Anyone in touch with people building system component by component realizes a great majority of these are AMD based, since they get to see the processor price difference and know the price/performance problems with P4 vs. Athlon. I personally have been eagerly awaiting a version with AGP for AMD, and then I might buy in... I want to have an All-in-Wonder in one of these babies to integrate into my entertainment system...
IIRC, it not only strengthens the 'shrink wrap' licenses, but also sets a default liability. So just saying 'this comes with no warranty' in a file somewhere does not free the programmer of responsibilty. Commercial companies don't care as they are explicitly allowed to shirk this liability through their shrink-wrap license, but open source applications can't really have a shrink-wrap license, so no matter how much you deny liability, you are stuck with it..
If by 'pretty good' you mean 'total piece of shit', I agree. Most of the Visual Studio suite is pretty decent, but Source Safe comes off exactly as what it is, a software package bought out by MS for the express purpose of saying Visual Studio has SCM, and then doing next to nothing with it to make it work better. As an admin in a shop of people begging to leave Source Safe for something like Rational, or even CVS, I know how it is... Subversion may offer the core needed to build a replacement that keeps both developers and the businessmen happy...
I've been running gcc 3.1 compiled gentoo for a while now. Very nice and offers some tangible speedups. However, the costs:
Browser Plugins: Flash plugin required me to write a small compatibility library to mimick some the old libstdc++ mangled memory allocation schemes. This will probably not appear in a Mandrake desktop, as they will likely provide a -compat library without the user knowing. Never have gotten java plugins to work... Just haven't figured it out for blackdown, ibm's, nor suns.... Realplayer plugin problem same as Flash. Right now I am just lacking java...
Build: Some programs won't build out of the box. Some due to bad code, but mostly due to strange build configuration. For example, basiliskII's build fails at one point when gcc is used to link object files generated by g++ and bombs because some g++ symbols are unknown to gcc, switching that gcc to g++ makes that step go by... Others I've had issues with include PixiePlus, mame, and openoffice.
Others may have issues. I don't use crossover as vanilla wine fills all my needs, and I have nothing in the way of commercial software aside from games, which all *worked*, (every quake, civctp). The biggest problem I've had is again, c++ browser plugins...
I can't stand VS.net, I started using it and then tired of the crap and found VS6 much more straightforward. In either case I'm glad I rarely need Visual Studio, as most of my development is either *nix or cross-platform.
Complicates the cache issue? How? Basically, they are making two complete hard drive assemblies on a common assembly, but one can only is designed from the ground up to be read-only, from the head to the interface, providing the most possible levels of protection to get through. If the drive head is physically incapable of modifying data, then no bug in any software or firmware can be explotied to allow writes.
I personally think this is useless for other reasons, completely redundant with networked strategies to do the same thing, at least with competent administrators. However, I guess this thing would be good for places with less-than-stellar administration. Then, they could have the web server's only connection to the data through that safe harddrive cable. If an organization is just serving up static data, then they probably aren't sophisticated enough of an organization to afford stellar admin people, so I guess this thing has a place.... No where near me though.
Perhaps that is promising, but I'm still not sure what Helix is meant to encompass. Wading through the marketspeak doesn't yield anything definite. Is the Helix 'client' a full-featured player, or the client-end of a content delivery framework which feeds retrieved data into a binary-only plugin or codec? I'll wager the core media encodings will remain binary-only, even if/when a good deal of the platform is open. If stripped down to the point where only the codec itself is binary-only and the interfaces to that codec are open, that would be very nice too, but I doubt they can implement a pure, open platform for decoding all media content (Sorenson and Microsoft will allow their stuff to be open when hell freezes over), and since their competitors won't be allowing open codecs, there is little incentive for them to open up their codec...
I'll wager that even if the 'client' is a media player, a closed core will handle the decoding, but if the player doesn't suck or the interfaces to that core are open, I'll be content...
I'm glad they are making a cross-platform effort, but am not glad that their cross-platform strategy comes off as more of a peripheral 'as they feel like it' effort rather than a serious effort. The non-Windows players are hidden from www.real.com and lacking in development effort. RealOne for linux is in alpha, while it has been out for a long time with windows. I only care about this because RealOne will make use of the Xvideo extension. In general, Real's player has bad controls (in terms of keyboard particluarly) and I want an alternative player, or a centralized player that can play whatever I want, be it DVD, avi, mov, mpg, or rm, and Real's (as well as apple's) efforts aren't allowing this sort of solution to gel, particularly with non-windows,non-mac systems. MPlayer and xine and others are trying damn hard to acheive this, Real and Apple be damned, but with a more cooperative approach (as this Helix, on second glance, *might* represent), I could acheive my multimedia dream under linux.
I think his point is valid. Even if a cracker roots a server, he can't change stuff if a separate system dictates permission. Given proper firewalling rules and/or login configuration, the second system is just as untouchable as the host system in this multi-hard dirve configuration. The key difference would be performance (the two systems don't care what the other is doing), but with the typical bandwidth bottleneck, hard drive IO is the least of your concerns, particularly with static content.
I saw no mention of open sourcing the player, or even mention of anything that would necessarily do encoding... It seems the portion being open sourced would be transport and delivery, and even the license to that would be restrictive.
This is a move to appeal to providers starting to go on the fence regarding Windows Media, not to end-users. This will offer the most tweakable solution to vendors and potentially a way to make the most efficient server in town, without opening the realmedia format. They know full well the minute they do that alternative players will pop up that don't show ads, that don't 'call home' to real, and offer an easy to use interface that is easier to get at than real will allow its player to be, and their player software will fall by the wayside. Without the ads/private information provided by their player, their business model would fall apart and they know it. They've been thus far relying on being encumbant to power their usage, and in the face of the Microsoft behemoth, that is going away.
They are perfectly willing to sacrifice ad revenue and excessive personal info collection for their linux player for the sake of claiming to be more cross-platform (even their developing RealOne player doesn't do ads and doesn't even offer the 'gold pass' option). If they were sure that the information would only be used for linux and other 'fringe system' decoding, they would gladly open up their format. They don't seem to like developing their non-windows versions as they take forever to do so, and that would mean they no longer had to worry. Mplayer has had some success with their binary codecs, and I doubt they could care. The current linux user base would provide more bad than useful data anyway in their forms and not tend to be influenced by Real's marketing. Linux is just a nice platform to say 'look, look, does Windows media or quicktime work natively with linux? we must be better, we are cross-platform!"
But the second ZoomPlayer had realmedia support, Real could kiss RealOne's popularity goodbye, and then next stop, chapter 11....
When I had AI, my professor was working on a project using the unreal tournament engine. The idea was to have scripted events triggers, and an evolving storyline affected by the viewer who is allowed the most freedom the engine allows. The demo shown in class was an aquarium where the 'plaques' were generated on the fly using some basic facts, in whatever language the user requested. The plaques never looked the same twice and always offered different facts. The english sentences generated were pretty good and convincing enough to think things were scripted, but weren't. Character actions and dialog are not scripted, but instead they have a knowledge base and goals and rules of how goals can be acheived that influence their speech and action. Truly a gargantuan project but really neat to see it in action. Agents showed visible pauses at decision points in the demo. I have a friend who still works on the project with the college. Now this sort of stuff could make for some really interesting experiences when brought to its ultimate incarnation. Currently it requires three high-end systems to run, and even then the pauses are rather annoying, but given time this could be really exciting. Imagine an RPG with this technology where NPCs have goals and knowledge bases instead of scripts....("Sigh... Times are tough..."). That would kick serious ass.
I know part of the reason for using small DVDs was anti-piracy (though dvd-rs in that form factor should be available before long, if not already...), but perhaps some stuff in the article relates to the choice of such a small disc. With the small disc, there is a much more consistant speed with a constant rotation rate from innermost to outermost track. And that speed could be close to normal DVDs at the outermost tracks, since it could be spun faster with lower risk and noise... Just an offtopic thought...
But not much more. He said what we all know, commercial apps are not so plentiful under linux and many users are scared off.
However, to say there is no future on anything but embedded and headless servers is extremely stupid. Maybe not for the common user, but among professional users who *do* care about the stability of the underlying OS and who *know* where to go to get the apps, linux is great. And not just computer professionals, I know people from various science disciplines using it as well, and also friends of mine run linux even if non-techinical, because they can ask me for help and I can usually give it quickly. The desktop is alive and well, but not for Joe Schmoe, but among professionals it is gaining considerable share... The move to an NT based kernel has appeased some, but not all Windows users sick of the underlying instability. MacOSX has a great thing going, but the price is too high. I'm sure MacOSX could stamp out linux desktops, as they offer all that does and more as far as desktop use is concerned, but the price is too high and they couldn't care less about winning anything but the Windows market...
Frankly, I think his stance is more influenced by the decline of enlightenment's popularity (and his resultant decline in fame) and potentially some business interest in his coding with regards to embedded applications. I would dare say there are just as many disadvantages in the embedded arena for linux as the desktop, since systems like QNX are much more adapted to the environment than linux...
A Windows 98 'workalike'? Windows 98 doesn't really 'work' per say, so to workalike seems to be a apardox...
But Seriously, why name a 9x system as a goal rather than an NT based one as a goal, 9x's core is very flawed, and since they would be starting from scratch anyway...
I have never ever had an AMD CPU go bad, this FUD annoys me. I've had models from their 286 way back when, to the Athlon XP, and not one of them ever showed a hint of a problem. The K6 series underperformed, but still ran solid as could be... More than I can say for the two intel processors I bought, one which had the fdiv bug and the other which had the F00F bug.... I know, two exceptions to intel quality, but don't pretend AMD is a piece of crap while glossing over Intel's flaws..
Same thing happened with me, but I just knew some issues to twek, the exception being of course the XV atribute lockup which would bite people in the ass, but the opther stuff an 'average' user might not ever notice.
The top 20 list admits to being flawed, and the data there is pretty scarce, hardly a story.
To be on topic, I'm using a little Presario 725US w/ 1.4 GHz. Athlon. Pretty good price/performance, and works well with linux out of the box...
Of course, to get the most out of it, I have patched the kernel for PowerNow and ACPI, to extend battery life, reduce heat, and lower fan noise. Also applied a 'kacpid' patch to kernel to recover lost acpi interrupts. Because it has no builtin support for suspend, I also have the swsup patch applied. The sound also required a patch to actually work. The savage chipset driver with XFree I replaced with a more up-to-date version with better performance, but no matter what the driver locks up when xv attributes are set, so I have to patch xine to run it and RealOne is out of the question...
Ok, so it isn't *that* great out of the box but it was a hell of a lot cheaper than the competition out there and the end result is a solid system, at a price of 1200 new (at the time, after rebate).
Time for another wacky ogg format for stilss:) On a more serious note, presuming that they do not infringe on other patents, would jpeg2000 or a 'lossy' png be a feasible replacement. Mainly I was wordering about png, since jpeg2000 is obviously a workable solution if not patent infringing.
Incorrect, while it may be in pratcial turns difficult to enforce such a patent now, they do not forfeit their rights over it by not enforcing it. That is true for things such as copyright, but patents are set in stone, that is why the process costs so damn much, it lets you pull sleazy tricks like this. Patents need reform, patent holders *should* forfeit ownership if they neglect to take any action whatsoever for such a long time while it becomes a standard.
Someone signed up for an account with the name 1 and posted a hoax story? Big deal. Surprising that 1 was still available, but it obviously is what happened. I mean come on, the page did not look professional nor apple-like, and why would it have the/1/ subdir in there huh, or did no one notice it is in a place where a common account could post?
Though not a side door, many Dell's ship where you can open things up that wide and easy.... Hard to do in a DIY system, but OEMs are certainly capable of such systems...
I disagree. While graphical installers make things better for the novice, for the advanced user it has little to offer and constricts in so many ways. People don't just choose to use a terminal over a gui just to be l337, I and many more people use it because it is faster to get stuff done and also it can be highly configured and scripted to do things automatically.
I mean, a gui has to give all the options through visual elements, meaning a lot of buttons/menus to navigate. The only way to keep any reasonably powerful application from being unwieldy is to stick options into many submenus. While this helps those who do not know what options or available, or need prompting to remember what to do, for people who do know and remember, the command line allows entering it directly. Hell, with a nice shell and good completion functions, tab completion could show what can be done at any point. GUIs are nice to complement command line utilities, but people using the command line are not simply doing it to be 'elite'.
Slashdot Mirror
Anyone in touch with people building system component by component realizes a great majority of these are AMD based, since they get to see the processor price difference and know the price/performance problems with P4 vs. Athlon. I personally have been eagerly awaiting a version with AGP for AMD, and then I might buy in... I want to have an All-in-Wonder in one of these babies to integrate into my entertainment system...
IIRC, it not only strengthens the 'shrink wrap' licenses, but also sets a default liability. So just saying 'this comes with no warranty' in a file somewhere does not free the programmer of responsibilty. Commercial companies don't care as they are explicitly allowed to shirk this liability through their shrink-wrap license, but open source applications can't really have a shrink-wrap license, so no matter how much you deny liability, you are stuck with it..
If by 'pretty good' you mean 'total piece of shit', I agree. Most of the Visual Studio suite is pretty decent, but Source Safe comes off exactly as what it is, a software package bought out by MS for the express purpose of saying Visual Studio has SCM, and then doing next to nothing with it to make it work better. As an admin in a shop of people begging to leave Source Safe for something like Rational, or even CVS, I know how it is... Subversion may offer the core needed to build a replacement that keeps both developers and the businessmen happy...
quake3 works fine with gcc3.1, trust me :)... dunno about the rest..
I've been running gcc 3.1 compiled gentoo for a while now. Very nice and offers some tangible speedups. However, the costs:
Browser Plugins:
Flash plugin required me to write a small compatibility library to mimick some the old libstdc++ mangled memory allocation schemes. This will probably not appear in a Mandrake desktop, as they will likely provide a -compat library without the user knowing.
Never have gotten java plugins to work... Just haven't figured it out for blackdown, ibm's, nor suns.... Realplayer plugin problem same as Flash. Right now I am just lacking java...
Build:
Some programs won't build out of the box. Some due to bad code, but mostly due to strange build configuration. For example, basiliskII's build fails at one point when gcc is used to link object files generated by g++ and bombs because some g++ symbols are unknown to gcc, switching that gcc to g++ makes that step go by... Others I've had issues with include PixiePlus, mame, and openoffice.
Others may have issues. I don't use crossover as vanilla wine fills all my needs, and I have nothing in the way of commercial software aside from games, which all *worked*, (every quake, civctp). The biggest problem I've had is again, c++ browser plugins...
What you are referring to is XFree86 4.2.0, which has the foundation of X11R6.6, so be happy.
I can't stand VS .net, I started using it and then tired of the crap and found VS6 much more straightforward. In either case I'm glad I rarely need Visual Studio, as most of my development is either *nix or cross-platform.
Complicates the cache issue? How? Basically, they are making two complete hard drive assemblies on a common assembly, but one can only is designed from the ground up to be read-only, from the head to the interface, providing the most possible levels of protection to get through. If the drive head is physically incapable of modifying data, then no bug in any software or firmware can be explotied to allow writes.
I personally think this is useless for other reasons, completely redundant with networked strategies to do the same thing, at least with competent administrators. However, I guess this thing would be good for places with less-than-stellar administration. Then, they could have the web server's only connection to the data through that safe harddrive cable. If an organization is just serving up static data, then they probably aren't sophisticated enough of an organization to afford stellar admin people, so I guess this thing has a place.... No where near me though.
Perhaps that is promising, but I'm still not sure what Helix is meant to encompass. Wading through the marketspeak doesn't yield anything definite. Is the Helix 'client' a full-featured player, or the client-end of a content delivery framework which feeds retrieved data into a binary-only plugin or codec? I'll wager the core media encodings will remain binary-only, even if/when a good deal of the platform is open. If stripped down to the point where only the codec itself is binary-only and the interfaces to that codec are open, that would be very nice too, but I doubt they can implement a pure, open platform for decoding all media content (Sorenson and Microsoft will allow their stuff to be open when hell freezes over), and since their competitors won't be allowing open codecs, there is little incentive for them to open up their codec...
I'll wager that even if the 'client' is a media player, a closed core will handle the decoding, but if the player doesn't suck or the interfaces to that core are open, I'll be content...
I'm glad they are making a cross-platform effort, but am not glad that their cross-platform strategy comes off as more of a peripheral 'as they feel like it' effort rather than a serious effort. The non-Windows players are hidden from www.real.com and lacking in development effort. RealOne for linux is in alpha, while it has been out for a long time with windows. I only care about this because RealOne will make use of the Xvideo extension. In general, Real's player has bad controls (in terms of keyboard particluarly) and I want an alternative player, or a centralized player that can play whatever I want, be it DVD, avi, mov, mpg, or rm, and Real's (as well as apple's) efforts aren't allowing this sort of solution to gel, particularly with non-windows,non-mac systems. MPlayer and xine and others are trying damn hard to acheive this, Real and Apple be damned, but with a more cooperative approach (as this Helix, on second glance, *might* represent), I could acheive my multimedia dream under linux.
I think his point is valid. Even if a cracker roots a server, he can't change stuff if a separate system dictates permission. Given proper firewalling rules and/or login configuration, the second system is just as untouchable as the host system in this multi-hard dirve configuration. The key difference would be performance (the two systems don't care what the other is doing), but with the typical bandwidth bottleneck, hard drive IO is the least of your concerns, particularly with static content.
I saw no mention of open sourcing the player, or even mention of anything that would necessarily do encoding... It seems the portion being open sourced would be transport and delivery, and even the license to that would be restrictive.
This is a move to appeal to providers starting to go on the fence regarding Windows Media, not to end-users. This will offer the most tweakable solution to vendors and potentially a way to make the most efficient server in town, without opening the realmedia format. They know full well the minute they do that alternative players will pop up that don't show ads, that don't 'call home' to real, and offer an easy to use interface that is easier to get at than real will allow its player to be, and their player software will fall by the wayside.
Without the ads/private information provided by their player, their business model would fall apart and they know it. They've been thus far relying on being encumbant to power their usage, and in the face of the Microsoft behemoth, that is going away.
They are perfectly willing to sacrifice ad revenue and excessive personal info collection for their linux player for the sake of claiming to be more cross-platform (even their developing RealOne player doesn't do ads and doesn't even offer the 'gold pass' option). If they were sure that the information would only be used for linux and other 'fringe system' decoding, they would gladly open up their format. They don't seem to like developing their non-windows versions as they take forever to do so, and that would mean they no longer had to worry. Mplayer has had some success with their binary codecs, and I doubt they could care. The current linux user base would provide more bad than useful data anyway in their forms and not tend to be influenced by Real's marketing. Linux is just a nice platform to say 'look, look, does Windows media or quicktime work natively with linux? we must be better, we are cross-platform!"
But the second ZoomPlayer had realmedia support, Real could kiss RealOne's popularity goodbye, and then next stop, chapter 11....
When I had AI, my professor was working on a project using the unreal tournament engine. The idea was to have scripted events triggers, and an evolving storyline affected by the viewer who is allowed the most freedom the engine allows. The demo shown in class was an aquarium where the 'plaques' were generated on the fly using some basic facts, in whatever language the user requested. The plaques never looked the same .("Sigh... Times are tough..."). That would kick serious ass.
twice and always offered different facts. The english sentences generated were pretty good and convincing enough to think things were scripted, but weren't. Character actions and dialog are not scripted, but instead they have a knowledge base and goals and rules of how goals can be acheived that influence their speech and action. Truly a gargantuan project but really neat to see it in action. Agents showed visible pauses at decision points in the demo. I have a friend who still works on the project with the college. Now this sort of stuff could make for some really interesting experiences when brought to its ultimate incarnation. Currently it requires three high-end systems to run, and even then the pauses are rather annoying, but given time this could be really exciting. Imagine an RPG with this technology where NPCs have goals and knowledge bases instead of scripts...
I know part of the reason for using small DVDs was anti-piracy (though dvd-rs in that form factor should be available before long, if not already...), but perhaps some stuff in the article relates to the choice of such a small disc. With the small disc, there is a much more consistant speed with a constant rotation rate from innermost to outermost track. And that speed could be close to normal DVDs at the outermost tracks, since it could be spun faster with lower risk and noise... Just an offtopic thought...
But not much more. He said what we all know, commercial apps are not so plentiful under linux and many users are scared off.
However, to say there is no future on anything but embedded and headless servers is extremely stupid. Maybe not for the common user, but among professional users who *do* care about the stability of the underlying OS and who *know* where to go to get the apps, linux is great. And not just computer professionals, I know people from various science disciplines using it as well, and also friends of mine run linux even if non-techinical, because they can ask me for help and I can usually give it quickly. The desktop is alive and well, but not for Joe Schmoe, but among professionals it is gaining considerable share... The move to an NT based kernel has appeased some, but not all Windows users sick of the underlying instability. MacOSX has a great thing going, but the price is too high. I'm sure MacOSX could stamp out linux desktops, as they offer all that does and more as far as desktop use is concerned, but the price is too high and they couldn't care less about winning anything but the Windows market...
Frankly, I think his stance is more influenced by the decline of enlightenment's popularity (and his resultant decline in fame) and potentially some business interest in his coding with regards to embedded applications. I would dare say there are just as many disadvantages in the embedded arena for linux as the desktop, since systems like QNX are much more adapted to the environment than linux...
A Windows 98 'workalike'? Windows 98 doesn't really 'work' per say, so to workalike seems to be a apardox...
But Seriously, why name a 9x system as a goal rather than an NT based one as a goal, 9x's core is very flawed, and since they would be starting from scratch anyway...
I have never ever had an AMD CPU go bad, this FUD annoys me. I've had models from their 286 way back when, to the Athlon XP, and not one of them ever showed a hint of a problem. The K6 series underperformed, but still ran solid as could be... More than I can say for the two intel processors I bought, one which had the fdiv bug and the other which had the F00F bug.... I know, two exceptions to intel quality, but don't pretend AMD is a piece of crap while glossing over Intel's flaws..
Same thing happened with me, but I just knew some issues to twek, the exception being of course the XV atribute lockup which would bite people in the ass, but the opther stuff an 'average' user might not ever notice.
The top 20 list admits to being flawed, and the data there is pretty scarce, hardly a story.
To be on topic, I'm using a little Presario 725US w/ 1.4 GHz. Athlon. Pretty good price/performance, and works well with linux out of the box...
Of course, to get the most out of it, I have patched the kernel for PowerNow and ACPI, to extend battery life, reduce heat, and lower fan noise. Also applied a 'kacpid' patch to kernel to recover lost acpi interrupts. Because it has no builtin support for suspend, I also have the swsup patch applied. The sound also required a patch to actually work. The savage chipset driver with XFree I replaced with a more up-to-date version with better performance, but no matter what the driver locks up when xv attributes are set, so I have to patch xine to run it and RealOne is out of the question...
Ok, so it isn't *that* great out of the box but it was a hell of a lot cheaper than the competition out there and the end result is a solid system, at a price of 1200 new (at the time, after rebate).
Time for another wacky ogg format for stilss :) On a more serious note, presuming that they do not infringe on other patents, would jpeg2000 or a 'lossy' png be a feasible replacement. Mainly I was wordering about png, since jpeg2000 is obviously a workable solution if not patent infringing.
IAANAL (I Am Also Not a Lawyer), but if memory serves, that is true for Copyright, maybe trademark, but patents do not have that restrictiobn
Incorrect, while it may be in pratcial turns difficult to enforce such a patent now, they do not forfeit their rights over it by not enforcing it. That is true for things such as copyright, but patents are set in stone, that is why the process costs so damn much, it lets you pull sleazy tricks like this. Patents need reform, patent holders *should* forfeit ownership if they neglect to take any action whatsoever for such a long time while it becomes a standard.
Someone signed up for an account with the name 1 and posted a hoax story? Big deal. Surprising that 1 was still available, but it obviously is what happened. I mean come on, the page did not look professional nor apple-like, and why would it have the /1/ subdir in there huh, or did no one notice it is in a place where a common account could post?
Though not a side door, many Dell's ship where you can open things up that wide and easy.... Hard to do in a DIY system, but OEMs are certainly capable of such systems...
I disagree. While graphical installers make things better for the novice, for the advanced user it has little to offer and constricts in so many ways. People don't just choose to use a terminal over a gui just to be l337, I and many more people use it because it is faster to get stuff done and also it can be highly configured and scripted to do things automatically.
I mean, a gui has to give all the options through visual elements, meaning a lot of buttons/menus to navigate. The only way to keep any reasonably powerful application from being unwieldy is to stick options into many submenus. While this helps those who do not know what options or available, or need prompting to remember what to do, for people who do know and remember, the command line allows entering it directly. Hell, with a nice shell and good completion functions, tab completion could show what can be done at any point. GUIs are nice to complement command line utilities, but people using the command line are not simply doing it to be 'elite'.