Beyond the "no money" implication, there's the issue that "free software" implies that the software is free, rather than that the user is free. I don't know about you, but I want the software I use to behave exactly as I tell it, and not demand vacations, the right to work for someone else, a fair wage, and so forth ("Free speech" is an idiom, which is why there's no alternative to "speech" in "free as in speech"). In fact, my experience with open source software is that it is more firmly under my control than closed source. I've seen a translation in which the translator took "Logiciel libre" to mean public domain software, lacking any other idea as to how "logiciel" could be "libre".
The term "free software", taken literally with the intended meaning of "free", means "liberated software", when "liberating software" is what people really mean. In fact, I think that upper management might like to use "liberating software", since they seem to feel trapped by their proprietary software.
Most of those games weren't that great 15 years ago, too. Zelda was good, and they liked it. SF2 was good, and they liked it about as much as any fighting game. Defender's got too many buttons, 720's too hard to control, Galaga's just like a bunch of similar games, etc.
I was expecting them to dismiss the old games based on the dated graphics, but they seem to have actually given each game a fair shot and enjoyed the games or found them annoying just like we did back then.
You're right to fault the communication rather than the documentation. The 2.6.4 config help for CONFIG_USB_MOUSE (which gives usbmouse) says "If even remotely unsure, say N." The help for CONFIG_USB_HID says "If unsure, say Y." Once you say what the problem is, it's trivial to find why it's wrong, but finding the solution in the first place is difficult (especially if the symptom is "basic mouse functionality works, but not more"). It's particularly difficult to realize that you're doing something wrong when the answer used to be that Linux didn't support the extra features.
As for improving the communication, it would be nice if information could be used in checking over installers out of the kernel documentation. Since people tend to say M for everything possible, it would be really nice if there were a way to get from a module name to the help for it, and also get the suggestion automatically. Then you could take a module list and say, "Is there anything discouraged that I'm loading? Is there anything encouraged that I'm not loading?"
The estimates are all based on different amounts of data. As you perform more measurements, the values will change significantly, especially for the first few measurements. If you consider the odds they show when they're televising poker games, they vary substantially. The early odds aren't any less reliable, but there's a lot of information left to get.
Assuming the astronomers can figure things out reliably this far in advance, one way to thing about it is that a 1/45 chance of getting hit means that there was a 44/45 chance of the odds dropping to very small when they determined pretty much for sure where it was going. Now there's a 1/56,000 chance that tomorrow it'll be 1/1, which is obviously much less likely.
Really, they ought to report the estimated distance it will pass from earth and the standard deviation in their estimate, which would give a much clearer picture of the difference between 1/56K based on very little data and 1/56K based on a lot of data.
Actually, we don't want to destroy the rock; we want to divert it. If we destroy it, we'll just get hit by lots of smaller pieces (which will burn up more due to the greater surface area, but not by enough to really matter). On the other hand, pushing it over a bit while it's pretty far away will cause it to miss us entirely (assuming it would have hit us otherwise). Attach a rocket and push it gently aside, like you're pushing a car.
I think these days, the backbones are primarily using IPv6, tunnelling IPv4 traffic over it, because IPv6 has advantages for routing and backbone providers don't have to deal with end users whose software isn't capable of dealing with IPv6. On the other hand, the backbones don't have a lot of non-router machines on them, so CERNET2 might be the biggest in terms of users using IPv6 directly.
You might want to send in the strace of cdrecord not working for you to the mailing list; all of the known things from the 2.6.8 breakage have been fixed, so it will probably never work on your hardware unless you report it.
As anyone who's played a racing car game in the past twenty years can tell you, when you pass a checkpoint, you get more time. As long as they keep completing new areas, they'll keep getting bonuses. On the other hand, if the Spirit team doesn't get moving, they're going to have to put in another quarter pretty soon.
I would guess that people searching for educational things wouldn't all search for the same educational things, and would search for particular educational things. People will search for "games" when they want to play something and don't care what, but people tend to care what they learn.
Re:The GPL/LGPL worries me....
on
Revising the GPL
·
· Score: 3, Informative
There is nothing in the GPL 2.0 which allows future versions to be used. The FSF (in the non-license portion of the GPL document) suggests licensing your work under "the GPL v2.0 or any later version published by the FSF", but there's no reason you have to do this. In particular, the Linux kernel is mostly under only the GPL v2.0 (portions of it are available under other licenses as well). Linus did it this way primarily because of the concerns you raise.
The LGPL can be explained without the term "linking": you can distribute a non-(L)GPL binary so long as it is possible to replace any LGPL portions without needing to do anything that the recipient can't do. If anything, it is static linking which will disappear, making the LGPL easier to follow (if you change the portions you got from somewhere else, you have to release these changes).
When some company manages to claim a patent on a feature of any software, regardless of the license, it becomes illegal to distribute it. The GPL is not special in this respect. Patents are not an issue for the GPL, which is a copyright license. They are, however, an issue for Free Software, because there's another entity which might restrict your freedom, and it's a thornier issue, because the holder of a patent is less likely to have needed to agree to a license than the holder of a copyright (since copyrights are often on derived works of some sort).
I've personally been at a relatively small company where we got a ~$100,000 Oracle database after one Oracle salesman showed up to talk to the developers for an afternoon. If you have a whole lot of data going through your database (millions of rows added/day), you want a big server, and Oracle is the right thing to make good use of it. Any company that's a retail chain or providing services to one is going to have this order of data, which is plenty of business for Oracle.
I think Oracle is right in saying that the competition helps them get new sales, because the availability of MySQL means that database apps for more purposes get written at the low end, and then big sites realize that they could use those apps (or more scalable versions of them) with an Oracle database. Or sites start out with OSS databases and then the load overwhelms the hardware that the database can use effectively.
I think the hot drink cup I used recently had a patented material, a copyrighted paint job, a trademarked logo, and contained a trade secret beverage. Software is unusual in that the aspects covered by copyright are important to its function, so it is easier to conflate the techniques by which the software works with the particular implementation in use.
There's not really anything to suggest that Solaris is going to perform in the market any differently in the near future than it's done in the past. They'd had x86 before, and it wasn't popular; the open source thing will only matter if someone forks it (or Sun discontinues it). They are continuing to add new features, but that isn't surprising, and anything that people actually find useful can be added to Linux (or, rather, an equivalent can be added).
Solaris certainly could become a major competitor for Linux at some point, but there would surely be some signs that it's getting somewhere first. For now, it's probably wise for Linus to ignore Solaris, and assume that someone will tell him if there's a reason to look at it.
Considering that it just openned in Japan, they probably haven't done any work on an English version yet. American distributors are annoyingly slow about signing anime, even Miyazaki's work, and it would be too much of a risk to actually do English voices without any particular plans to release it in an English version. On the other hand, I bet the English script will be relatively true to the original this time.
Actually, Access is two things, neither of which is the same as, for example, MySQL. The first is a user-friendly front end to a database. The second is a program to manipulate data files, which, unlike in most SQL databases, are self-contained and can be transferred to other people without any particular hassle. OOo 2.0 actually includes one of each (making them one program is a dumb idea): the GUI is the program they're talking about in the article, and the back-end is hsqldb, which is a respectable independant project.
The point of the second thing is to be like how people often use spreadsheets; you can have a lot of similar ones in different files, rather than a single big one as part of your infrastructure. The advantage over a spreadsheet is that you can actually keep it consistant and organized.
Re:GTK != Grand Theft Potassium
on
GTK 2.6.0 Released
·
· Score: 2, Funny
The "you die if you fall into water" property got merged into GTA, so the potassium version didn't matter much.
Her first thing is actually not at all about the miniseries. It's about what the director of the miniseries says about the book. The director of a movie shouldn't claim to be true to the book without having the author of the book tell him what the book is about. It's okay for a movie not to follow the book, but the director shouldn't claim it follows the book if it doesn't. Peter Jackson didn't claim that the Scouring of the Shire wasn't in the book, or that all of the songs from the books appeared in the movies, because he was aware that these things were different.
The second piece is a complaint about the miniseries, but it's one that didn't need to be the way it was. It's possible and not even particularly expensive to make characters have any skin color you want. It is ironic that Sci Fi has shows with blue people, but the red-brown people in this are white. If they couldn't use an actor for Ged with red-brown skin, they could at least have painted him the right color.
What's the point of photographing the people who buy stamps? It's not like, when a stamp is used to commit a crime, you can track it back to the photo by serial number. Unless...
Anyone taken a very close look at a stamp recently?
The page I'm looking at right now (slashdot's "lite" comment editor) doesn't have fixed-width tables, but has a banner ad wider than my window (I actually like to be able to have two browser windows side-by-side on my screen, and I find it easier to read reasonably narrow columns such that I don't have to track my eyes horizontally all that far in a line). When you're looking at an image in Firefox, it defaults to scaling it to fit in the window, and you can click on it to make it full-size. I'd like something similar with HTML pages, except that it only does the width and it doesn't shrink the fonts (unless, perhaps, it really couldn't fit a word from each column of a table, in which case it will need to scale the text; but then I'd like to see the overview and zoom in).
I think that browsers should scale all pixel values by whatever factor is needed to make the page fit in the window, rather than actually obeying them.
He says that the number one task that people use a web broser for is reading. Sure, it's obvious, but people seem to forget about it. In fact, he says this and then doesn't say anything about reading in the rest of his essay.
The biggest feature I want out of a web browser is that it not need horizontal scroll bars on any normal page regardless of the window size. It's okay if I'm looking at a large image or a large real table (think consumer reports) and I have clicked something to tell it that I want to look at the details.
Actually, 5 academics wrote an uberlint, tested it on the kernel (because the kernel has a lot of code and active developers who can tell them if it's generating false positives), reported a lot of real bugs, and are doing a startup with the tool. The tool is a static analyzer that makes pessimal assumptions about the data, so it finds the data-dependant bugs, plus theoretical "bugs" which are fine, but only for non-trivial reasons. For some of these cases, the reasons it works were added to the tool; for others, the code was made more explicit (because someone might change the code such that the reason doesn't work any more).
The bugs they found were entirely the sort of thing that simple static checkers don't find, and normal input doesn't trigger. It wouldn't be too surprising to have a lot of these in a project that basically worked. (I bet the GNU utilities have more than 30/KSLOC, given how easy they are to crash with weird input)
I'd call that a bug. Either the code is something that can never run, even if someone changes a constant or enables it, in which case it shouldn't be there, or the code is intended to be run in some situation which is presently impossible but would be needed if some other changes were made, in which case it should work correctly.
In fact, the body of an "if (0)" is generally considered reachable, unlike "if (1 > 2)", because people only write the former if they sometimes change the 0. Java, which generally prohibits unreachable code, specifically permits "if (false)" for this reason.
The case where you have to allow this sort of thing is:
int a = 1; int b;
while (1) {
if (a == 2)
return b;
b = 16;
a = 2; }
In this case, every line is actually used, but only in safe orders. If there's something more complicated going on in the loop (such as looking for a pair of sequence numbers which don't match the previous set but match each other), this code structure could actually be justified. A good static checker should be able to identify that the only paths that reach the return also set b first.
A good static checker will also catch the bug you found, because it tracks the limits of the possible values of the variables, so having the bug occur rarely doesn't matter. Of course, if the report were to give Thanksgiving instead of Thursday on Thanksgiving, it would then want one byte more memory than is actually needed, because it wouldn't necessarily know that Thanksgiving can never go in the string with the longest month name. Of course, that's probably just as well, because this arrangement isn't going to be immediately obvious to the next programmer, either.
I have a different version of top, but I'm sort of suspicious of the "VIRT" column. There's something kind of weird if only 57% of the non-shared size of your program is resident, there aren't any big inactive parts, and the program isn't completely bogged down swapping. I'd believe the 92M, with libraries being 35M of that (which may or may not count, depending on whether you have othe GTK+ programs you'd be using anyway). I have 38M resident, which probably reflects a smaller cache in memory.
Considering the clear mandate against software patents, shouldn't the EU Parliament be able to go on the offensive against software patents? Pass a bill making it fraud to charge someone with infringing a patent with software. Criminalize and prosecute ownership of the existing software patents and any future ones, legally grants or otherwise. It's very difficult in a complex government system to prevent something from being passed by well-funded and sneaky groups, but it's also difficult to prevent a majority with a mandate from getting something else passed.
Slashdot Mirror
Beyond the "no money" implication, there's the issue that "free software" implies that the software is free, rather than that the user is free. I don't know about you, but I want the software I use to behave exactly as I tell it, and not demand vacations, the right to work for someone else, a fair wage, and so forth ("Free speech" is an idiom, which is why there's no alternative to "speech" in "free as in speech"). In fact, my experience with open source software is that it is more firmly under my control than closed source. I've seen a translation in which the translator took "Logiciel libre" to mean public domain software, lacking any other idea as to how "logiciel" could be "libre".
The term "free software", taken literally with the intended meaning of "free", means "liberated software", when "liberating software" is what people really mean. In fact, I think that upper management might like to use "liberating software", since they seem to feel trapped by their proprietary software.
Most of those games weren't that great 15 years ago, too. Zelda was good, and they liked it. SF2 was good, and they liked it about as much as any fighting game. Defender's got too many buttons, 720's too hard to control, Galaga's just like a bunch of similar games, etc.
I was expecting them to dismiss the old games based on the dated graphics, but they seem to have actually given each game a fair shot and enjoyed the games or found them annoying just like we did back then.
You're right to fault the communication rather than the documentation. The 2.6.4 config help for CONFIG_USB_MOUSE (which gives usbmouse) says "If even remotely unsure, say N." The help for CONFIG_USB_HID says "If unsure, say Y." Once you say what the problem is, it's trivial to find why it's wrong, but finding the solution in the first place is difficult (especially if the symptom is "basic mouse functionality works, but not more"). It's particularly difficult to realize that you're doing something wrong when the answer used to be that Linux didn't support the extra features.
As for improving the communication, it would be nice if information could be used in checking over installers out of the kernel documentation. Since people tend to say M for everything possible, it would be really nice if there were a way to get from a module name to the help for it, and also get the suggestion automatically. Then you could take a module list and say, "Is there anything discouraged that I'm loading? Is there anything encouraged that I'm not loading?"
The estimates are all based on different amounts of data. As you perform more measurements, the values will change significantly, especially for the first few measurements. If you consider the odds they show when they're televising poker games, they vary substantially. The early odds aren't any less reliable, but there's a lot of information left to get.
Assuming the astronomers can figure things out reliably this far in advance, one way to thing about it is that a 1/45 chance of getting hit means that there was a 44/45 chance of the odds dropping to very small when they determined pretty much for sure where it was going. Now there's a 1/56,000 chance that tomorrow it'll be 1/1, which is obviously much less likely.
Really, they ought to report the estimated distance it will pass from earth and the standard deviation in their estimate, which would give a much clearer picture of the difference between 1/56K based on very little data and 1/56K based on a lot of data.
Actually, we don't want to destroy the rock; we want to divert it. If we destroy it, we'll just get hit by lots of smaller pieces (which will burn up more due to the greater surface area, but not by enough to really matter). On the other hand, pushing it over a bit while it's pretty far away will cause it to miss us entirely (assuming it would have hit us otherwise). Attach a rocket and push it gently aside, like you're pushing a car.
I think these days, the backbones are primarily using IPv6, tunnelling IPv4 traffic over it, because IPv6 has advantages for routing and backbone providers don't have to deal with end users whose software isn't capable of dealing with IPv6. On the other hand, the backbones don't have a lot of non-router machines on them, so CERNET2 might be the biggest in terms of users using IPv6 directly.
You might want to send in the strace of cdrecord not working for you to the mailing list; all of the known things from the 2.6.8 breakage have been fixed, so it will probably never work on your hardware unless you report it.
As anyone who's played a racing car game in the past twenty years can tell you, when you pass a checkpoint, you get more time. As long as they keep completing new areas, they'll keep getting bonuses. On the other hand, if the Spirit team doesn't get moving, they're going to have to put in another quarter pretty soon.
I would guess that people searching for educational things wouldn't all search for the same educational things, and would search for particular educational things. People will search for "games" when they want to play something and don't care what, but people tend to care what they learn.
There is nothing in the GPL 2.0 which allows future versions to be used. The FSF (in the non-license portion of the GPL document) suggests licensing your work under "the GPL v2.0 or any later version published by the FSF", but there's no reason you have to do this. In particular, the Linux kernel is mostly under only the GPL v2.0 (portions of it are available under other licenses as well). Linus did it this way primarily because of the concerns you raise.
The LGPL can be explained without the term "linking": you can distribute a non-(L)GPL binary so long as it is possible to replace any LGPL portions without needing to do anything that the recipient can't do. If anything, it is static linking which will disappear, making the LGPL easier to follow (if you change the portions you got from somewhere else, you have to release these changes).
When some company manages to claim a patent on a feature of any software, regardless of the license, it becomes illegal to distribute it. The GPL is not special in this respect. Patents are not an issue for the GPL, which is a copyright license. They are, however, an issue for Free Software, because there's another entity which might restrict your freedom, and it's a thornier issue, because the holder of a patent is less likely to have needed to agree to a license than the holder of a copyright (since copyrights are often on derived works of some sort).
Hopefully, there will be a jogger in the corner of some shot wearing spherical shoes, just for those of us who know the description from the book.
I've personally been at a relatively small company where we got a ~$100,000 Oracle database after one Oracle salesman showed up to talk to the developers for an afternoon. If you have a whole lot of data going through your database (millions of rows added/day), you want a big server, and Oracle is the right thing to make good use of it. Any company that's a retail chain or providing services to one is going to have this order of data, which is plenty of business for Oracle.
I think Oracle is right in saying that the competition helps them get new sales, because the availability of MySQL means that database apps for more purposes get written at the low end, and then big sites realize that they could use those apps (or more scalable versions of them) with an Oracle database. Or sites start out with OSS databases and then the load overwhelms the hardware that the database can use effectively.
I think the hot drink cup I used recently had a patented material, a copyrighted paint job, a trademarked logo, and contained a trade secret beverage. Software is unusual in that the aspects covered by copyright are important to its function, so it is easier to conflate the techniques by which the software works with the particular implementation in use.
There's not really anything to suggest that Solaris is going to perform in the market any differently in the near future than it's done in the past. They'd had x86 before, and it wasn't popular; the open source thing will only matter if someone forks it (or Sun discontinues it). They are continuing to add new features, but that isn't surprising, and anything that people actually find useful can be added to Linux (or, rather, an equivalent can be added).
Solaris certainly could become a major competitor for Linux at some point, but there would surely be some signs that it's getting somewhere first. For now, it's probably wise for Linus to ignore Solaris, and assume that someone will tell him if there's a reason to look at it.
Considering that it just openned in Japan, they probably haven't done any work on an English version yet. American distributors are annoyingly slow about signing anime, even Miyazaki's work, and it would be too much of a risk to actually do English voices without any particular plans to release it in an English version. On the other hand, I bet the English script will be relatively true to the original this time.
Actually, Access is two things, neither of which is the same as, for example, MySQL. The first is a user-friendly front end to a database. The second is a program to manipulate data files, which, unlike in most SQL databases, are self-contained and can be transferred to other people without any particular hassle. OOo 2.0 actually includes one of each (making them one program is a dumb idea): the GUI is the program they're talking about in the article, and the back-end is hsqldb, which is a respectable independant project.
The point of the second thing is to be like how people often use spreadsheets; you can have a lot of similar ones in different files, rather than a single big one as part of your infrastructure. The advantage over a spreadsheet is that you can actually keep it consistant and organized.
The "you die if you fall into water" property got merged into GTA, so the potassium version didn't matter much.
Her first thing is actually not at all about the miniseries. It's about what the director of the miniseries says about the book. The director of a movie shouldn't claim to be true to the book without having the author of the book tell him what the book is about. It's okay for a movie not to follow the book, but the director shouldn't claim it follows the book if it doesn't. Peter Jackson didn't claim that the Scouring of the Shire wasn't in the book, or that all of the songs from the books appeared in the movies, because he was aware that these things were different.
The second piece is a complaint about the miniseries, but it's one that didn't need to be the way it was. It's possible and not even particularly expensive to make characters have any skin color you want. It is ironic that Sci Fi has shows with blue people, but the red-brown people in this are white. If they couldn't use an actor for Ged with red-brown skin, they could at least have painted him the right color.
What's the point of photographing the people who buy stamps? It's not like, when a stamp is used to commit a crime, you can track it back to the photo by serial number. Unless...
Anyone taken a very close look at a stamp recently?
The page I'm looking at right now (slashdot's "lite" comment editor) doesn't have fixed-width tables, but has a banner ad wider than my window (I actually like to be able to have two browser windows side-by-side on my screen, and I find it easier to read reasonably narrow columns such that I don't have to track my eyes horizontally all that far in a line). When you're looking at an image in Firefox, it defaults to scaling it to fit in the window, and you can click on it to make it full-size. I'd like something similar with HTML pages, except that it only does the width and it doesn't shrink the fonts (unless, perhaps, it really couldn't fit a word from each column of a table, in which case it will need to scale the text; but then I'd like to see the overview and zoom in).
I think that browsers should scale all pixel values by whatever factor is needed to make the page fit in the window, rather than actually obeying them.
He says that the number one task that people use a web broser for is reading. Sure, it's obvious, but people seem to forget about it. In fact, he says this and then doesn't say anything about reading in the rest of his essay.
The biggest feature I want out of a web browser is that it not need horizontal scroll bars on any normal page regardless of the window size. It's okay if I'm looking at a large image or a large real table (think consumer reports) and I have clicked something to tell it that I want to look at the details.
Actually, 5 academics wrote an uberlint, tested it on the kernel (because the kernel has a lot of code and active developers who can tell them if it's generating false positives), reported a lot of real bugs, and are doing a startup with the tool. The tool is a static analyzer that makes pessimal assumptions about the data, so it finds the data-dependant bugs, plus theoretical "bugs" which are fine, but only for non-trivial reasons. For some of these cases, the reasons it works were added to the tool; for others, the code was made more explicit (because someone might change the code such that the reason doesn't work any more).
The bugs they found were entirely the sort of thing that simple static checkers don't find, and normal input doesn't trigger. It wouldn't be too surprising to have a lot of these in a project that basically worked. (I bet the GNU utilities have more than 30/KSLOC, given how easy they are to crash with weird input)
In fact, the body of an "if (0)" is generally considered reachable, unlike "if (1 > 2)", because people only write the former if they sometimes change the 0. Java, which generally prohibits unreachable code, specifically permits "if (false)" for this reason.
The case where you have to allow this sort of thing is:In this case, every line is actually used, but only in safe orders. If there's something more complicated going on in the loop (such as looking for a pair of sequence numbers which don't match the previous set but match each other), this code structure could actually be justified. A good static checker should be able to identify that the only paths that reach the return also set b first.
A good static checker will also catch the bug you found, because it tracks the limits of the possible values of the variables, so having the bug occur rarely doesn't matter. Of course, if the report were to give Thanksgiving instead of Thursday on Thanksgiving, it would then want one byte more memory than is actually needed, because it wouldn't necessarily know that Thanksgiving can never go in the string with the longest month name. Of course, that's probably just as well, because this arrangement isn't going to be immediately obvious to the next programmer, either.
I have a different version of top, but I'm sort of suspicious of the "VIRT" column. There's something kind of weird if only 57% of the non-shared size of your program is resident, there aren't any big inactive parts, and the program isn't completely bogged down swapping. I'd believe the 92M, with libraries being 35M of that (which may or may not count, depending on whether you have othe GTK+ programs you'd be using anyway). I have 38M resident, which probably reflects a smaller cache in memory.
Considering the clear mandate against software patents, shouldn't the EU Parliament be able to go on the offensive against software patents? Pass a bill making it fraud to charge someone with infringing a patent with software. Criminalize and prosecute ownership of the existing software patents and any future ones, legally grants or otherwise. It's very difficult in a complex government system to prevent something from being passed by well-funded and sneaky groups, but it's also difficult to prevent a majority with a mandate from getting something else passed.