anything in consumer space is only standards complient if it uses the Win32 API
On the x86 PC platform that may be true, but that is not what we're talking about here. Consoles (and set-top boxes, and networked `appliances' or whatever name they have tomorrow) are still OS-neutral, as long as it does the job it is OK. So now what counts is which OS does the job in a satisfying manner both for the user as well as for the designer, manufacturer and software developer. Look at Netpliance's i-Opener (QNX) for an example of a device which is totally `consumer space' oriented, yet lacks even a single bit of Win32 software. The hardware platform supports it, but it is not used, instead they chose QNX. Why? Probably for reasons of size (small), stability (good) and maybe licensing fees (no idea really). Linux can also be made small, it can be very stable in the right configuration, and licensing fees are a non-issue.
So i'd say there Win32 (or Winwhatever) is not the de-facto API for these applications...
Well, maybe they plan to target their console at those who wish to do a little more with it than 'just play games'. If so, using a standards compliant operating systems for which a whole body of good, free software already exists makes a lot of sense. So using Linux (or *BSD) might not be such a bad idea...
Re:How about a nice ROM Monitor instead?Seriously,
on
Linux BIOS
·
· Score: 1
Seriously, this BIOS stuff has got to go. Nobody uses 16-bit operating systems anymore; why are we still booting 32-bit (and soon 64-bit) operating systems using a 16-bit BIOS?
Well, if you're using an x86 family processor, the moment you're turning it on you're actually running a 16-bit processor. The processor has to be explicitly told to behave like a 32-bit processor (or `switched to protected mode' if you prefer). This is so you can run your ancient segmented:memory_model software on that same box.
One simple answer: vote with your wallet. Don't buy DVD.
Now before you all put me on the griddle and call me a Luddite, figure this: nobody forces you to buy DVD-players, discs, etc. They `only' outlaw every other means of distribution using their armada of lawyers and lobbyists. Now of course this implies that you will not be able to access recent `entertainment' products. Personally I couldn't care less, but if you're into that kind of stuff (watching product placement, being submitted to all kinds of subversive `buy me' tricks, etc.) you've got a problem. This problem can be solved if you are rich (by buying the gadgets mentioned in this article, or getting your own lawyer to tee off with the MPAA's lawyers, etc), clever (have a peek at the miserable excuse they use for content obfuscation, reverse engineer it, use it) or from a jurisdiction which does not give a rats ass about US law (by just stealing and copying the lot).
But I still think the best solution is to give up on the instant-spam-in-a-can entertainment from those shiny platters and get a book, go out, meet people, whatever. And if you *really* want to see that latest movie, go to the cinema. You'll get all the widescreen and panscanless imagery you can swallow...
Is this a vulnerability in the Active technology? No. This vulnerability results because of a manual error in marking the particular control at issue.
Manual error? But why then does the "Show Me" function need to be disabled to negate this threat? Or was this entire funcionality the result of a "manual marking error"? Or might it be that ActiveX does not offer fine-grained control over who is allowed to do what to which data? In other words, a "design problem" with ActiveX?
Hmmm... Maybe you North-Americans are getting a bit *too* attached to your creature comforts?
Airco is not nearly as widely spread here in Europe, and still people live, work and play just as you do. Sometimes it gets hot, yes, but then you just open a window or dress lightly. Or use a fan instead of a full-flegded air conditioner. Saves energy and does not cause respiratory diseases...
If you mean `on the short term', the BSD license is easier for corps. They can just grab what they want, stuff it in their product and sell sell sell. But by doing that they miss out on one of the major benefits of free software, the `free support' that often comes with it. By keeping their software closed, they *will* fall behind other companies who opened their stuff to the world. If the closed-source-corp has some piece of whiz-bang hardware which runs circles around the competition, they will probably succeed anyway since the benefits of the superior hardware offset the drawbacks of the less reliable driver support for many buyers. If, however, the hardware is on par with that offered by the competition, they'd better make sure their software staff keeps up with everything out there and fixes bugs faster than you can spot them, or they will lose.
I chose a Matrox card for this reason. Not because it Quakes better or worse than nVidia (I have no idea if it does, I'm not a Quaker). but because I know I can get it running on my box no matter which OS or kernel I run. Same thing for the other cards in there.
Allright, they manage to retrofit Otto's invention with some fancy gizmo's to squeeze some more miles out of liquified dinosaurs and trees. Good for them, but this is not what the world (except for those with interests in oil companies) is waiting for I'd say.
How about getting serious with fuel cells, hydrogen storage and distribution, efficient hydrogen production (not much sense in using an environmentally friendly fuel if that fuel itself is produced using an environmentally unfriendly process)? Several companies here in Europe (Daimler-Benz, Volvo, Fiat) have had fuel-cell powered vehicles in development for quite some time now. The technology seems to work, the range is good, the performance adequate.
IF this is true, I still don't understand how Microsoft thinks they have any business releasing software with Internet functionality anymore. Intranet, sure. Internet? No way.
Erm, might I remind you that the vast majority of security breaches come from the INSIDE of the network/company/megacorp? This is why the concept of perimeter defence utterly fails to protect resources against those most willing and able to exploit flaws (and this is a security flaw, no matter whether is was put there intentionally or not). So Microsoft (or any other producer of shoddy wares) has not more business releasing software for use on the internal network than they have on the perimeter or outside of it.
Of course, they in fact generate a lot of business doing exactly that, but that is another story.
I never said it before and I won't repeat it, but SMP certainly DOES make sense depending on your application(s). If you `only' play games or MSOffice your way through the day, it will probably not help very much. If you (like me) use your box to compile largish projects on a regular basis, SMP will either give you a shorter build cycle (using the -jn (2 = n = younameit) flag for GNU make) or give you a responsive box while the compile is chugging along. Add in some extra memory and you're set. Of course, the viability of this hinges on the availability of a reasonably priced SMP board (enter the ABit BP-6), processors (add your favourite Celeron, possibly the new FC-PGA processors from Intel using a convertor or a motherboard modification - only for HW hackers though) and an OS which knows how to use the extra processor (Linux, various flavours of BSD, BeOS, WinNT/2000 (but the latter still suffers from various driver problems in SMP mode)).
Oh, and if you're so inclined, you can also overclock this setup. I run it at default speed using lower CPU voltage, your choice...
So, SMP is not (yet) for the masses, but it certainly is a viable and economically sound choice for some.
Re:Interesting side effects?
on
G3 Solar Storm
·
· Score: 1
Yes, there was aurora in the Netherlands. And how! The entire northern part of the sky was at times ablaze with red flares and streaks, the northern horizon shrouded in a gauzy band of greenish light. This was observed from the top of my house on the outskirts of Lelystad. It lasted for something like 45 minutes and was a spectacle to behold. According to the magnetometers the storm is subsiding, so I doubt whether tonight will be as spectacular, but who knows?
ld-linux.so is the shared library loader, part of the Gnu C library. Are you sure it is Enlightenment which triggers this `bug' and not RedHat's version of Netscape Communicator/Navigator? The latter is usually started through a script which calls the library loader directly to set up an environment for Netscape. This way, the library loader shows up in the process list eating a lot of memory. I had the exact same `problem' on my 256 MB box, and `solved' it by using my own version of Netscape. Which still hogs memory as all Netscape's do, but at least it does not preload half the contents of/usr/lib... Cheers//Frank
The shareware concept - try before you buy - has had limited success in the software market. I wonder, maybe a similar concept in music distribution can be used as well. People will still copy music without paying the fee, but this will probably always happen (even in a fully digitized and encrypted world where the data is only decrypted once it reached the output device).
There is a lot of people who are willing to pay a reasonable price for the music they want to listen to, as long as it is clear that the revenues go to the artists instead of the usual suspects (the `fat guy in a Lexus' types who seem to populate the major labels). Total revenues would not be as large as they are now, but it would not be the artists who feel this in their pockets.
Comments? Would this work, or is it an Utopian view which will never come to pass because people prefer a free lunch?
The current (7.0) license for real's software explicitly forbids the use of shared objects outside real's own applications. They go as far as to name the DLL's and libraries in the license. So this is currently out of the question...
Well, I've tried my Maxtor 40G drives on the UDMA66 (HPT366) controller, and was surprised by the fact that they achieved around 25 M/s sustained performance in the hdparm disk read test. On the UDMA33 (PIIX4) controller this drops to around 13 M/s. The hdparm test is far from perfect, but is gives some indication of IO performance. The problem with those Maxtor drives on the HPT366 (the UDMA66 controller used on the Abit boards) is not lack of performance, but lack of stability. For some reason the system locks up under sustained disk load, which is rather inconvenient and counterproductive. So I put my drives on the `slower' UDMA33 controller, thereby negating the need for this distribution...
Now that is an intriguing idea. I think we can trust the distributed.net folks, but it *would* be rather easy for them to use their clienst for other purposes besides the usual contest-cracking. The clients are distributed in binary form, so it is not easy to determine what they're doing. The data the clients fetch from the server may be part of a cracking contest data block. Then again, it may also be part of whatever encrypted data block they want to know the contents of.
[Trust noone. Keep your laser handy. The computer is your friend.]
Oh, and the authors of the `original' Jurassic Park 3D-filetoy (called fsn, vavailable at SGI but only for IRIX 5.x, get it ) are working on something similar (but functional this time), which will be open source AFAIK. This was posted on {linuxtoday|freshmeat|?} some time ago, and I don't have the URL handy. A search should turn up something I guess?
Gun Laws? Who wants guns? Better live in a place where you don't need the damn things in the first place (like the country I live in - the Netherlands) and be done with it. Now if only we had those mountains and parks and space it'd be bliss... Oh yeah, and maybe somewhat less than 50%-60% income tax... Any Canadians out there hiring Unix guru's? Got Unix, will travel...
This is odd... Scroll down the page, and look at the `Linux Kernel Usage' table. Why are FreeBSD, OpenBSD and Solaris listed in this table? And what impact does this have on the `maximum uptime' score for Linux?
The only illegal thing done here is to have reverse-engineered a poorly-written software decoder to extract a key. However, it would also have been possible to brute-force test keys until one was found, although it would have taken a while.
And even that is not illegal in most countries, AFAIK. So, if no illegal acts were committed, what ground do these lawyerks have to stand on? I say let them eat their suit and suits and begone forever. It will be a better world without them.
Operating system stability is NOT dependent on application stability. If your applications bring down your OS, there's something wrong with the OS. Sure, you can slow things down to a crawl (intentional or otherwise) even on a `correct' OS, but it should NOT be possible for user-level applications to bring down the OS. If it does, you've either found a bug and/or a design flaw. In the Windows 9[5|8] world it's mostly design flaws (DLL hell, etc) which cause the instability. In the Unix world most of those flaws have been ironed out by now. There are bugs, and there will probably will be bugs forever, but there are far fewer glaring design faults in Unix then there are in `consumer Windows'. Windows NT is probably somewhere in between, with a more solid foundation (probably more solid than the original Unix foundation) but a lot of `implementation issues' (euphemism for bugs) and `speed tradeoffs' (video not fast enough? Move it to the kernel!). WinCE is terra incognita for me in this respect, so I'll refrain from commenting on it.
And that forkomb example you give can be defused quite easily on modern Unixen. Process limits, resource limits and `fork bomb detector gizmos' (as recently seen on Freshmeat) take care of this.
AOL's counter should be interesting. I know some of the guys at AOL--hell, one of 'em is probably the single smartest networking guy I've ever met. We're talking about a company that, in response to Microsoft's attempts at circumventing their network security, consistently and repeatedly
exploited security holes in Microsoft's AIM client, and likely threatened to announce exploits for their client unless Microsoft caved in. (They did.)
BEEEEP error.... AOL exploited holes in their own client, not in Microsoft's client.
AFAIK, StarOffice is written in C++ to the StarView toolkit. The latter is Stardivision's own cross-platform toolkit which has enabled them to port the software to several platforms without too much trouble. Unleash objdump and nm on the StarOffice libraries to get some insight in the internals of StarView.
Slashdot Mirror
On the x86 PC platform that may be true, but that is not what we're talking about here. Consoles (and set-top boxes, and networked `appliances' or whatever name they have tomorrow) are still OS-neutral, as long as it does the job it is OK. So now what counts is which OS does the job in a satisfying manner both for the user as well as for the designer, manufacturer and software developer. Look at Netpliance's i-Opener (QNX) for an example of a device which is totally `consumer space' oriented, yet lacks even a single bit of Win32 software. The hardware platform supports it, but it is not used, instead they chose QNX. Why? Probably for reasons of size (small), stability (good) and maybe licensing fees (no idea really). Linux can also be made small, it can be very stable in the right configuration, and licensing fees are a non-issue.
So i'd say there Win32 (or Winwhatever) is not the de-facto API for these applications...
Well, maybe they plan to target their console at those who wish to do a little more with it than 'just play games'. If so, using a standards compliant operating systems for which a whole body of good, free software already exists makes a lot of sense. So using Linux (or *BSD) might not be such a bad idea...
Well, if you're using an x86 family processor, the moment you're turning it on you're actually running a 16-bit processor. The processor has to be explicitly told to behave like a 32-bit processor (or `switched to protected mode' if you prefer). This is so you can run your ancient segmented:memory_model software on that same box.
One simple answer: vote with your wallet. Don't buy DVD.
:-)
Now before you all put me on the griddle and call me a Luddite, figure this: nobody forces you to buy DVD-players, discs, etc. They `only' outlaw every other means of distribution using their armada of lawyers and lobbyists. Now of course this implies that you will not be able to access recent `entertainment' products. Personally I couldn't care less, but if you're into that kind of stuff (watching product placement, being submitted to all kinds of subversive `buy me' tricks, etc.) you've got a problem. This problem can be solved if you are rich (by buying the gadgets mentioned in this article, or getting your own lawyer to tee off with the MPAA's lawyers, etc), clever (have a peek at the miserable excuse they use for content obfuscation, reverse engineer it, use it) or from a jurisdiction which does not give a rats ass about US law (by just stealing and copying the lot).
But I still think the best solution is to give up on the instant-spam-in-a-can entertainment from those shiny platters and get a book, go out, meet people, whatever. And if you *really* want to see that latest movie, go to the cinema. You'll get all the widescreen and panscanless imagery you can swallow...
Now you may ignite your flamethrowers...
Microsoft states in their FAQ:
Is this a vulnerability in the Active
technology? No. This vulnerability results
because of a manual error in marking the
particular control at issue.
Manual error? But why then does the "Show Me" function need to be disabled to negate this threat? Or was this entire funcionality the result of a "manual marking error"? Or might it be that ActiveX does not offer fine-grained control over who is allowed to do what to which data? In other words, a "design problem" with ActiveX?
May I introduce George, a Gnome app written in Perl...
Hmmm... Maybe you North-Americans are getting a bit *too* attached to your creature comforts?
Airco is not nearly as widely spread here in Europe, and still people live, work and play just as you do. Sometimes it gets hot, yes, but then you just open a window or dress lightly. Or use a fan instead of a full-flegded air conditioner. Saves energy and does not cause respiratory diseases...
Cheers//Frank
[North-European without airco...]
Sure,
If you mean `on the short term', the BSD license is easier for corps. They can just grab what they want, stuff it in their product and sell sell sell. But by doing that they miss out on one of the major benefits of free software, the `free support' that often comes with it. By keeping their software closed, they *will* fall behind other companies who opened their stuff to the world. If the closed-source-corp has some piece of whiz-bang hardware which runs circles around the competition, they will probably succeed anyway since the benefits of the superior hardware offset the drawbacks of the less reliable driver support for many buyers. If, however, the hardware is on par with that offered by the competition, they'd better make sure their software staff keeps up with everything out there and fixes bugs faster than you can spot them, or they will lose.
I chose a Matrox card for this reason. Not because it Quakes better or worse than nVidia (I have no idea if it does, I'm not a Quaker). but because I know I can get it running on my box no matter which OS or kernel I run. Same thing for the other cards in there.
Allright, they manage to retrofit Otto's invention with some fancy gizmo's to squeeze some more miles out of liquified dinosaurs and trees. Good for them, but this is not what the world (except for those with interests in oil companies) is waiting for I'd say.
How about getting serious with fuel cells, hydrogen storage and distribution, efficient hydrogen production (not much sense in using an environmentally friendly fuel if that fuel itself is produced using an environmentally unfriendly process)? Several companies here in Europe (Daimler-Benz, Volvo, Fiat) have had fuel-cell powered vehicles in development for quite some time now. The technology seems to work, the range is good, the performance adequate.
Now THAT is what I call progress.
Erm, might I remind you that the vast majority of security breaches come from the INSIDE of the network/company/megacorp? This is why the concept of perimeter defence utterly fails to protect resources against those most willing and able to exploit flaws (and this is a security flaw, no matter whether is was put there intentionally or not). So Microsoft (or any other producer of shoddy wares) has not more business releasing software for use on the internal network than they have on the perimeter or outside of it.
Of course, they in fact generate a lot of business doing exactly that, but that is another story.
I never said it before and I won't repeat it, but SMP certainly DOES make sense depending on your application(s). If you `only' play games or MSOffice your way through the day, it will probably not help very much. If you (like me) use your box to compile largish projects on a regular basis, SMP will either give you a shorter build cycle (using the -jn (2 = n = younameit) flag for GNU make) or give you a responsive box while the compile is chugging along. Add in some extra memory and you're set. Of course, the viability of this hinges on the availability of a reasonably priced SMP board (enter the ABit BP-6), processors (add your favourite Celeron, possibly the new FC-PGA processors from Intel using a convertor or a motherboard modification - only for HW hackers though) and an OS which knows how to use the extra processor (Linux, various flavours of BSD, BeOS, WinNT/2000 (but the latter still suffers from various driver problems in SMP mode)).
Oh, and if you're so inclined, you can also overclock this setup. I run it at default speed using lower CPU voltage, your choice...
So, SMP is not (yet) for the masses, but it certainly is a viable and economically sound choice for some.
Yes, there was aurora in the Netherlands. And how! The entire northern part of the sky was at times ablaze with red flares and streaks, the northern horizon shrouded in a gauzy band of greenish light. This was observed from the top of my house on the outskirts of Lelystad. It lasted for something like 45 minutes and was a spectacle to behold. According to the magnetometers the storm is subsiding, so I doubt whether tonight will be as spectacular, but who knows?
[proudly posting from a fresh mozilla CVS build]
See? Netscape is available in the English, Japanese and Marketing language! :-)
ld-linux.so is the shared library loader, part of the Gnu C library. Are you sure it is Enlightenment which triggers this `bug' and not RedHat's version of Netscape Communicator/Navigator? The latter is usually started through a script which calls the library loader directly to set up an environment for Netscape. This way, the library loader shows up in the process list eating a lot of memory. I had the exact same `problem' on my 256 MB box, and `solved' it by using my own version of Netscape. Which still hogs memory as all Netscape's do, but at least it does not preload half the contents of /usr/lib... Cheers//Frank
The shareware concept - try before you buy - has had limited success in the software market. I wonder, maybe a similar concept in music distribution can be used as well. People will still copy music without paying the fee, but this will probably always happen (even in a fully digitized and encrypted world where the data is only decrypted once it reached the output device).
There is a lot of people who are willing to pay a reasonable price for the music they want to listen to, as long as it is clear that the revenues go to the artists instead of the usual suspects (the `fat guy in a Lexus' types who seem to populate the major labels). Total revenues would not be as large as they are now, but it would not be the artists who feel this in their pockets.
Comments? Would this work, or is it an Utopian view which will never come to pass because people prefer a free lunch?
The current (7.0) license for real's software explicitly forbids the use of shared objects outside real's own applications. They go as far as to name the DLL's and libraries in the license. So this is currently out of the question...
Well, I've tried my Maxtor 40G drives on the UDMA66 (HPT366) controller, and was surprised by the fact that they achieved around 25 M/s sustained performance in the hdparm disk read test. On the UDMA33 (PIIX4) controller this drops to around 13 M/s. The hdparm test is far from perfect, but is gives some indication of IO performance. The problem with those Maxtor drives on the HPT366 (the UDMA66 controller used on the Abit boards) is not lack of performance, but lack of stability. For some reason the system locks up under sustained disk load, which is rather inconvenient and counterproductive. So I put my drives on the `slower' UDMA33 controller, thereby negating the need for this distribution...
Now that is an intriguing idea. I think we can trust the distributed.net folks, but it *would* be rather easy for them to use their clienst for other purposes besides the usual contest-cracking. The clients are distributed in binary form, so it is not easy to determine what they're doing. The data the clients fetch from the server may be part of a cracking contest data block. Then again, it may also be part of whatever encrypted data block they want to know the contents of.
[Trust noone. Keep your laser handy. The computer is your friend.]
http://threedsia.sourceforge.net/
Oh, and the authors of the `original' Jurassic Park 3D-filetoy (called fsn, vavailable at SGI but only for IRIX 5.x, get it ) are working on something similar (but functional this time), which will be open source AFAIK. This was posted on {linuxtoday|freshmeat|?} some time ago, and I don't have the URL handy. A search should turn up something I guess?
Gun Laws? Who wants guns? Better live in a place where you don't need the damn things in the first place (like the country I live in - the Netherlands) and be done with it. Now if only we had those mountains and parks and space it'd be bliss... Oh yeah, and maybe somewhat less than 50%-60% income tax... Any Canadians out there hiring Unix guru's? Got Unix, will travel...
This is odd... Scroll down the page, and look at the `Linux Kernel Usage' table. Why are FreeBSD, OpenBSD and Solaris listed in this table? And what impact does this have on the `maximum uptime' score for Linux?
And even that is not illegal in most countries, AFAIK. So, if no illegal acts were committed, what ground do these lawyerks have to stand on? I say let them eat their suit and suits and begone forever. It will be a better world without them.
[phew... had to get it out of my system...]
Aaargh!
NO.
Operating system stability is NOT dependent on application stability. If your applications bring down your OS, there's something wrong with the OS. Sure, you can slow things down to a crawl (intentional or otherwise) even on a `correct' OS, but it should NOT be possible for user-level applications to bring down the OS. If it does, you've either found a bug and/or a design flaw. In the Windows 9[5|8] world it's mostly design flaws (DLL hell, etc) which cause the instability. In the Unix world most of those flaws have been ironed out by now. There are bugs, and there will probably will be bugs forever, but there are far fewer glaring design faults in Unix then there are in `consumer Windows'. Windows NT is probably somewhere in between, with a more solid foundation (probably more solid than the original Unix foundation) but a lot of `implementation issues' (euphemism for bugs) and `speed tradeoffs' (video not fast enough? Move it to the kernel!). WinCE is terra incognita for me in this respect, so I'll refrain from commenting on it.
And that forkomb example you give can be defused quite easily on modern Unixen. Process limits, resource limits and `fork bomb detector gizmos' (as recently seen on Freshmeat) take care of this.
Cheers//Frank
BEEEEP error.... AOL exploited holes in their own client, not in Microsoft's client.
FYI...
AFAIK, StarOffice is written in C++ to the StarView toolkit. The latter is Stardivision's own cross-platform toolkit which has enabled them to port the software to several platforms without too much trouble. Unleash objdump and nm on the StarOffice libraries to get some insight in the internals of StarView.