have made that possible, and why it's a problem now.
You might not have bought the song either way, but your getting the enjoyment of said song. That has some value to you. If it didn't, you wouldn't have downloaded the song. Or at the very least you wouldn't have put it up on display for others to download. So clearly the song has value to the people whom are putting it up for download, and it has value to some of the people downloading it.
Those songs aren't naturally occuring objects. They don't just grow on trees.
Do you understand the Lockean princepals that are the rational and foundation of our current system of gov't? Know why people can claim they own land? Go read John Locke's "Second Treaties on Government". Very good book, you'll learn a lot about ethics and princepals in it. A lot about the justification for gov't, ownership, and property rights.
They can claim ownership by working the land, and improving it. You can claim ownership because you've put forth the effort to change and create the land from it's natural state. That's how one claims ownership of such things. I think that the individual artist, and the corporations behind them have put in the work to create the objects, and thus have a moral right to their ownership of the music.
Personally, I think that P2P networks should be left alone. They are fine constructions that have legitimate purposes. I think that if the RIAA is going to go after individuals who are copyright infringers, that's great. I think the people they went after recently who created search methods is wrong. I think those people should have been left alone. I think the RIAA should just crucify several copyright infringers in court, and keep doing it, until people realize the risks. It should continue doing so, one after another. It's illegal, it's copyright infringement, it's against the law. That's all there is to it.
Copyright is a *WONDERFUL* thing. It's what makes the GPL tick. It's what makes being a writer, and a programmer a viable proposition. It's what makes so very many occupations work.
Copyright intentionally makes scarcity of non-scarce objects, for the specific purpose of creating economic value. Did you catch that? It was an intentional construction, put forth by the founding fathers, who clearly thought about the matter at length. I think fair use is a good thing. I think making backup copies is a good thing. I think copyright is a good thing (not as currently implemented in the US or internationally by the Berne Convention).
Now, I think that copyright is a screwed up deal. I think they are entirely too long, and that they are fundamentally broken in that respect. However, you should respect copyrights. If you don't, there are innumerable things you enjoy which will disappear, precisely because they are what creates the economic incentive to create. That's why the founding fathers created copyright.
If you've got a problem with the Music business, stop consuming their product, support a different product. The music business doesn't have a right to a business model, so they can't just randomly sue people
Clearly it's novel, if absolutely no one did it, while such technology existed for a long time, and no one did it. It most definitely has utility, it is useful.
Now the last standard is this: non-obvious
Now go all the way to the bottem of that, and we might see what the problem is...Looks like the patent office, has a problem with web-based technology. If you can't find documentation where you take an idea from documents A, B, and C, and get the idea in the patent. What if the problem is, it's the god damn web, and it's all oral tradition?
You might have a case for non-obvious, but clearly a judge didn't agree with Barnes and Noble on the point. Now it might be obvious to most people now, but was it obvious in 1996-7 when they applied for that patent. Not sure, I didn't know enough about the web at the time to have been considered an expert in the area. At the time, I didn't know what a cookie was.
I think the Internet is too much of a land grab at the moment, and that too many simple basic ideas right now are coming out, which make things tricker on the internet.
Again, I'm still standing up being a devil's Advocate. I really, really wish, that it was just over.
I think that concept of saying: "Put it on my tab", existed in the real world for a long time, and that this is merely an application of such an idea to the internet. I think that the referral business relationship where by Amazon pays you a piece of the pie if the referral comes from your website, and the associated technology is stupid. I think that saying, well I'm going to automate my process using a computer, shouldn't be patentable, however, doing that falls well within the definition of what is patentable.
I'm trying to say what Lessig says so much better: Read it here
Remember a patent is a supposed to be specific invention. If you can have a check out in a different way then they patented, you should win the patent case. Hell, patent it, so you can tell Amazon, that clearly the patent office thinks it's unique.
People did actually expend an immense amount of effort to find prior art. I think you'll actually have to give that one to Amazon. At least they didn't patent selling shit over the web. So my guess is that, they really did concieve of a single click patent. I believe that most trivial coding idioms will have prior art. They aren't idioms until they are in common usage.
So, what would it take to implement a single click check out not using their patented method? I'm not really sure, I'd have to go read the patent closely. If somebody wants to piss them off, just extend all of their patents trivially, so if they ever try and do anything more then what they've patented they have to come see you about it. It'd be lots of fun.
FYI: I'm playing devils advocate here mostly. I think the Single click patent is stupid. I think that they patented something obvious, not something novel. However, patents of this nature have existed for a long time in other spheres, it's just that people are starting to see them in computers.
Kirby
They don't have a copyright on the algorithm, They have copyright on the specific version they implemented. I'm guessing that patents (the actual patent document) are considered part of the public domain, so the original description of the algorithm is probably not copyrighted at all.
Sorry.
I know copyright is completely different then patenting, both of which are different then trademarks.
I'm an idiot for typing the wrong word on slashdot... I'm not unique or special....
Kirby
I won't be shocked to find out that, this whole deal is that SCO copied a BSD driver, and then Linux did, and now they have absolutely identical code in them.... Uhhh, but BSD owns the copyright SCO, how are you suing for copyright infringement. Oh well...
Kirby
Okay, now I might have to get a several hubs. Lets assume I got HP Pro-Curve 80 port capable switch and filled them all the way to the gills, I'd need 3 for 200 drops, and they'd cost no more then $8K fully loaded if I remember correctly. Again that's MSRP, not a price I'd use if I did this for a living. I'd need probably another $3K in hardware (screws, face plates, Communication racks, and what not). Total cost for install, for a completely network that is ready to rock and roll:
$6K (wires female ends) + $1K (tools) + $5K (spectrometer) + $8K (switches) + $3K (misc parts) =
$23K of which, $6K are one time costs to get into the business. So for a stock install, the total cost is $17K + Plus 2man weeks. Assuming they couldn't negocaite a decent price. It'd take me 2 weeks to do all the work personally, and I'm not very speedy at it. I'd bid it at no more then $30K-$35K if I was doing it personally. In two weeks, that's roughly $13-$18K in my pocket. Not great for a contractor, but nothing to cry about either. What'd they do that I didn't price out, where'd I miss the price too badly. I intentionally over-estimated everything on price. I could do the job for parts in a lot less then $6K is my guess. Nobody runs that many 300 foot runs. If I'd gotten cheaper female jacks in the cabling closet, it's have been better. Oh, and my prices are about 3-4 years old. I believe everything is cheaper now.
Heck, I'll bet I could run fiber to 200 rooms, and put copper to fiber transceivers in on both ends, or fiber on one end, and used fiber capable hubs they quoted, and still make a decent profit. The transceivers might be a little price, last time I read about it, I thought they ran, $45 a piece, but still 45 * 200 = $9K. So $80K is pretty expensive, assuming there really was conduit installed.
Kirby
Second, quality of service is trival to handle now out of band from TCP/IP. QoS is also a rats nest of stuff that will drive up costs. Differentiation of traffic is a neato way to have extra costs foisted on you by the upstream ISP. Multi-homing can be done today by anybody who knows about TCP/IP. Now getting your own AS number and having the BGP protocol broadcast that two ISP's can get to your network is a problem.
IPv6 might have a magic solution for this, I haven't read up on it enough yet. As far as I know they just added 2 octects to the routeable IP space (while they added another 6 octects to the local space). So essentially, you only have to route to the first 6 bytes of the IP, after that, it's the local link/router's job to actually deliver the packets from there.
Kirby
As long as you do the compiling yourself, I don't believe there is anything you can't do with GPL'ed source. The GPL only comes into play when you start to distributed the source to other seperate legal entities.
Kirby
Kirby
I suppose if it's documented to only work in certain cases, that's acceptable, however, the the code that calls it without checking for the input is then broken, and buggy. It should be fixed. If it can't be checked before calling the functionality, then the functionality better work for all inputs. That's good software. Stuff that just assumes that unsafe input will never, ever be put in, is a bug. A security hole. It's not reusable code. Reusable code, checks inputs. Reusable code fails gracefully. Reusable code, returns error codes indicating invalid inputs. Reusable code doesn't have security flaws in it.
Distributing code that won't handle all input cases for use in a public distributed computing project for the sake of speed is irresponsible, and stupid. Now, I'm a lot more likely to just never run one of the distributed projects then to risk security flaws if they are willing to sacrifice security for their speed. Security should be the winning factor in all concerns when writting software. When trading security for speed, is an option don't take it. Security or ease of use, take security. Security or correctness, re-write the software using a new protocol, or new algorithm, but still take security and document the correctness flaw. Right now I only run them on machines that don't have any valuable information on them, but I'd prefer they not be used in a DDos, so it'll probably get stripped off all my machines.
Now, NVidia might do lots of nasty things, but they still write the best damn 3D video driver on the planet (for consumer level hardware under Windows). Now it might only work because they structure their hardware so it works more effeiciently with software, so they can lower costs, or because they have some Algorithm's expert doing stuff nobody else has every thought of, or because they figured out how to get a few key loops to fit in less then 3 lines of P4/AMD L1-Cache. NVidia's driver is in fact a competitive edge. Giving up that structure to the public domain is bad for NVidia stock holders. It's bad for NVidia, but good for the public. Having all drivers of that kind of quality available to study for new drivers would be good for the public, it would improve quality, and speed of video hardware, and lower the cost. It'd be great for the public.
It'd be great if Coke gave up the formula for Coca-Cola too, but I don't expect them to do it, and I'd surely hope they don't for the sake of their stock holders. I wish NVidia would give up the secrets. I wish they did open source the driver. I wish for a lot of things, that aren't going to happen.
Kirby
Kirby
There can't be an Engineer in software. As Alan Cox has written on several occasions, software engineering is roughly in the alchemay stages in relation to state of the art chemical engineering.
In a real engineering, process there are things to check, verification, and known facts which can be double checked against the standing design. The design checked against the implementation.
Building a bridge is a much easier process then say writting a secure OS UNIX clone. It's easier in the sense, that it's been done so many repeatable times since the Romans. The understanding of the underlying structures of a bridge are well know. All intereactions between the people on the bridge, and the bridge are known (hard to calculate them all, but reasonable well known). Recently (in the last 50 years), we learned a new trick about bridges. That the cross sectional area, and wind can interact with forces that will tear down a bridge. Okay, that as far as I know was the last major mistake that was a complete unknown in bridge building. I believe it happened sometime in the 50s. You know right after the first couple of computers we're built. We don't know anything writting software in comparison to bridge building.
Building a Secure UNIX or UNIX clone has yet to be done once. The understanding of all the various layers of software spans when writting an GUI application all the way down to the quantum effects that happen in silicon, I believe they claim 12-15 orders of magnitude of understanding. That's a lot, possibly more then any other intellectual endeavor ever undertaken.
About the only places that can stand up and say, they follow an Engineering process are places that are SEI certified Level 5. They have a repeatable, measurable process by which they do things. That starts to sound like Engineering. Real Engineering is very hard, very tedious, and very boring. Most code jockeys I know, couldn't do it. It'd drive them nuts. You've got absolutely no right to be called an Engineer on an off the shelf software application. If you write software for a company that foists on you a horrible, gut renching process, of checks, double checks, that involves throwing away everything that doesn't meet their excating standards, now you talking about it. Where every single index is documented to not be possible to overflow the array, that's Engineering. I've never seen anybody ever do it, but if it we're Engineering, it would have to be done. Even after all that, you still have to be working in an area where what your doing is extremely well understood, and has been done lots of times before. Where all interactions between every module is well researched, and well understood. Where all the compenents have well defined qualities, that react in a statistically predictable manner in all situations.
Have you considered what solar flares would do to your software? Do you have a fault tolerant memory storage scheme to account for the bit flips that will occur because of it? Do you use an operating system, and do you understand all of interactions between the various parts? Have you documented why everything works the way it does? Do you document every single change with a full risk analysis of why it's being done, and how it only improves the reliability of the system. People in Engineering do that. People who worry about stuff like that, are Engineers. It's not that Engineer's do
One example story was here: Stats for Feb
Posted by timothy, submitted by Dan, with a link to bsdforums.
Here is another story: Stats for January
Look at that posted by timothy, submitted by Dan with a link to bsdforums.
Like I said, I didn't mind the first or second time it was posted, but it seems silly to post them on a monthly basis. I suppose I can start submitting the number of LOC a base Linux kernel has.
There are lots of interesting stories about BSD, this isn't one of them.
I'd rather see them post interesting things about BSD, I'm sure they are out there, this isn't one of them. Statistical updates about NetBSD that consist of a single number seem dull to me. If they had a list of interesting packages, rather then a list of 64 packages of which there are 3 or 4 sets of 6 inter-related packages, so the number isn't particularly meaningful anyways...
Kirby
While I don't mind the occasional, hey look at this new cool software package, this seems completely out of place on Slashdot. Monthly postings of the number of packages just seems incredibly bizarre. It's not like you are posting a new revision, a new feature, a new major release, major changes. In short, this isn't news worthy, probably not even to people who use NetBSD on a daily basis. If there we're some compelling packages it would be better. It's not like they just finished porting KDE or Gnome, or MS Office to it. That would qualify as slashdot worthy.
I thought the first time it was done last month, it was weird but okay for slashdot, however, I can't see the use in posting it every month. What is there to discuss? It is merely a statistic, there isn't anything compelling in the story to discuss, hence my meta-discussion of if it is worthy of discussion.
Kirby
Now, if you are trying to avoid paying for the ICA client, but you are willing to pony up the money to Microsoft, then there is a client named rdesktop that does the Microsoft remote desktop protocol (RDP). It was reverse Engineered from scratch, and supposedly is reasonable stable. So now, you can run this on Linux desktops, but you still have to pay Microsoft a bunch of money for the apps (just because they are all running on one server, doesn't get you out of paying them for as many concurrent users as there could be).
Now, if you have to have Microsoft Applications, but not a Microsoft desktop, you might want to see the guys who develop the Crossover stuff. Now you can run a lot of Windows Apps on a Linux box that has a Wine processes running remote. The product is called Crossover Office Server Edition I don't follow the legality of this, so get a real good lawyer before you try it out. looks like CodeWeavers is saying, you get to pay Microsoft a bunch of money.
This is probably roughly the same quality, but now your talking about using X for your network transport. Which is a little awkward for remote users, as they will have to run an X server. Cygwin ships with one for Windows desktops.
Now, if all you want is a bunch of desktops you can run remote from a linux server. Get a bunch of machines that can act like X-Terminals. A bunch of old cheap PC's with a good NIC will do the job, as long as the NIC will do PXE, or netbooting of some flavor. Go get PXES from sourceforge and run it. It will net boot, and run rdesktop, a Citrix ICA client, or run as an X Terminal for you. It is very good, and runs pretty well. This is what the city of Largo, FL does. They claim it's great, grand, glorious and best of all, dirt cheap.
I don't understand your requirements. They appear to be directly contradictory. We have to have cool stuff from Microsoft, but we can't afford to pay for it. My guess is the cost of the Citrix Clients isn't nearly as bad as the cost of all of the copies of MS apps you sound like you want to run. Anyways, these are some pretty decent ways to get remote desktops. However, with Microsoft, you don't really get a break on the pricing that way, it does simplify administration of the desktop, and makes replacing broken hardware much easier.
Kirby
However your usage of the word "your" in that sentence is completely wrong. The word "your" implies ownership. In fact, it is nothing more then the possessive form of the word "you". Does he own anything in that sentence? He doesn't own Ali G, the dengerate faction, or the verb representing. So clearly the possesive form is absolutely wrong in that sentence. Did you mean the contraction you're? No because that wouldn't make sense in that sentence either, and it would be a gross error on the order of using "a" instead of "an" after editing a sentence. Did you just construct a piss poor English sentence after being a grammer Nazi? Yes, it would appear you did. The correct wording would have been:
If you'd like, I can explain in excruciating detail about why that usage is better then the sentence you worte. I'd be a prick for doing so, but I will happily donate my time to an English lesson for you. Perhaps you would post under a named account, so I can go nit pick every post in your history.
I reword sentences in my posts often, and periodically that means I have mistakenly not corrected the corresponding articles, tenses of the words I should use. If English wasn't so sensitive to such things I'd be much happier, to me that are irrelavent for the most part. I'm attempting to convey information, not get it published in a book, or try and prove my superiour skill in the written form of formal English. This is informal language used on a public website. The situation doesn't dictate to me, that I should take the time to proof read it in detail, when my concept is clear. If it had been a clear cut mistake that was repeated several times in the post you would have a point. However, I've seen published books with more glaring mistakes that that. Those people were actually trying to be grammatically correct in their usage of the English language. I however wasn't.
Feel free to correct all the subtle mistakes I have made in this post. I'm not very careful about checking my posts for grammatical errors, so I won't be surprised to find out I made several.
Kirby
Which means my guess in this post is reasonable close.
The regular ISO's they will see as something nice to do for the community, and have it be the test bed for new features, and the "beta/gamma" release of the upcoming Enterprise series. Then the new enterprise releases will have lots of software that has been tested and released on all kinds of hardware and they will have a very good chance of making a very, very stable release for the enterprise lineup.
Because they have the stable release, I believe you'll see fewer, and fewer X.1 and X.2 releases, and you'll see a lot more .0 releases. Somewhat because it will be enticement to have people buy the Enterprise line, and somewhat to keep the "beta/gamma" testing on the bleeding edge. It's a pretty clever scheme all in all. If they can pull it off, and keep the bugs in the .0 releases down, and put out a .1 to solve big problems in .0, I'll happily use the standard ISO on my desk, and happily pay the money for the Enterprise lineup for my servers.
Kirby
The next edition of RedHat I believe is supposed to include the new kernel threads stuff, with the glibc that supports it (hence re-implementing pthreads), it has a new compiler, and the new glibc. So probably the applications aren't binary compat with 8.0, so this is now 9.0. The price you pay for upgrading. It's not like the upgrade path doesn't work, and it's not like upgrading past these things will be vastly superior on Gentoo.
They are pushing out new big things, if you want to stay current, then upgrade to it. What's the big deal about the major version number? I really don't see why your panties are in a bunch with RedHat. Gentoo will do most of the same crapola to your machine that Redhat does when you upgrade, it just won't have a major version number change. Big whoop.
Kirby
I knew several people who used to say deleting spam isn't that hard, 5 years ago. After having the same e-mail address for 7-8 years, they can't empty out e-mail often enough to get rid of the spam on machines with 10Mbyte quotas for e-mail. At last count, they got 3-4 hundred a day.
Filtering spam leads to the possiblity of false positives, which is bad. The extra bonus is that, you don't have to pay 3 cents to send e-mail, but your mail will get rejected if the receiver expects 3 cents to read it.
I believe that in the end, the best solution is one that involves keyrings, and verifiable e-mails. I'll happily receive e-mail from anyone who attaches a PGP digest. I can build up a ring of people whom I trust to identify spams, and spam e-mails. If they register that key as a spam sender, my mail reader will filter it for me automatically, and my server will reject the e-mail when it sees the sees the key is that of a known spammer.
The cost of generating new keys, and the distributed process of identifing known spammers will solve the problem pretty easily. The hard part is being the first guy to transition to it. It'd be even harder then transitioning to IPv6. Which means it'll never happen.
Kirby
Kirby
Kirby
The point of encrypting a drive is so that if the drive mechanism itself is stolen your safe. There is some saftey in having data encrypted on a life filesystem, but the saftey of the files is limited unless you encrypt each file/directory/partition with a different key. If they all use the same key, I'll just compromise the OS and read it if the filesystem is mounted. Otherwise, I just compromise the OS and wait around for you to access the encrypted filesystem, and steal it after you put the key in.
Encrypting the filesystem is also handy for drive disposal. If you always write to the encrypted drive, then you can just give the drive to anyone you want and they can't get any data off it. Which means that hospitals or other places with sensitive could feel a lot safer that medical data isn't being given away every time they auction old computers off, or everytime a laptop gets lost.
Swap has a lot of stuff on it. Lets say I know that your using ssh-agent, this stores the decrypted private keys in memory so it's never written on the disk (the private key is normally stored encrypted via a password on disk). So if I pull the plug on the computer, it was in RAM so it's gone now I have to know your password. However, if I can load up enough stuff into memory to force that to be written to swap, now if I pull the plug, I just steal the swap disk (or boot into single user mode and copy the swap partition). Granted this is predicated on me having physical access to the machine, but if swap isn't secured, I've got a legitimate attack on your machine to get me enough information that I can be you, and all I had was a regular user account (or access to an open console), and physical access.
Kirby