I clearly remember when the Pentium (original 60-MHZ version) came out, that was the big selling point was the capability of watching videos on it. In fact, I've got a CD I picked up back then that had the Beatles movie A Hard Days Night on it, and it played fine on my old 486.
One additional thing. This can be implemented using the existing system, by an individual bank if they desire. Since the account number begins with a bunch of zeros, the hash (or a fingerprint of it at least) can be stored where those zeros are. So the first 10 digits become the hash fingerprint, and the remaining digits are the actual account number.
Actually, all you need is to have the check number, account & routing number hashed with a secret key known only by your bank & check printing company. Have this unique hash (unique to each check number) appear on the bottom of each check, and the bank can refuse to clear it if it doesn't match up, or if a check number is duplicated. You end up with the same paper check everyone is used to, but with a bit more security. The only way fraud can happen is if someone steels your blank checks, or duplicates an uncashed check and changes the amount.
You can go a step further and include a maximum dollar amount with the hash, and have separate checkbooks for small vs. large amounts
In my case, I have an interpreter for a small language I wrote, which is currently released it under the GPL. Now part of it's design is that it can be used as an embedded language attached to other programs (in this case, one would link to a library implementing my language). So in this instance I would like for any Free/Open Source app (as defined by the OSI or FSF) to be able to link against my code. But for the time being I want to reserve rights when it comes to non-Free software.
What about when you want your software to be combined with other Free / Open Source software (even code that may be incompatible with the GPL), but don't want to leave the door wide open to non-Free software? Would the GPL-with-exceptions license work?
That is something I never understood. Normally, two different companies can hold a trademark on a common word if they are in different industries, which is why we can have a Linux OS and Linux Soap. So how could Apple Records have any standing to sue Apple Computers? Of course, they didn't "win" the suit, as it was settled out of court, but still...
Ok it looks like the bigger boot.iso image (129-meg) contains an "install.img" file. With a smaller boot.iso image (that you create by using the "making an installation boot cd" instructions), you get prompted for the server containing the Fedora installation files, in which you would feed "http://download.fedoraproject.org/pub/fedora/linux/releases/10/Fedora/i386" (or your own local server if you mirrored the install directory locally). The large boot.iso (which is the same as "fedora-10-i386-netinst.iso") is hard-coded to go after a random public Fedora mirror, so it is not suitable for working off a private installation copy.
The boot.iso image I mentioned was a minimal one (5 meg) on the Red Hat Enterprise linux distros (rhel4 and rhel5). But it looks like it is a 129-meg file on Fedora 10. This matches the "fedora-10-i386-netinst.iso" image found in the regular iso directory. I haven't looked at this yet, but based on the name it should do the same net install. If not, you can generate a small boot iso by grabbing the files from "http://download.fedoraproject.org/pub/fedora/linux/releases/10/Fedora/i386/os/isolinux/" directory and following the "making an installation boot cd" section of the Red Hat Enterprise installation guide (available on redhat.com -- support -- documentation). This will create a 2-meg iso image.
If you want, I'll try both and report back the differences.
Actually you can kind of do that with Fedora. Instead of downloading the full ISO, just download the 5-meg ISO image which contains just the anaconda installer. Then select a fedora mirror as your install server, and choose the package groups that you need as they are presented to you.
Only thing is they don't make it obvious where to grab the small ISO image. You have to browse through the the repository a bit. For example: "/pub/fedora/linux/releases/10/Fedora/i386/os/images/boot.iso"
I would rather see a home and business versions, with expansion packs on top of that. So you can start with home basic, then add on ultimate if you find that you need it (otherwise you have to choose up front, not knowing if the extra features on ultimate will be needed).
Another way of looking at it is: RAM is manufactured in units based on powers of 2, since anything else would cause issues for programmers (at least ASM programmers, back in the early days). And it may cause issues with other hardware (CPU, etc) which can have optimizations for powers of 2 (i.e. bit shifts instead of multiply & divide). So rounding off the power-of-2-based numbers in RAM to the closest matching SI units is a convenience for people (who wants to say "536.87 MB ram" instead of "512 MB").
But hard drives are not manufactured in power-of-2 units. Instead, they are sector-cylinder-head units (at least originally). Only the sector size is a power-of-2 unit, everything else is arbitrary, which leads to a hard drive size not exactly fitting either power-of-2 or -10 units. Even the old 5-meg hard drives weren't exactly 5*2^20 or 5*10^6.
On the other hand, hard drives are used to store data that is held in RAM. So in a way, if the amount of RAM taken up is (for convenience) specified in base-2 units, then it would make sense to specify the hard drive space used in the same base. Which is why it makes sense for the file manager to report all sizes in base-2.
Would one of the UFS variations be suitable for flash drives? And also better portability (almost all OS's support UFS by now). Would performance be better?
It's not so much that IE is integrated with windows, but instead Windows ships with a bunch of toolkits, one of which is a toolkit containing library routines for Web access. This includes establishing HTTP connections, doing the low level HTTP get, and a rendering engine to do something with HTML documents.
So this toolkit is used as a primary component of IE. Parts of it is also used by various other OS components, such as Windows Update (uses the HTTP libraries), the File manager (HTML rendering engine), etc. And, obviously, it is the toolkit that is used to build IE. So what Microsoft means by "removing IE will break Windows" is removing IE and it's associated librarys/toolkit will break the other components. But the part of IE that contains "main()" could be removed without affecting anything else.
But now we have another problem. Is a web browser a stand alone application, or is it a necessary part of a modern OS, same as a file manager and command shell? This gets down to the basic debate of what an OS is. Here's my definition: OS Kernel -- the low level component that connects applications to hardware devices (device drivers), and defines / maintains data structures on those devices (think "file system" layer). OS Utilities -- programs that allow a user to manipulate data structures the Kernel maintains (such as a file manager), and programs that facilitate user interaction with the hardware the kernel interfaces to (such as a utility to talk to a modem, or send a file to a printer port).
So an operating system is composed of the Kernel and OS Utilities. An Operating Environment (OE) is a combination of an OS with a set of applications that facilitate performing tasks that nearly all users of that computer would need to do. So text editors and paint programs fall in this category (although a text editor may straddle this category and OS Utilities).
Now the question is, where does an ftp client, telnet, ssh, etc. fit in? And does a web browser fit the same category as ftp?
This may be a better deal now, but a few years ago the wireless carriers were constantly coming out with better plans than their competitors, not to mention that the selection of handsets from the carriers differed significantly. It was a good day when my Nextel contract was up several years ago (just after Sprint bought them out).
why would you have to extend? Just buy a replacement phone at full price. Personally, I don't find it worth it to save $100 or so by committing to spend 24 x $90/month.
Instead of having a hard cutover date, have the stations reduce their wattage on the analog signals over the course of several months. That way, instead of people loosing their signal, they will get a slightly degraded picture one month... then a slightly more degraded picture the next month, until it gets to a point where they realize that they need the digital converter box.
Both Sharp (with the Zaurus line) and Motorola (with their EZX phone platform) used modified QT (actually, QTE, QT Embedded). Not sure what their reasons were though.
I can agree on the bindings issue. I've started writing gtk+ bindings for my small scripting language (lang2e.sourceforge.net), and all that was required was for me to add a callback mechanism (that is, a way to pass a pointer function written in the scripting language to a gtk function). But so far I haven't figured out a clean way of adding qt or kde bindings. Especially since my language doesn't have much in the way of object-oriented features yet.
Does anyone have any good pointers on writing language bindings for qt?
There's a few ad- and parasite-blocking hosts files out there -- they contain a list of known ad servers and redirect them to localhost. They were initially targeting malicious servers, but now include most ad servers too. Only problem is you get a bunch of error messages in the place of the ads, but you can for the most part leave flash turned on.
I can top that. I new someone once who only knew Lotus 123 (OK it was back in the 90's). She would send out text memos written out as a series of spreadsheet text cells.
Then the same applies to free software. Very few free software programmers would put out anything if they knew up front that no one would contribute code back, or ideas, or any sign of appreciation. Most free software is simply from coders scratching each others backs.
My biggest issue with video conferencing is that no matter where you place the camera, if you are looking at the person on the screen, to them it looks like you are looking away from them. Put the camera on top of the screen, and your eyes are pointing down from the camera which makes it look like you are lying. Put the camera to the side, and it looks like your thoughts are wandering.
If someone can come up with a camera embedded in the middle of the screen, that would be awesome.
Slashdot Mirror
I clearly remember when the Pentium (original 60-MHZ version) came out, that was the big selling point was the capability of watching videos on it. In fact, I've got a CD I picked up back then that had the Beatles movie A Hard Days Night on it, and it played fine on my old 486.
One additional thing. This can be implemented using the existing system, by an individual bank if they desire. Since the account number begins with a bunch of zeros, the hash (or a fingerprint of it at least) can be stored where those zeros are. So the first 10 digits become the hash fingerprint, and the remaining digits are the actual account number.
Actually, all you need is to have the check number, account & routing number hashed with a secret key known only by your bank & check printing company. Have this unique hash (unique to each check number) appear on the bottom of each check, and the bank can refuse to clear it if it doesn't match up, or if a check number is duplicated. You end up with the same paper check everyone is used to, but with a bit more security. The only way fraud can happen is if someone steels your blank checks, or duplicates an uncashed check and changes the amount.
You can go a step further and include a maximum dollar amount with the hash, and have separate checkbooks for small vs. large amounts
In my case, I have an interpreter for a small language I wrote, which is currently released it under the GPL. Now part of it's design is that it can be used as an embedded language attached to other programs (in this case, one would link to a library implementing my language). So in this instance I would like for any Free/Open Source app (as defined by the OSI or FSF) to be able to link against my code. But for the time being I want to reserve rights when it comes to non-Free software.
What about when you want your software to be combined with other Free / Open Source software (even code that may be incompatible with the GPL), but don't want to leave the door wide open to non-Free software? Would the GPL-with-exceptions license work?
That is something I never understood. Normally, two different companies can hold a trademark on a common word if they are in different industries, which is why we can have a Linux OS and Linux Soap. So how could Apple Records have any standing to sue Apple Computers? Of course, they didn't "win" the suit, as it was settled out of court, but still...
Ok it looks like the bigger boot.iso image (129-meg) contains an "install.img" file. With a smaller boot.iso image (that you create by using the "making an installation boot cd" instructions), you get prompted for the server containing the Fedora installation files, in which you would feed "http://download.fedoraproject.org/pub/fedora/linux/releases/10/Fedora/i386" (or your own local server if you mirrored the install directory locally). The large boot.iso (which is the same as "fedora-10-i386-netinst.iso") is hard-coded to go after a random public Fedora mirror, so it is not suitable for working off a private installation copy.
The boot.iso image I mentioned was a minimal one (5 meg) on the Red Hat Enterprise linux distros (rhel4 and rhel5). But it looks like it is a 129-meg file on Fedora 10. This matches the "fedora-10-i386-netinst.iso" image found in the regular iso directory. I haven't looked at this yet, but based on the name it should do the same net install. If not, you can generate a small boot iso by grabbing the files from "http://download.fedoraproject.org/pub/fedora/linux/releases/10/Fedora/i386/os/isolinux/" directory and following the "making an installation boot cd" section of the Red Hat Enterprise installation guide (available on redhat.com -- support -- documentation). This will create a 2-meg iso image.
If you want, I'll try both and report back the differences.
Actually you can kind of do that with Fedora. Instead of downloading the full ISO, just download the 5-meg ISO image which contains just the anaconda installer. Then select a fedora mirror as your install server, and choose the package groups that you need as they are presented to you.
Only thing is they don't make it obvious where to grab the small ISO image. You have to browse through the the repository a bit. For example: "/pub/fedora/linux/releases/10/Fedora/i386/os/images/boot.iso"
I would rather see a home and business versions, with expansion packs on top of that. So you can start with home basic, then add on ultimate if you find that you need it (otherwise you have to choose up front, not knowing if the extra features on ultimate will be needed).
Actually I meant UDF.
Another way of looking at it is:
RAM is manufactured in units based on powers of 2, since anything else would cause issues for programmers (at least ASM programmers, back in the early days). And it may cause issues with other hardware (CPU, etc) which can have optimizations for powers of 2 (i.e. bit shifts instead of multiply & divide). So rounding off the power-of-2-based numbers in RAM to the closest matching SI units is a convenience for people (who wants to say "536.87 MB ram" instead of "512 MB").
But hard drives are not manufactured in power-of-2 units. Instead, they are sector-cylinder-head units (at least originally). Only the sector size is a power-of-2 unit, everything else is arbitrary, which leads to a hard drive size not exactly fitting either power-of-2 or -10 units. Even the old 5-meg hard drives weren't exactly 5*2^20 or 5*10^6.
On the other hand, hard drives are used to store data that is held in RAM. So in a way, if the amount of RAM taken up is (for convenience) specified in base-2 units, then it would make sense to specify the hard drive space used in the same base. Which is why it makes sense for the file manager to report all sizes in base-2.
Would one of the UFS variations be suitable for flash drives? And also better portability (almost all OS's support UFS by now). Would performance be better?
It's not so much that IE is integrated with windows, but instead Windows ships with a bunch of toolkits, one of which is a toolkit containing library routines for Web access. This includes establishing HTTP connections, doing the low level HTTP get, and a rendering engine to do something with HTML documents.
So this toolkit is used as a primary component of IE. Parts of it is also used by various other OS components, such as Windows Update (uses the HTTP libraries), the File manager (HTML rendering engine), etc. And, obviously, it is the toolkit that is used to build IE. So what Microsoft means by "removing IE will break Windows" is removing IE and it's associated librarys/toolkit will break the other components. But the part of IE that contains "main()" could be removed without affecting anything else.
But now we have another problem. Is a web browser a stand alone application, or is it a necessary part of a modern OS, same as a file manager and command shell? This gets down to the basic debate of what an OS is. Here's my definition:
OS Kernel -- the low level component that connects applications to hardware devices (device drivers), and defines / maintains data structures on those devices (think "file system" layer).
OS Utilities -- programs that allow a user to manipulate data structures the Kernel maintains (such as a file manager), and programs that facilitate user interaction with the hardware the kernel interfaces to (such as a utility to talk to a modem, or send a file to a printer port).
So an operating system is composed of the Kernel and OS Utilities. An Operating Environment (OE) is a combination of an OS with a set of applications that facilitate performing tasks that nearly all users of that computer would need to do. So text editors and paint programs fall in this category (although a text editor may straddle this category and OS Utilities).
Now the question is, where does an ftp client, telnet, ssh, etc. fit in? And does a web browser fit the same category as ftp?
This may be a better deal now, but a few years ago the wireless carriers were constantly coming out with better plans than their competitors, not to mention that the selection of handsets from the carriers differed significantly. It was a good day when my Nextel contract was up several years ago (just after Sprint bought them out).
why would you have to extend? Just buy a replacement phone at full price. Personally, I don't find it worth it to save $100 or so by committing to spend 24 x $90/month.
Instead of having a hard cutover date, have the stations reduce their wattage on the analog signals over the course of several months. That way, instead of people loosing their signal, they will get a slightly degraded picture one month... then a slightly more degraded picture the next month, until it gets to a point where they realize that they need the digital converter box.
Both Sharp (with the Zaurus line) and Motorola (with their EZX phone platform) used modified QT (actually, QTE, QT Embedded). Not sure what their reasons were though.
I can agree on the bindings issue. I've started writing gtk+ bindings for my small scripting language (lang2e.sourceforge.net), and all that was required was for me to add a callback mechanism (that is, a way to pass a pointer function written in the scripting language to a gtk function). But so far I haven't figured out a clean way of adding qt or kde bindings. Especially since my language doesn't have much in the way of object-oriented features yet.
Does anyone have any good pointers on writing language bindings for qt?
There's a few ad- and parasite-blocking hosts files out there -- they contain a list of known ad servers and redirect them to localhost. They were initially targeting malicious servers, but now include most ad servers too. Only problem is you get a bunch of error messages in the place of the ads, but you can for the most part leave flash turned on.
Reminds me of Sun's Sunfire demo http://www.youtube.com/watch?v=cmM4o17UvLU
And there was a point in time when wireless bluetooth headsets looked dorky. They still do, however they have become somewhat accepted.
I can top that. I new someone once who only knew Lotus 123 (OK it was back in the 90's). She would send out text memos written out as a series of spreadsheet text cells.
Then the same applies to free software. Very few free software programmers would put out anything if they knew up front that no one would contribute code back, or ideas, or any sign of appreciation. Most free software is simply from coders scratching each others backs.
My biggest issue with video conferencing is that no matter where you place the camera, if you are looking at the person on the screen, to them it looks like you are looking away from them. Put the camera on top of the screen, and your eyes are pointing down from the camera which makes it look like you are lying. Put the camera to the side, and it looks like your thoughts are wandering.
If someone can come up with a camera embedded in the middle of the screen, that would be awesome.