As I was reading through this article, one statement really jumped out at me:
I think instead of new applications, a significant development will be localization and personalization of Linux distributions for different uses....
-Rahul Chopra
I think this is significant because it's one area where Linux and FOSS software has a clear advantage over MS -- one which I don't think they can ever overcome. Microsoft has internationalized their products for a great number of other languages and locales, but as was seen recently in Israel, if the internationalization is difficult (L-to-R scripts), and the market is relatively small, it just doesn't make economic sense for them to do it.
Open Source, on the other hand, works according to another economic model, one which is not limited by profit-loss ratios and ROI. If you have people interested in it, you can create an internationalized version of a package for any audience. Now, there are still complicated technological issues (such as some of the really complex scripting systems in many of the smaller markets like SE Asia), but once we get past some of the difficult hurdles of creating truly flexible font and glyph servers and text rendering systems, we will see Linux and FOSS expanding into places where MS cannot hope to go. True, these won't bring in gobs of cash for Linux developers and ISV's, but I think we will see steady progress made. We will soon see Linux as the foundation for technological, and ultimately economic freedom for the majority of the world's governments and citizens.
The Borne Shell is what I would call, The Mother of All Shells. Other shells (ksh, bash) are usually a superset of/bin/sh. Csh (and its ilk) is another beast entirely, but I never really considered it a REAL shell for doing programming in.
Basically, if you have a *NIX like system, you are pretty much guaranteed to have a/bin/sh, and it will always work the same. Bash is common, but not universal. Ksh is proprietary, though there are open source versions (pdksh), but you won't find it everywhere. Csh is an abomination. Other shells are interesting and useful, but you'll never know if they are on any given system, or if they will run the same.
If you want a script that will run the same everywhere, then use/bin/sh.
(Of course, this isn't even getting into differences between "awk", "nawk", "gawk" and all the other tools and their various incarnations that you end up using in shell scripts, but at least you can have some hope that the underlying shell will be predictable.)
I don't see a link to the actual article, but I wonder if the author has any real evidence of this forking, or is he simply working from arm-chair philosophies? Sure, it's possible in theory, but where are the examples? In the case of a fork, does one side quickly die off, and the other branch simply re-absorb the developers? Is forking no more than evolution at work, with the strongest (i.e., the strategy most supported by users) ultimately surviving in the end? Is the proprietary model an example of one person's vision of how a project should work being forced on users? (examples: horrendous implementations like "Bob" and "Clippy"?) Is the proprietary approach an example of the useless features and software bloat caused by isolated software developers, dwelling in their ivory towers, huddled around clueless "focus groups"?
Sure, it's all well and good for a bunch of "researchers" to sit around and pontificate on the "Dangers" of one development approach or another, but until I see some hard numbers and indications of actual long term effects, I'm not impressed.
> Its not a matter of throwing them away. A simple computer program can reformat a card from VFAT into ext2...
Heh.... I would posit that, for 95% of people, performing a low level format on a memory card is equivalent to "having to throw it out." Most people want to pop their Flash card into a convenient USB connected reader or some such device and have the OS simply mount it as another drive so they can copy things off it like a floppy. You might be able to provide low level drivers for that, but I have a feeling you're going to have a hard time getting Joe User to get it working.
> a patent must to be vigeriously defended, just like a copyright.
Actually, as has been pointed out in numerous posts, this is NOT true. This is only for Trademarks, and perhaps copyrights.
> I don't see how they were skating on thin ice.
The people skating on thin ice were the people USING the patented technology without proper permission. The fact that MS wasn't enforcing the patent meant they could get away with it, for the time being, and implement the FAT on their devices. However, now that MS has chosen to enforce the patent, the manufacturers are caught between a rock and a hard place...
Actually, 16K and 32K limits were mostly because memory was so expensive then. The real limit, before DOS (?) came up with a way around it, was 64K, since that was the highest number you could directly reference with a 16bit integer (2^16-1 = 65535). I worked on "top end" CP/M systems, which ran on a Z/80 processor, and long ints were 16 bit. Hence, the most memory we could put in them, regardless of cost, was 64K
Sounds like ext2 is already there, with simple implementations that would fit nicely inside some small embedded device.
However, that's sort of a moot point. The problem is the hundreds of thousands of devices out there already, not to mention the millions of memory cards used to connect to those devices. You expect everyone to throw away all those nifty 256Meg cards they just bought six months ago and buy new memory for their camera/PDA/smartphone or whatever?
> If you already know how to do it then you can safely ignore this offer.
That would be true if they were selling a specification for the file system. A specification just helps you figure out how to interface with another system or device, so you don't have to figure it out on your own.
In this case, this is a license to use the patented technology. In other words, if you are using a FAT file system in some device, then you've been skating along on legal thin ice. Well, the thin ice just broke under you, and you now have to anti up big $$$ to continue to use it, especially if you are distributing devices with an embedded FAT filesystem (e.g. Flash memory).
(I'm still wondering how this will affect the Linux kernel, since it has support for FAT file systems. I wonder if Linux is going to have to drop the support, of if we'll be able to slip in under the "interoperability" loophole.)
Seems like this is a repeating pattern. Site is running Server A, but redirects to Server B. I wonder how common this is, especially with servers running different platforms.
Of course, it makes for a convenient way to keep pulling on the thread, and then quit when you find what you want to find -- in this case, a server running IIS.
The Google search didn't find me much, but the Kerneltrap article has gobs of useful information, as does the people.redhat.com reference. Sounds like the big difference is the fact that you have to upgrade a bunch of tools, like the modutils, in order to get it to work. I don't want to try this on my main development system yet (i.e., the box I need to use every day to do the work I actually get paid for), but I should be able to set aside a test RH9 box to try it on.
Thanks again!
(Just have to make sure I make good backups of any of my critical data before moving forward...;-)
One of the things I've noticed with the 2.6 kernels is that they are significantly different from the 2.4 kernels in their setup. I've been downloading and building kernels since the 2.2 series, and though I've tried to build perhaps a couple dozen instances of the 2.5/2.6 kernel, I have never been able to get one up and running!!
Does anyone here have a pointer to a decent HOWTO type document on steps required to take a system currently running a 2.4 kernel (like, for instance, a RH 9 release) and migrate it to a 2.6 kernel?
I'd love to help out with testing, but after blowing upwards of 80 hours beating my head on a brick wall trying to get it to compile and boot up, I've pretty much given up.
Is this simply a compressed version of the internal XML? Even OpenOffice compresses the data before saving it to file. Otherwise, the document files would end up being HUGE!
> then all of the FUD about linux in general and the GPL specifically, stays out there without being ruled on.
I agree with you that IBM buying SCO would be a lousy solution to the problem, but I don't think it would be as bad as you make it out to be. If IBM buys SCO, it gets the rights to all that nifty UNIX code that SCO now owns. What would happen if IBM simply turned around and released all that SCO code under the GPL? For one thing, everyone and his brother could do their own code analysis to see if SCO was telling the truth. Also, any useful code that was in the ancient UNIX source would now be available for anyone to use (under the usual GPL restrictions). True, it wouldn't be as tidy a vindication as we'd all like to see, but it could certainly put a lot of interesting code out there in GPL land.
(Of course, there may be other complications, such as proprietary third party code that even SCO/IBM couldn't release because it belongs to OTHER people, but whatever got out there would make for some interesting reading.)
> They are mechanically capable of performing the maneuvres, however carefully scripted, hence the problem is now in intelligence...
Yes and no.
The easy part to agree about is the fact that creating an "AI" to deal with the complexities of the real world will be several orders of magnitude more complex than simply walking across a flat floor. That's a given.
However, even with a simple knowledge of human anatomy, there is still a huge amount of complexity built into our structure in order to deal with things like uneven ground, loose stones, cracks in the pavement, walking into the wind, navigating a steep slope, etc. Consider the number of "degrees of freedom" in a human joint, or even the bottom of your foot! Part of the problem is the AI -- figuring out how to shift your weight if one foot suddenly slips -- but there is still a lot to do in the mechanics of the various joints, and coming anythere close to the flexibility of the human body.
Of course having said that, the question is, what is the goal? If you want a robot that can navigate a complex path but over a more or less predictable surface, dealing with fairly run-of-the-mill obsticles such as doorways, stairs, small obstructions and simple slopes (e.g., a factory floor), then you're probably pretty close on the mechanical end of things. The problem then really does boil down to the AI, especially if you want these things to be USEFUL for something, not just pretty walking machines!
> interesting that the mechanical challenges are no longer the stumbling block.
Not so fast... Remember that these robots are still operating in highly controlled environments, performing a small number of carefully scripted and highly constrained maneuvers. Granted, these are no small feat, but they are still a long way from navigating complex and unpredictable environments -- something humans and other "biologicals" do all the time without even thinking.
Again, these robots have made huge strides (pun intended), but they have a long way to go before they can go out and take a stroll through a field or walk down a busy city sidewalk in traffic.
> What is to stop microsoft from flooding the (alternative) market with it's own Linux clone?
Microsoft is.
Seriously, there's nothing stopping MS from coming out with its own distribution, other than the fact that they would be undercutting their own market more than they would be undercutting RedHat or anyone else. If it's the black market you're talking about (or the "gray market"), and MS putting out boxed sets that claim to be official Red Hat software, then I doubt that MS is worried enough about RedHat's income enough to risk anything that close to being illegal. (They generally only engage in blatantly illegal activities when there's a lot of money to be made off it.) If they take out RH, then there are several hundred other distributors to jump into the gap.
> The only way to protect yourself from unwanted outside connections is with correct crypto code.
SECURITY 101: The only way to really protect yourself from unwanted connections from the outside world is to unplug from the network. Of course, that's hard if you're trying to build a Web Service. Even that isn't a guarantee if you can't provide physical security to prevent access to the system console. There's a handy little floppy boot disk I've seen that will break into any Windows box made, though it won't help you if the file system is encrypted. I have a feeling there are similar exploits possible on Linux or other UNIX systems if you can get to the physical box.
Point being, security is a question of choices and compromises. What series of choices (such as leaving a ssl port open or closed) gives you an acceptable risk, and still allows you to do what you need to do?
I can't read ESR's mind, but I strongly suspect that this is intended to provide ammunition for a counter-suit claiming that SCO has pirated GPL code and illegally embedded it into their software. The comparison isn't a proof of copied code, but it could indicate "hot spots" and provide sufficient basis for a sort of "search warrant" to force SCO to show its code. If it turns out, as it did in the famous AT&T vs. BSD case, that SCO has been whole-sale ripping off other people's code, then things might turn really uncomfortable for McB and his cronies...
Do I smell a Court Order in the works? If you really can do this without divulging the original code, then someone could conceivably convince a judge to issue an order to have a "neutral third party" create the MD5 sums on SCO's codebase, giving us a chance to look for pirated GPL code hidden inside of SCO proprietary products, without having to look directly at the SCO code.
I know about HAL in the story "2001", but I don't remember anything that could be called a "bug". HAL was operating according to the instructions of its original programmers, instructions which the actual astronauts had no knowledge of. This led to HAL killing off several of the crew, but other than that, I don't remember it actually malfunctioning. It was programmed to proceed to it's target at all costs, and that's what it did.
What am I missing?
(The linked articles didn't give any hints either.)
> Their complaint is that the MEDIA is out to get them.
Actually yes. The complaint is specifically that other entities, such as Red Hat, Novell and the FSF are acting as (paid) proxies for IBM, and that, "they have Eric Raymond on their payroll."
SCO is still trying to smear IBM in the Court of Public Opinion, to try and pressure them into settling outside of a real court, where they know they will be out-gunned, out-financed and backed up by flimsy arguments and scant facts.
I agree with you that IBM is probably pulling some strings, especially in the media, but as Eric said, "Those attacks have been happening because our community is outraged. IBM didn't have to talk to us or suborn us or bribe us or anything else."
> Now, if there are bans or restrictions on UCE/spam, they could be applied to me if someone decided they didn't like me.
Well, this is exactly why the process of making laws is so difficult. It'd be nice if we didn't need any laws, and people would just act nice! The reality is that people don't act nice, and powerful people beat up on or take advantage of less powerful people. In theory, laws are written to protect us from the violent or dishonest. But, it's impossible to write laws that always apply all the time, and can never be misinterpreted.
So, that's why we have both the Legislative branch (i.e., people writing the laws), and the Courts, who interpret the laws. Complain all you want, but the court system is designed to take imperfect laws and apply them in a case by case basis. Sure, there is no way, by just looking at the legislation, to tell the difference between a scum-of-the-earth spammer and a legitimate free speech case, but we have judges and juries and yes, even lawyers to try to apply some kind on intelligence to the laws and apply them in a just and fair manner.
Not sure if that makes you feel any better, but at least that's the theory...
> How do you get out of the mandatory 60hr weeks so you can ride with decent daylight?
Unfortunately, I don't know a good solution for the 60 hour part, but I know that lights (BIG lights) help the riding at night part. When I was commuting in the Winter, I did a lot of riding at night. I did pretty well, though you are right about keeping your eyes (and ears) open at all times. Sometimes you have to pick your routes carefully, and wear the most gawd-awful bright clothes. Wear anything that makes you stick out like a sore thumb.
Actually, riding at dusk is probably worse than riding in full dark, because the lights stand out more when it's dark. So, you might do better waiting a little longer (after rush hour) and then getting the biggest lights you can afford.
(Funny part is, people always thought I was insane for riding in the Winter, but I always found it one of the best times to ride. The simple fact that there's this loony out there in the snow, riding his or her bicycle raises the "weirdness factor" to the point where people actually notice you, and generally give you lots of room as they pass. Just look out for the snowplows...)
Slashdot Mirror
Open Source, on the other hand, works according to another economic model, one which is not limited by profit-loss ratios and ROI. If you have people interested in it, you can create an internationalized version of a package for any audience. Now, there are still complicated technological issues (such as some of the really complex scripting systems in many of the smaller markets like SE Asia), but once we get past some of the difficult hurdles of creating truly flexible font and glyph servers and text rendering systems, we will see Linux and FOSS expanding into places where MS cannot hope to go. True, these won't bring in gobs of cash for Linux developers and ISV's, but I think we will see steady progress made. We will soon see Linux as the foundation for technological, and ultimately economic freedom for the majority of the world's governments and citizens.
Basically, if you have a *NIX like system, you are pretty much guaranteed to have a /bin/sh, and it will always work the same. Bash is common, but not universal. Ksh is proprietary, though there are open source versions (pdksh), but you won't find it everywhere. Csh is an abomination. Other shells are interesting and useful, but you'll never know if they are on any given system, or if they will run the same.
If you want a script that will run the same everywhere, then use /bin/sh.
(Of course, this isn't even getting into differences between "awk", "nawk", "gawk" and all the other tools and their various incarnations that you end up using in shell scripts, but at least you can have some hope that the underlying shell will be predictable.)
Sure, it's all well and good for a bunch of "researchers" to sit around and pontificate on the "Dangers" of one development approach or another, but until I see some hard numbers and indications of actual long term effects, I'm not impressed.
Heh.... I would posit that, for 95% of people, performing a low level format on a memory card is equivalent to "having to throw it out." Most people want to pop their Flash card into a convenient USB connected reader or some such device and have the OS simply mount it as another drive so they can copy things off it like a floppy. You might be able to provide low level drivers for that, but I have a feeling you're going to have a hard time getting Joe User to get it working.
Actually, as has been pointed out in numerous posts, this is NOT true. This is only for Trademarks, and perhaps copyrights.
> I don't see how they were skating on thin ice.
The people skating on thin ice were the people USING the patented technology without proper permission. The fact that MS wasn't enforcing the patent meant they could get away with it, for the time being, and implement the FAT on their devices. However, now that MS has chosen to enforce the patent, the manufacturers are caught between a rock and a hard place...
Actually, 16K and 32K limits were mostly because memory was so expensive then. The real limit, before DOS (?) came up with a way around it, was 64K, since that was the highest number you could directly reference with a 16bit integer (2^16-1 = 65535). I worked on "top end" CP/M systems, which ran on a Z/80 processor, and long ints were 16 bit. Hence, the most memory we could put in them, regardless of cost, was 64K
Sounds like ext2 is already there, with simple implementations that would fit nicely inside some small embedded device.
However, that's sort of a moot point. The problem is the hundreds of thousands of devices out there already, not to mention the millions of memory cards used to connect to those devices. You expect everyone to throw away all those nifty 256Meg cards they just bought six months ago and buy new memory for their camera/PDA/smartphone or whatever?
That would be true if they were selling a specification for the file system. A specification just helps you figure out how to interface with another system or device, so you don't have to figure it out on your own.
In this case, this is a license to use the patented technology. In other words, if you are using a FAT file system in some device, then you've been skating along on legal thin ice. Well, the thin ice just broke under you, and you now have to anti up big $$$ to continue to use it, especially if you are distributing devices with an embedded FAT filesystem (e.g. Flash memory).
(I'm still wondering how this will affect the Linux kernel, since it has support for FAT file systems. I wonder if Linux is going to have to drop the support, of if we'll be able to slip in under the "interoperability" loophole.)
Seems like this is a repeating pattern. Site is running Server A, but redirects to Server B. I wonder how common this is, especially with servers running different platforms.
Of course, it makes for a convenient way to keep pulling on the thread, and then quit when you find what you want to find -- in this case, a server running IIS.
Thanks again!
(Just have to make sure I make good backups of any of my critical data before moving forward... ;-)
Does anyone here have a pointer to a decent HOWTO type document on steps required to take a system currently running a 2.4 kernel (like, for instance, a RH 9 release) and migrate it to a 2.6 kernel?
I'd love to help out with testing, but after blowing upwards of 80 hours beating my head on a brick wall trying to get it to compile and boot up, I've pretty much given up.
Is this simply a compressed version of the internal XML? Even OpenOffice compresses the data before saving it to file. Otherwise, the document files would end up being HUGE!
I agree with you that IBM buying SCO would be a lousy solution to the problem, but I don't think it would be as bad as you make it out to be. If IBM buys SCO, it gets the rights to all that nifty UNIX code that SCO now owns. What would happen if IBM simply turned around and released all that SCO code under the GPL? For one thing, everyone and his brother could do their own code analysis to see if SCO was telling the truth. Also, any useful code that was in the ancient UNIX source would now be available for anyone to use (under the usual GPL restrictions). True, it wouldn't be as tidy a vindication as we'd all like to see, but it could certainly put a lot of interesting code out there in GPL land.
(Of course, there may be other complications, such as proprietary third party code that even SCO/IBM couldn't release because it belongs to OTHER people, but whatever got out there would make for some interesting reading.)
(Hey - it's even under 3 lines!) -0- --0 000
Hummm... I thought that was supposed to be, "There Must be Fifty Ways to Leave Your Data..."
Yes and no.
The easy part to agree about is the fact that creating an "AI" to deal with the complexities of the real world will be several orders of magnitude more complex than simply walking across a flat floor. That's a given.
However, even with a simple knowledge of human anatomy, there is still a huge amount of complexity built into our structure in order to deal with things like uneven ground, loose stones, cracks in the pavement, walking into the wind, navigating a steep slope, etc. Consider the number of "degrees of freedom" in a human joint, or even the bottom of your foot! Part of the problem is the AI -- figuring out how to shift your weight if one foot suddenly slips -- but there is still a lot to do in the mechanics of the various joints, and coming anythere close to the flexibility of the human body.
Of course having said that, the question is, what is the goal? If you want a robot that can navigate a complex path but over a more or less predictable surface, dealing with fairly run-of-the-mill obsticles such as doorways, stairs, small obstructions and simple slopes (e.g., a factory floor), then you're probably pretty close on the mechanical end of things. The problem then really does boil down to the AI, especially if you want these things to be USEFUL for something, not just pretty walking machines!
Not so fast... Remember that these robots are still operating in highly controlled environments, performing a small number of carefully scripted and highly constrained maneuvers. Granted, these are no small feat, but they are still a long way from navigating complex and unpredictable environments -- something humans and other "biologicals" do all the time without even thinking.
Again, these robots have made huge strides (pun intended), but they have a long way to go before they can go out and take a stroll through a field or walk down a busy city sidewalk in traffic.
Microsoft is.
Seriously, there's nothing stopping MS from coming out with its own distribution, other than the fact that they would be undercutting their own market more than they would be undercutting RedHat or anyone else. If it's the black market you're talking about (or the "gray market"), and MS putting out boxed sets that claim to be official Red Hat software, then I doubt that MS is worried enough about RedHat's income enough to risk anything that close to being illegal. (They generally only engage in blatantly illegal activities when there's a lot of money to be made off it.) If they take out RH, then there are several hundred other distributors to jump into the gap.
SECURITY 101: The only way to really protect yourself from unwanted connections from the outside world is to unplug from the network. Of course, that's hard if you're trying to build a Web Service. Even that isn't a guarantee if you can't provide physical security to prevent access to the system console. There's a handy little floppy boot disk I've seen that will break into any Windows box made, though it won't help you if the file system is encrypted. I have a feeling there are similar exploits possible on Linux or other UNIX systems if you can get to the physical box.
Point being, security is a question of choices and compromises. What series of choices (such as leaving a ssl port open or closed) gives you an acceptable risk, and still allows you to do what you need to do?
I can't read ESR's mind, but I strongly suspect that this is intended to provide ammunition for a counter-suit claiming that SCO has pirated GPL code and illegally embedded it into their software. The comparison isn't a proof of copied code, but it could indicate "hot spots" and provide sufficient basis for a sort of "search warrant" to force SCO to show its code. If it turns out, as it did in the famous AT&T vs. BSD case, that SCO has been whole-sale ripping off other people's code, then things might turn really uncomfortable for McB and his cronies...
Do I smell a Court Order in the works? If you really can do this without divulging the original code, then someone could conceivably convince a judge to issue an order to have a "neutral third party" create the MD5 sums on SCO's codebase, giving us a chance to look for pirated GPL code hidden inside of SCO proprietary products, without having to look directly at the SCO code.
What am I missing?
(The linked articles didn't give any hints either.)
Actually yes. The complaint is specifically that other entities, such as Red Hat, Novell and the FSF are acting as (paid) proxies for IBM, and that, "they have Eric Raymond on their payroll."
SCO is still trying to smear IBM in the Court of Public Opinion, to try and pressure them into settling outside of a real court, where they know they will be out-gunned, out-financed and backed up by flimsy arguments and scant facts.
I agree with you that IBM is probably pulling some strings, especially in the media, but as Eric said, "Those attacks have been happening because our community is outraged. IBM didn't have to talk to us or suborn us or bribe us or anything else."
Well, this is exactly why the process of making laws is so difficult. It'd be nice if we didn't need any laws, and people would just act nice! The reality is that people don't act nice, and powerful people beat up on or take advantage of less powerful people. In theory, laws are written to protect us from the violent or dishonest. But, it's impossible to write laws that always apply all the time, and can never be misinterpreted.
So, that's why we have both the Legislative branch (i.e., people writing the laws), and the Courts, who interpret the laws. Complain all you want, but the court system is designed to take imperfect laws and apply them in a case by case basis. Sure, there is no way, by just looking at the legislation, to tell the difference between a scum-of-the-earth spammer and a legitimate free speech case, but we have judges and juries and yes, even lawyers to try to apply some kind on intelligence to the laws and apply them in a just and fair manner.
Not sure if that makes you feel any better, but at least that's the theory...
Unfortunately, I don't know a good solution for the 60 hour part, but I know that lights (BIG lights) help the riding at night part. When I was commuting in the Winter, I did a lot of riding at night. I did pretty well, though you are right about keeping your eyes (and ears) open at all times. Sometimes you have to pick your routes carefully, and wear the most gawd-awful bright clothes. Wear anything that makes you stick out like a sore thumb.
Actually, riding at dusk is probably worse than riding in full dark, because the lights stand out more when it's dark. So, you might do better waiting a little longer (after rush hour) and then getting the biggest lights you can afford.
(Funny part is, people always thought I was insane for riding in the Winter, but I always found it one of the best times to ride. The simple fact that there's this loony out there in the snow, riding his or her bicycle raises the "weirdness factor" to the point where people actually notice you, and generally give you lots of room as they pass. Just look out for the snowplows...)