how by being open source going to save millions of non technical user privacy ?
Easy: their geeky friends would tell them "Don't download Kazaa: it's dangerous for you. Instead, download OpenKazaa: it's faster, safer, and still interoperates with the original version."
In different words, non-technical users benefit because a few technical users would quickly make a version without the security problems, something that can't be done for closed source code.
Where is the shielding???
on
Tool Box PC
·
· Score: 2
It's a nice case, but where is the shielding? This thing looks like an unshielded plastic box. PCs need shielding--without them, they interfere strongly with amateur radio, emergency services, and medical devices.
(Note that you can get something similar in the lunchbox format (e.g., here and many other places).
Sony is large enough to hire the smart monopoly economists (yes, there is such a thing) to figure out for them how to maximize profit. And you can bet that this is the point at which they are doing so.
TIFF has a tile format as well. Is there something better about the JPEG2000 or PhotoCD formats over the TIFF one? Especially given that TIFF tool support is quite broad.
Both JPEG2000 and PhotoCD are not merely tiled, they are hierarchical: you can extract a complete, scaled down view of the image without every looking at the high resolution version. You can also make some global adjustments (color, cropping, etc.) without touching every pixel. And for viewing detail, only small portions need to be decompressed.
From what I read a few months ago, adding CMYK support to the GIMP, while desirable for the whole graphic arts industry, would be heroic amounts of work.
It doesn't look that hard to me. The GIMP already has support for arbitrary numbers of channels. CMYK is just a four-channel image. All you would need to do is write code that loads and saves it, and some code that converts between CMYK and RGB on the fly for display and RGB-based filters.
I think the main reason this hasn't happened is because nobody who understands color can figure out why anybody would want to do such a thing, or even what "viewing or editing CMYK" is even supposed to mean. Just convert to RGB and use RGB tools. Conversions to CMYK are best left to the device drivers that actually do the printing. What the Gimp needs much more than CMYK is 16bit support, IMO.
Well, the Gimp also has the problem for us of requiring lots of work to get it running on HP-UX, and we would also need something for Windows. The gimp's web page does not inspire confidence in their Windows stability--or are they just being modest?
I have had no problems with the Gimp on Windows, but it is a slightly older version. Why not just download it and give it a try? As for HP-UX, "configure" generally works pretty well, but HP-UX is also pretty weird.
Your best bets for high-end graphics (unless you want to pay an arm and a leg) are MacOSX, Solaris, and (for free or research software) Linux.
What makes you say that? A correctly encrypted password management system makes security better because it allows you to choose lots of different good passwords for different accounts.
Even if you can control the logins on the major operating systems, your users will still encounter other passwords everywhere. I think rather than trying to control the uncontrollable, a better solution is to get them Palm Pilots with encrypting password managers.
I don't think I would need a script to click 100 times. And the thought that clicking on a banner ad 100 times might get you hauled into court on federal computer tampering charges is pretty chilling.
I still don't see the fraud. What kind of legal contract am I under that I could defraud anyone? I didn't agree to anything just by going to the web page. It's a public web page. If someone is foolish to make a contract that pays $4/click to someone else whenever someone clicks on their web page, well, they deserve what they get. I don't have a contract with them.
Some bad (i.e., obvious/software/business method) patents are on annoying inventions. That's a real dilemma. On the one hand, I don't like to see the USPTO grant bad patents. On the other hand, if patents on spam techniques, search result reordering, and other annoying marketing and business practices get enforced vigorously and hence get used less, our lives get easier. Can someone please patent animated GIFs and the BLINK tag?
While touch pads are pretty awful as pointing devices, if you want absolute quiet, they may be your best bet. Many of them can be configured so that when you gently tap on them, it counts as a click.
It's probably more useful to hook up a camera to a GPS system. That way, not only can you snap a picture every mile, you can also record where exactly it was snapped without having to make guesses.
Forget about the fact that Microsoft's attempts at restricting open source implementations are bogus. Let's ask: why even waste time on CIFS? Is enabling Windows clients to talk to non-Windows servers even the right thing to do? And if it is, why use CIFS? It seems to me that developing an installable file system driver that speaks a high-quality open protocol (NFS doesn't qualify) would result in something that's both easier to maintain and would work better than trying to reverse engineer Microsoft's latest protocol hiccup.
Microsoft can license their code in whatever way they like--the more restrictive, the better as far as I'm concerned.
This isn't about licensing code, however. It's an attempt by Microsoft to restrict what you can do with code you hold the copyright to based on having read a document that they have written. And that's not OK. The GPL is completely different--if you want to make a proprietary implementation that works like a piece of GPL'ed code, the GPL doesn't stop you.
It's difficult to see what legal basis these restrictions are supposed to have. Trademarks and patents don't apply. So, either it's a trade secret or a copyright issue. I think it's hard to argue that something published to half the world and already widely known is a "trade secret". And copyright defines in and of itself what you can and cannot do with a document (believe me, publishers would love to impose restrictions like "if you read this, our license requires you to buy the two sequels as well").
Note that this is not analogous to the GPL. The GPL is a license that lets you do things with copyrighted software that copyright law prohibits you from doing unless you get a license. But that approach doesn't work for a protocol specification because the copyright is on the specification, not the protocol, and so you generally don't need a license in order to implement the protocol even if the specification is copyrighted. (It may be possible to copyright protocols, but that seems like a long shot for something like CIFS that's mostly open anyway.)
This license gaffe is curiously analogous to Microsoft software. Usually, Microsoft copies other people's software badly. This time, Microsoft copied other people's licenses (the GPL) badly, trying to impose onerous restrictions on people who merely read a document. Microsoft's incompetence manages to reach new lows day after day, and I suppose that is an amazing achievement in itself.
The short answer is: use a command line tool like "tifftopnm" to convert into RGB color space; it should do the conversion pixel-wise, so it doesn't need much memory at all. The resulting file will be about 400Mbytes uncompressed. Once you are in RGB space, many image viewers should work (ImageMagick probably won't--it's not very efficient). The Gimp in particular uses a tiling cache.
If you are going to do this sort of thing a lot, don't use TIFF. There are a few tiled formats around for GIS applications, and and hierarchical formats like JPEG2000 or PhotoCD should also be a reasonable choice (but not much tools support yet).
Of course, rather than writing your own, if you really need CMYK, why not add CMYK support to the Gimp? The Gimp already does the tiling and has lots of other useful features.
100 years from now, people will be looking back at Bill Gates and say the same thing: look, he ran a gigantic research and development lab, creating many of the inventions we take for granted today, and being responsible for the creation of thousands of patents. Never mind that almost all the technology Microsoft puts out was invented elsewhere.
Of course, Gates is not Edison, but think about how today's events are going to look in the future. That may give you a bit of a better idea of what to think of the past.
This is true. However, with Qt (in general), you can develop something ONCE for numerous platforms: Linux/Unix, Linux-PDA (Zaurus/iPAQ), Windows, MacOS.
That's a separate issue. If the Sharp ran X11 as its primary window system, you could still run Qt on it, but you could also use many other toolkits. The problem with the Sharp is not its use of Qt for its applications, it's its use of the Qt/Embedded drawing engine instead of X11. The choice of Qt/Embedded excludes other toolkits without giving users any additional functionality.
In fact, Sharp can still fix that: move from Qt/Embedded to Qt/X11.
The iMac also runs Ogg Vorbis if you like, together with your favorite encoders and players. People have been talking about an Ogg Vorbis plug-in for iTunes, but I don't know whether anybody has bothered adapting it.
Why you would want to use Ogg isn't clear to me. Obviously, the iMac comes with an MP3 license. At the bitrates where you would want to encode home audio, I don't think there is a big difference between the two. And the MP3 patents are going to expire sooner or later--what matters is that the MP3 format is open and well-documented, and has numerous open source implementations.
The consumer market is pretty much covered by Palm and PocketPC. Yes, those operating systems are pretty limited, but they have lots of software available for it, and mass market developers worry about making stuff work on them.
Where Sharp could have been a winner is in vertical markets: special purpose apps for markets like finance, medicine, and research. But their choice of Qt/Embedded for the GUI killed that--many vertical apps have already been developed, either two existing X11 toolkits or to Win32 APIs. Both of those can be supported (and share the same screen) with X11. With Qt/Embedded, you have to develop everything again for that one toolkit (it doesn't even help if you get VNC or X11 running on the thing somehow because Qt/Embedded won't integrate with it). And, besides, what developer wants to spend a lot of time becoming proficient at yet another toolkit?
Overall, an iPaq running Linux is probably still your best bet for a handheld Linux machine. You can even dual-boot them with WinCE now (just like your desktop--isn't that fun).
What is so annoying about efforts like the Zaurus is that their poor business decisions and marketing give all of Linux a bad reputation.
Here is a possible scenario. Watermarking only works if everybody is using it and everybody is detecting it--otherwise, people will just end up using the non-watermarked codecs. The two companies will use patented technology for watermarking. They will then go the MPAA and RIAA and similar cartels and unite with them to pressure Congress to adopt their watermarking scheme. The end result is tidy for them: DivX and Fraunhofer get complete control over codecs, and copyright holders can completely control who does what with any content, whether it's their or independently created. And Microsoft will likely like it too, because they can afford to license such mandated technology and enforce its inclusion in their software.
Excluded are open source software developers, researchers, and independent creators of content.
They cost around $1000 (the CRT-based one), have a slot-loading CD drive, and come with a nice color screen. I think the Graphite one looks pretty good--nicer than a big, black box. And iTunes is very convenient for both capturing and playing back MP3s. If you want a remote control, there are a couple of companies offering USB remote controls. And, of course, it has network interfaces, USB, and FireWire.
Of course, this dedicated box may be a little cheaper, but then it also isn't quite as versatile.
Over the last few hundred years, the writing plume has displaced the chisel and stone tablet. This is unfortunate, because chisels and stones are really much more natural than writing plumes. Plumes were really a concession to portability, but the documents people author with them are ephemeral. Writing plumes also deprive writers of the pleasant feel of hammering away at a phrase, relieving stress and giving people exercise.
New Titanium Chisel Technology (TCT) promises to change all that. Using ultra-modern, flexible "stone" tablets made out of composites, together with titanium chisels, people can now enjoy the convenience and pleasure of writing using the human-centered technologies while still enjoying portability close to that of writing plumes and parchment.
Titanium Chisel Corporation is a fast-growing, high-technology, privately held startup, located in Granite Valley, Nevada, where most of the fast-growing, exciting stone and stone-age technology companies of the world are located.
Re:Valgrind and memory leaks
on
KDE 3.0 is Out
·
· Score: 2
Oh, I have seen plenty of memory leaks in my C++ programs--from other people's libraries, often C libraries. That's why I have to look. But I haven't found any in my code.
Re:Valgrind and memory leaks
on
KDE 3.0 is Out
·
· Score: 2
Will a few 100kloc do?
C++ is very different from C in this regard because it has constructors, destructors, access control, user-defined assignment, and exceptions. If you use them right, it's easy to avoid memory leaks. If you write C code in C++ (i.e., invoke new/delete all over the place), of course you'll get memory leaks.
Slashdot Mirror
Easy: their geeky friends would tell them "Don't download Kazaa: it's dangerous for you. Instead, download OpenKazaa: it's faster, safer, and still interoperates with the original version."
In different words, non-technical users benefit because a few technical users would quickly make a version without the security problems, something that can't be done for closed source code.
(Note that you can get something similar in the lunchbox format (e.g., here and many other places).
Sony is large enough to hire the smart monopoly economists (yes, there is such a thing) to figure out for them how to maximize profit. And you can bet that this is the point at which they are doing so.
Both JPEG2000 and PhotoCD are not merely tiled, they are hierarchical: you can extract a complete, scaled down view of the image without every looking at the high resolution version. You can also make some global adjustments (color, cropping, etc.) without touching every pixel. And for viewing detail, only small portions need to be decompressed.
From what I read a few months ago, adding CMYK support to the GIMP, while desirable for the whole graphic arts industry, would be heroic amounts of work.
It doesn't look that hard to me. The GIMP already has support for arbitrary numbers of channels. CMYK is just a four-channel image. All you would need to do is write code that loads and saves it, and some code that converts between CMYK and RGB on the fly for display and RGB-based filters.
I think the main reason this hasn't happened is because nobody who understands color can figure out why anybody would want to do such a thing, or even what "viewing or editing CMYK" is even supposed to mean. Just convert to RGB and use RGB tools. Conversions to CMYK are best left to the device drivers that actually do the printing. What the Gimp needs much more than CMYK is 16bit support, IMO.
Well, the Gimp also has the problem for us of requiring lots of work to get it running on HP-UX, and we would also need something for Windows. The gimp's web page does not inspire confidence in their Windows stability--or are they just being modest?
I have had no problems with the Gimp on Windows, but it is a slightly older version. Why not just download it and give it a try? As for HP-UX, "configure" generally works pretty well, but HP-UX is also pretty weird.
Your best bets for high-end graphics (unless you want to pay an arm and a leg) are MacOSX, Solaris, and (for free or research software) Linux.
What makes you say that? A correctly encrypted password management system makes security better because it allows you to choose lots of different good passwords for different accounts.
Even if you can control the logins on the major operating systems, your users will still encounter other passwords everywhere. I think rather than trying to control the uncontrollable, a better solution is to get them Palm Pilots with encrypting password managers.
I don't think I would need a script to click 100 times. And the thought that clicking on a banner ad 100 times might get you hauled into court on federal computer tampering charges is pretty chilling.
I still don't see the fraud. What kind of legal contract am I under that I could defraud anyone? I didn't agree to anything just by going to the web page. It's a public web page. If someone is foolish to make a contract that pays $4/click to someone else whenever someone clicks on their web page, well, they deserve what they get. I don't have a contract with them.
Why would clicking the same gambling link 100 times be considered "fraud"? And what is algorithmically non-trivial about detecting that?
Some bad (i.e., obvious/software/business method) patents are on annoying inventions. That's a real dilemma. On the one hand, I don't like to see the USPTO grant bad patents. On the other hand, if patents on spam techniques, search result reordering, and other annoying marketing and business practices get enforced vigorously and hence get used less, our lives get easier. Can someone please patent animated GIFs and the BLINK tag?
While touch pads are pretty awful as pointing devices, if you want absolute quiet, they may be your best bet. Many of them can be configured so that when you gently tap on them, it counts as a click.
It's probably more useful to hook up a camera to a GPS system. That way, not only can you snap a picture every mile, you can also record where exactly it was snapped without having to make guesses.
Forget about the fact that Microsoft's attempts at restricting open source implementations are bogus. Let's ask: why even waste time on CIFS? Is enabling Windows clients to talk to non-Windows servers even the right thing to do? And if it is, why use CIFS? It seems to me that developing an installable file system driver that speaks a high-quality open protocol (NFS doesn't qualify) would result in something that's both easier to maintain and would work better than trying to reverse engineer Microsoft's latest protocol hiccup.
This isn't about licensing code, however. It's an attempt by Microsoft to restrict what you can do with code you hold the copyright to based on having read a document that they have written. And that's not OK. The GPL is completely different--if you want to make a proprietary implementation that works like a piece of GPL'ed code, the GPL doesn't stop you.
Note that this is not analogous to the GPL. The GPL is a license that lets you do things with copyrighted software that copyright law prohibits you from doing unless you get a license. But that approach doesn't work for a protocol specification because the copyright is on the specification, not the protocol, and so you generally don't need a license in order to implement the protocol even if the specification is copyrighted. (It may be possible to copyright protocols, but that seems like a long shot for something like CIFS that's mostly open anyway.)
This license gaffe is curiously analogous to Microsoft software. Usually, Microsoft copies other people's software badly. This time, Microsoft copied other people's licenses (the GPL) badly, trying to impose onerous restrictions on people who merely read a document. Microsoft's incompetence manages to reach new lows day after day, and I suppose that is an amazing achievement in itself.
If you are going to do this sort of thing a lot, don't use TIFF. There are a few tiled formats around for GIS applications, and and hierarchical formats like JPEG2000 or PhotoCD should also be a reasonable choice (but not much tools support yet).
Of course, rather than writing your own, if you really need CMYK, why not add CMYK support to the Gimp? The Gimp already does the tiling and has lots of other useful features.
Of course, Gates is not Edison, but think about how today's events are going to look in the future. That may give you a bit of a better idea of what to think of the past.
That's a separate issue. If the Sharp ran X11 as its primary window system, you could still run Qt on it, but you could also use many other toolkits. The problem with the Sharp is not its use of Qt for its applications, it's its use of the Qt/Embedded drawing engine instead of X11. The choice of Qt/Embedded excludes other toolkits without giving users any additional functionality.
In fact, Sharp can still fix that: move from Qt/Embedded to Qt/X11.
Why you would want to use Ogg isn't clear to me. Obviously, the iMac comes with an MP3 license. At the bitrates where you would want to encode home audio, I don't think there is a big difference between the two. And the MP3 patents are going to expire sooner or later--what matters is that the MP3 format is open and well-documented, and has numerous open source implementations.
Where Sharp could have been a winner is in vertical markets: special purpose apps for markets like finance, medicine, and research. But their choice of Qt/Embedded for the GUI killed that--many vertical apps have already been developed, either two existing X11 toolkits or to Win32 APIs. Both of those can be supported (and share the same screen) with X11. With Qt/Embedded, you have to develop everything again for that one toolkit (it doesn't even help if you get VNC or X11 running on the thing somehow because Qt/Embedded won't integrate with it). And, besides, what developer wants to spend a lot of time becoming proficient at yet another toolkit?
Overall, an iPaq running Linux is probably still your best bet for a handheld Linux machine. You can even dual-boot them with WinCE now (just like your desktop--isn't that fun).
What is so annoying about efforts like the Zaurus is that their poor business decisions and marketing give all of Linux a bad reputation.
Excluded are open source software developers, researchers, and independent creators of content.
Of course, this dedicated box may be a little cheaper, but then it also isn't quite as versatile.
New Titanium Chisel Technology (TCT) promises to change all that. Using ultra-modern, flexible "stone" tablets made out of composites, together with titanium chisels, people can now enjoy the convenience and pleasure of writing using the human-centered technologies while still enjoying portability close to that of writing plumes and parchment.
Titanium Chisel Corporation is a fast-growing, high-technology, privately held startup, located in Granite Valley, Nevada, where most of the fast-growing, exciting stone and stone-age technology companies of the world are located.
Oh, I have seen plenty of memory leaks in my C++ programs--from other people's libraries, often C libraries. That's why I have to look. But I haven't found any in my code.
C++ is very different from C in this regard because it has constructors, destructors, access control, user-defined assignment, and exceptions. If you use them right, it's easy to avoid memory leaks. If you write C code in C++ (i.e., invoke new/delete all over the place), of course you'll get memory leaks.