Can anybody recomment a good LiveCD for recovery purporse? Have have tried multiple, but they end up either being to colorfull (lots of stuff I don't need for recovery, KDE, graphical Grub, etc) or to minimalistic (lacking ssh, rsync, etc.).
Stuff I need would be mainly Grub (important!), ssh, rsync, cfdisk, sfdisk, parted, etc. and good driver support (ide controller, network cards).
### Try your "Google criteria" with Windows. You will get an equally confusing raft of crap that pops up.
Aehm, no. On Windows he might and up with different choices of software for Windows, with Linux he ends up with different choices of Linux itself, big difference. A piece of Windows software I can install, test, deinstall, no problem, no incompatbilities to worry about. With Linux however half the solution one finds require a different brand of Linux to work and its basically impossible for JoeSixpack to install it on his other brand of Linux.
The throuble isn't that there are dozens of different Linux distros around, but the fact that they are all mostly incompatible to each other and that software is most often only build to work on exactly one distro and even a specific version of it. Compiling from source and solving dependencies manually is not really the kind of answer that helps JoeSixpack.
### The only reason people can use Windows and MacOS, but they "can't" use Linux is that Linux doesn't come pre-installed on a desktop system from any of the major players. People never have to install OSX or Windows, and that's a major advantage.
That and because its trivial to install other software under Windows and MacOS, while its almost impossible for the normal user to install something under Linux when it doesn't come with the distro. Moving from 'I have seen some software on some webpage' to 'I can use the software' doesn't take much more then a few minutes under Windows/MacOSX, most of which is just download time and install time, both of which fully automatic, under Linux such a issue can turn into a day long job, full of manual work, webpage browsing, groups.google browsing for tips&tricks and workaround, dependency resolution and other stuff which is impossible todo for the average user and still a pain for the experienced one.
### This seems to pop every once and again, in different varieties: "there's too many distros/desktop projects/widget sets/web browsers/Hello Kitty squid cookies to choose from. Why can't we have just one?"
I don't think many people want 'just one' distro, two, three or maybe even four are for sure ok, LFS and SuSE for example have completly different goals so its ok to keep them seperate, but most of the rest is just trying to solve exactly the same problem. Its 95% NIH syndrome and maybe 5% real difference.
### Users pick different distros/desktops and so on because they have different needs and different preferences.
No, they pick different distros because they are so many of them arround, not because one does suit their needs better then the other. Most people simply stay attached to whatever they touched first. What is for example the major difference between SuSE, Redhat, Fedora and Mandrake? All of them seem to solve exactly the same problem and none really solves it a lot better then the other and all end up with half done solutions because they don't work together. XFree86 configuration for example still is a pain, actually today more then five years ago when we had XF86Setup, not because its especially hard to solve this problem, but because everybody tries to solve it on its own. There are dozens of XFree86 config helpers around, none however is complete, most of them are incompatible and half of them would overwrite the others work if combined. There is really NO benefit of having a dozen broken solutions vs one working one.
The real throuble is that it has become a habit to solve problems under Linux with creating yet another custom incompatible distro, which is just plain stupid. For each problem a custom distro solves it creates two other due to the incompatibility with other distros. Under other OSs people solve problems by writing software which works everywhere, under Linux however people prefer to stay incompatible with each other.
LSB might one day reduce the number of custom distros since people can then just package the software that they actually need to solve their problem and arn't forced to build a whole distro around their solution, but we are not their yet and probally won't be for a while.
As long as each Linux is basically completly incompatible to the next, except via recompile, I really don't see Linux having any chance of ever entering mainstream.
First of that thing is called SpaceshipOne, secondly it did go nowwhere near where the shuttle went. SpaceshipOne did a little hop out of the atmosphere and then got back, didn't even need a heatshield for that. Bringing something into a stable orbit is a whole different beast (100km vs 400km + heck a lot more speed). The NASA did basically the same as SpaceShipOne in the 1960s with its X-15.
That said, yes, the NASA could probally be a lot more cost effective, but just saying SpaceShipOne did for 20mio$ what the Shuttle does is way off and basically just wrong. SpaceShipOne will never be capable todo what the shuttle does, to accomplish that they have design a completle new vehicle.
### 5) Linux now has hundreds of game titles, thanks to PS3 development
Aehm, no. That conclusion is way of base. Even if Sony uses some custom Linux, that has no implications on Linux in general, it doesn't even have much implications on the PS3 itself. Sony can and most likly will lock down that Linux as much they can, they for sure don't want to make it easy for everybody to pirate games with a one-click solution, so no ssh into your PS3 and 'dd' to get your image file. And neither will they use standard OpenGL, their hardware will have special features which the game will want to use, I kind of doubt that they will wait till the ARB has accepted their extensions instead they will make it so that it works for the PS3 and give little to nothing about compatibility. They for sure won't make much of their API puplically available either.
Last not least, even if their Linux will be open and even if you can dd your PS3 DVDs into an ISO file or export them via NFS, you are still stuck with some weird architecture binary which will be borth nothing on your standard PC.
If Sony uses Linux they use it to have a stable TCP/IP stack or whatever, the rest will most likly be 100% custom stuff for the PS3 complele incompatible with anything you know from the PC side and the games will most likly end up using custom PS3 APIs instead of being proper POSIX apps.
Linux on PS3, if it ever comes, will be no more then a tiny implementation detail, completly irrelevant to anybody then some geeks.
Just because something has a Linux sticker on it doesn't mean it will be open, hackable or anything. PS2 Linux didn't even allow to hack the memory cards, PS3 Linux for sure won't allow that either.
WineX/Cedega is as close to mainstream Linux gaming as you will come in the next years.
Those interested in zoomable interface might want to have a look at Flexlay, it is basically a collection of OpenGL based layers and objects that can be placed and edited on an almost unlimited large workspace (as much as a float can hold). Its currently mainly usefull as a simple editor for 2d games (SuperTux, netPanzer), but also comes with a drawing component that allows todo simple paint operations (like Gimps brushtool) on an unlimited and non-pixel based canvas. Beside zooming and panning it also has support for rotating the drawing area in realtime, which give it quite a natural paper-like feel. There is support for graphic tablets too.
Those interested in it should check out the latest SVN version, since the last release doesn't contain any of the more interesting features.
PS: This is blunt self-advertisment, hope you can forgive me, but it kind of fitted here and might be interesting for some people. And by the way its GPL.
### You simply cannot do this sort of thing in realtime, which is what gaming requires.
You missed the point. The fact is that you CAN do all these things in realtime, not at the same resolution, not with the funcky raytracer enabled to get caustics and such, but normal, paralax mapping and all those other things that you can do with todays GPU will get you extremly close to what a ten year old render farm put out. Effects in Terminator2, Abyss and the like are quite doable in realtime today.
The thing that you however still can't do is get the animation and camera angles right. Movies have a director, a bunch of artists, camera mans, maybe a crew to do motion capture and whatever. All those people tweak the hell out of a few second long sequence to make it look cool, these people do the major work to make it look good. Now what have you in a game? A crappy AI, a very limited number of motion captured animation and Joe Sixpack gamer who is bunny hopping around the final boss. No matter how much rendering power you throw at it, Joe in bunny hopping mode will always look like Joe in bunny hopping mode and not like Arnold doing a cool fight at the end of Terminator2. The simple reason is that there is nobody to do any directing, no actors to do the job and no way to change the camera angle without turning the game into an uncontrolable something.
To make the difference a bit more clear, just compare GTA or GrandTurrismo with some car chase in say Ronin, sure Ronin is a real movie with real actiors and such so CPU power is still a bit lacking to get that, but the point is that in games you always sit behind the steering wheel, while in a movie you get all kinds of interesting camera angles, you simply can't get that without interrupting the gameplay quite a bit. There are some attemps to bring movies and games closer together by adding something along the line of 'inline' cutscenes (NFS, Burnout3) where you get an instant replay of crashes or jumps, but they still are far far away from a proper directed car chase in a movie.
Last not least, movies are 2 hour long, everything that doesn't surve a purpose gets cut out, games are 10-30 hours long and nothing gets cut out, if the player walks a few minutes or hours in circle since he can't find the exist, so be it, no director there to change it.
Thats a difference that simply fades away over time, a ten your old renderfarm can't really do much more then a PC from today. Sure a PC can't do exactly what the renderfarm can, but thanks to better algorithms it can get results extremly close to that of the renderfarm with much less wasted CPU/GPU (no need to render hundreds of polygons when a single one with normal or parallaxmapping will do). PC also don't need to do renderings at cinema useable resolutions which reduces the work they need to do even more.
If you compare the ten year old Toy Story with todays
Doom3 there really is not much difference when it comes to the rendering quality, a little to less polys here and a unsharp texture there, but thats something that will quickly disappear in the next years. Beside from that, such issues are exactly my point, in a movie a blury texture doesn't matter, you simply don't move the camera close enough so that the viewer notices, in a game players will stick there noise into the wall, so even if the PC renders better then the movie, the results will be worse. Interactivity ruins the results today far more then the rendering power.
Its actually less bullshit than one might think at first. The biggest difference between Jurassic Park and a game with dinosaures is not the polygon count, but that the game is interactive, while the movie is not.
In a movie you have a fixed set of camera angles and actions to be performed, if you could throw all your polygons, artists and CPU power to render those, you would get results close enough to the movie. However in a game you end up having neither a fixed camera angle, nor fixed actions, most stuff is up to the player. You just don't have enough artists to tweak each and every situation. One time the player might have a bazooker, next time a MG and next time he might want to crash into the dino with his jeep. So since you can't prescript all actions you have to let a physic engine and AI handle it, which in turn burns down valueable CPU, which you no longer can use for pushing polys around, in addition to that you no longer have an artists involved who can fine tune the stuff that happens on screen, so you might run into clipping errors or silly looking situations.
Overall it is simply impossible to get an interactive situation look as good as a movie, even if you have all the CPU power you need at hand you still lack the artists for the fine tuning and often have zero control over the camera angle.
Beside from that we already are in a situation where yesterdays cutscenes are tomorrows gameplay scenes, yet, most gameplay looks for more borring then the cutscenes we saw before.
### If more people wanted them, they would have bought one.
Its a question of money, people buy what is cheap and gets the job done, not what is best. Apple computers tended to be extremly expensive compared to the PC counterparts (even so that is changing with the Mac mini), so they only sold to a minority, that doesn't mean that they are worse. Of course today there is also a bunch of vendor log-in into play, you can't just swap architecture and OS without losing all your apps, so people stay with what they are familar with.
When it comes to source code I basically always use version control, when it are single-file things I tend to use RCS, when it gets more I switch to CVS or SVN, but when it comes to binaries (.xcf,.blend,.png, etc.) I tend to not use version control systems or at least use them to a much lesser degree. Reasons for this are mostly ease of use related, with text editors you often have build-in support for checkin/checkout, with graphical editors you seldomly have them. Being able to do a 'diff' between different versions is also relativly useless with binary files, so I end up doing manual versioning instead (file-1.png, file-2.png, file-with-special-light-2.png, etc.), since that is much easier to access from the 'editor' (Gimp, Blender) and also much easier to work with. If i want to have a look at and older version or want to compare to different versions I just load them up, would they be stored in version control it would be much more tricky to get access to them. A filesystem plugin that allows to transparently access the version control system from withing the filesystem itself without falling back to svn or cvs command line tools would surly help to fix this issue, but so far I havn't seen any of them in the wild.
Last not least there is also the problem with newer version control systems that they are meant for larger projects, managing a single dotfile with svn or cvs is quite a pain, doing it with rcs on the other side is relativly easy. Would be nice if they would provide some 'instant' mode that doesn't require a seperate repository, but just works like RCS in a dot-directory or something like this.
### but for what most people do with their computers, Linux is not only adequate - it's better
No, its still worse. Linux might be better for a bit of mail and internet or some business uses with limited application requirements, but thats as far as it goes, people do a lot of other things on their computers beside mail and internet browsing (games, tax programms, house planing, custom software from @work, etc.). And simply getting a video from the net is already a pain in the ass under Linux, yes I can use mplayer which can play a hack of a lot formats, but tell some computer newbie how to compile and install it properly so that even inline-videos on webpages work out of the box, won't be fun.
### The sad part is, people just don't know it.
It doesn't matter if they know it or not, its simply not a good replacement for Windows at the moment and probally won't for quite a bit of time into the future. If for nothing else, gaming is still the killer feature that Windows can do and Linux can't and no Doom and Unreal alone don't make Linux 'ready for gaming'.
Rule of thumb, don't listen to Linux users when...
on
Cooking With Linux
·
· Score: 2, Interesting
Rule of thumb, don't listen to Linux users when you want to know how good or bad it compares to Windows or other operating systems. I mean when I read:
"Michael Stutz has used Linux exclusively for over a decade."
and then:
"Q: Do you think that Linux has enough applications that people can completely leave Windows behind?
A: Absolutely"
and then
"Q: Is there anything you need to run Windows for?
A:..I refer to the area of e-mail viruses - they just don't make them for Linux like they do for Windows...."
Its clear that this man has really no clue at all. His Windows knowledge seems to come directly from the yellow press and his Linux knowledge seems pretty biased, after all if you use it for ten years exclusivly you might actually think that some of its issue just have to be this way and couldn't be solved otherwise. And neither seems he have much clue about what people are actually doing with Windows today. After all I think he his quite right with his limited viewpoint, todays Linux doesn't compare that badly against a ten year old Windows, sad truth is that Windows and its application has moved a lot forward, while Linux is still 10 years behind.
Linux has its niches and areas where it can show its benefits, but simply claiming that its a perfect and complete replacment for Windows today couldn't further away from reality and is only damaging Linux fame. Lies don't help, be honest about what Linux can do and especially about what it can't do, then you might have a chance that people will continue to listen to you and not just turn back and think of you as some Linux-fanboy.
Re:Many fields left where Linux is unsuitable
on
Cooking With Linux
·
· Score: 1
Sodipodis interface is one of the most ugly I have ever seen, an order of magnitude worse then Gimps ever was, I find it completly unusable. On the other side there is Inkscape, a fork of Sodipodi which has a quite lovly interface, nice to use an all. However from the feature point of view both are largly lacking, so is Sketch, yet another vector paint programm for Linux.
They are all ok for doing a bit of svg drawing and such, but for professional work they are pretty much unusable, worse then Gimp is compared to Photoshop. From the feature point of view they are somewhat inbetween CorelDraw3 and CorelDraw5, being somewhat better in some areas (anti-aliased rendering) and a lot worse in other (printing, colorseperation), those named CorelDraws have been released before 1995, world hasn't stand still till then. Another problem of Inkscape/Sodipodi seems to be that they are SVG editors and not ment as general purpose vector graphics editors, which means features not supported by SVG might have a harder turn to make it into the programm.
Its nothing special really, its more or less just syntactic sugar for passing anonymous functions/function pointers around and that is something that you can do and do in quite a few other languages too (Scheme[lambda], Lisp, C++ [STL], Python,... and even C [qsort]). Ruby just does it in a less typing intensive way, but the basic idea is not so much different then what you do in other languages.
### Image quality is generally considered to be better on an LCD, as each pixel is generated by a specific set of transistors in the screen, which produces a crisp image.
But only when the screen resolution matches the LCDs resolution, which of course isn't much an issue for desktop work, but for those who play a game once in a while this is a huge issue. Most newer games can adopt to different resolutions, however both older games and games that force you to switch to a lower resolution for performance reason are a real problem, the image quality for them ends up being far worse then CRT due to the scaling of the image. This is making LCDs quite a bit less flexible then CRTs when it comes to choosing your resolution, since you are basically stuck with whatever your LCD provides and the multiples of that of course.
### Is DirectX a better game programming platform than OpenGL?
First of you have to seperate, OpenGL is not a 'game programming platform', its a graphic API and really doesn't do anything more beside graphic, no sound, no joystick handling, nada, just polygon and line drawing, you actually can't even OpenGL a window with pure OpenGL.
DirectX on the other side is a whole suit for gaming, contains everything from sound to input to network and graphics. The Direct3D part of it is the one that is comparable with OpenGL.
Comparing OpenGL to Direct3D today ends up more or less as a matter of taste, some people prefer OpenGL, some Direct3D, in the past OpenGL was a bit of ahead, but Direct3D has catched up a lot.
The non-graphic part of DirectX is surly a bit of an advantage, since people have one thing to concentrate on and don't have to cross their fingers that the way they access the soundcard is going to work on another distro (Alsa vs OSS vs some sound daemon), but I don't think it is that much of an issue. Linux isn't changing the way it handles stuff every second day, so it shouldn't be to difficult to adopt to Linux. The real throuble is really more the market share. Its not so much the lack of Linux users, but more those who are interested in gaming. Those that are interested, simply dual-boot, those that are not, well, they wouldn't buy a game anyway. To make it short, there are just very little Linux users out of which a company could make money and to serve those the hassle to support Linux is just to big. MacOSX for example might have a smaller number of users than Linux, but a lot more people interested in actually buying games for MacOSX, thats why you see more games there and so little on Linux.
5) When there is a painless way to build LSB-conforming binaries and packages. lsbdev and lsbcc might sooner or later provide that, but last time I tried I wasn't really able to get much done at all.
You could use a LSB conforming RPM on both SuSE and Redhat, yes, but you could NOT just pick a random SuSE package and use that on Redhat.
LSB packages work basically completly independend from what the distro provides. If a distro is LSB conforming it only means that it can install LSB-rpms (by using alien on Debian for example), it does not mean that the distro itself consist of LSB-rpms.
Ok, well maybe not exactly doomed to fail, after all Live Isos are nice to play around with for a couple of minutes, but I really don't see all that much use for billions of live-isos popping out all over the place. Live isos are good for recovery purpose, to give a quick first look at Linux and maybe for some kiosk installations, but beside from that their are relativly useless.
Live-Isos are really not the solution to lots of Linux problems, instead of seeing multiple games live-isos pop-up all over the place I would MUCH prefer a repository of statically linked/self-containing free games packaged in a distribution independend way, since that would be something that I as a Linux user could make use of. Such a Linux-games iso on the other side is basically completly useless for me as a Linux users. And on the other side its equally useless for the Windows people, I mean which Windows user can you impress with SuperTux, TORCS and friends? They might not be the worst games ever, but the Windows world has tons and tons of better games to offer. Most Windows people will most likly just laugh a bit and turn back to Windows world where they can have real games.
It would mean that one Linux would finally be compatible to another one, as it stands now you have a very hard time getting a binary build on distro X to run on distro Y.
### Would you like to run my cell-phone version of embedded Linux on your desktop?
Nope, but thats not the point. For different purposes, different Linux distros might make sense, however today we have multiple dozens of distros that serve *exactly* the same purpose.
Binary xml wouldn't be just about getting the files smaller, but also about making the parsers simpler. Parsing a xml file today is quite complex and slow, sure it doesn't matter much for a webpage or two, but if you have larger amounts of data its really no fun at all, a proper binary XML standard might speed that up by an order of magnitude or two.
### Can anyone think of any way to ease this translation?
Emacs does this quite well. If you find some key combo that doesn't behave the way you want it, you type 'C-h k ' and instantly you will be pointed to the function that is bound to that key and its documentation, from there on you can now either click your way through the documentation to find related function or click the source link and you end up in the exact location in the source where that function is defined. There is also 'M-x apropos' with which you can search for functions by name and a whole lot of other nice things that allow you to browse the source.
In addition to all that Emacs allows you do heavily customize it without even touching the source itself at all, there are lots of hooks and variables that you can tweak or which you can use to insert your own code.
Emacs of course has a lot of problems, but when it comes to introspection and modification of the source code its still one of the best apps around there.
Slashdot Mirror
Can anybody recomment a good LiveCD for recovery purporse? Have have tried multiple, but they end up either being to colorfull (lots of stuff I don't need for recovery, KDE, graphical Grub, etc) or to minimalistic (lacking ssh, rsync, etc.).
Stuff I need would be mainly Grub (important!), ssh, rsync, cfdisk, sfdisk, parted, etc. and good driver support (ide controller, network cards).
### Try your "Google criteria" with Windows. You will get an equally confusing raft of crap that pops up.
Aehm, no. On Windows he might and up with different choices of software for Windows, with Linux he ends up with different choices of Linux itself, big difference. A piece of Windows software I can install, test, deinstall, no problem, no incompatbilities to worry about. With Linux however half the solution one finds require a different brand of Linux to work and its basically impossible for JoeSixpack to install it on his other brand of Linux.
The throuble isn't that there are dozens of different Linux distros around, but the fact that they are all mostly incompatible to each other and that software is most often only build to work on exactly one distro and even a specific version of it. Compiling from source and solving dependencies manually is not really the kind of answer that helps JoeSixpack.
### The only reason people can use Windows and MacOS, but they "can't" use Linux is that Linux doesn't come pre-installed on a desktop system from any of the major players. People never have to install OSX or Windows, and that's a major advantage.
That and because its trivial to install other software under Windows and MacOS, while its almost impossible for the normal user to install something under Linux when it doesn't come with the distro. Moving from 'I have seen some software on some webpage' to 'I can use the software' doesn't take much more then a few minutes under Windows/MacOSX, most of which is just download time and install time, both of which fully automatic, under Linux such a issue can turn into a day long job, full of manual work, webpage browsing, groups.google browsing for tips&tricks and workaround, dependency resolution and other stuff which is impossible todo for the average user and still a pain for the experienced one.
### This seems to pop every once and again, in different varieties: "there's too many distros/desktop projects/widget sets/web browsers/Hello Kitty squid cookies to choose from. Why can't we have just one?"
I don't think many people want 'just one' distro, two, three or maybe even four are for sure ok, LFS and SuSE for example have completly different goals so its ok to keep them seperate, but most of the rest is just trying to solve exactly the same problem. Its 95% NIH syndrome and maybe 5% real difference.
### Users pick different distros/desktops and so on because they have different needs and different preferences.
No, they pick different distros because they are so many of them arround, not because one does suit their needs better then the other. Most people simply stay attached to whatever they touched first. What is for example the major difference between SuSE, Redhat, Fedora and Mandrake? All of them seem to solve exactly the same problem and none really solves it a lot better then the other and all end up with half done solutions because they don't work together. XFree86 configuration for example still is a pain, actually today more then five years ago when we had XF86Setup, not because its especially hard to solve this problem, but because everybody tries to solve it on its own. There are dozens of XFree86 config helpers around, none however is complete, most of them are incompatible and half of them would overwrite the others work if combined. There is really NO benefit of having a dozen broken solutions vs one working one.
The real throuble is that it has become a habit to solve problems under Linux with creating yet another custom incompatible distro, which is just plain stupid. For each problem a custom distro solves it creates two other due to the incompatibility with other distros. Under other OSs people solve problems by writing software which works everywhere, under Linux however people prefer to stay incompatible with each other.
LSB might one day reduce the number of custom distros since people can then just package the software that they actually need to solve their problem and arn't forced to build a whole distro around their solution, but we are not their yet and probally won't be for a while.
As long as each Linux is basically completly incompatible to the next, except via recompile, I really don't see Linux having any chance of ever entering mainstream.
That said, yes, the NASA could probally be a lot more cost effective, but just saying SpaceShipOne did for 20mio$ what the Shuttle does is way off and basically just wrong. SpaceShipOne will never be capable todo what the shuttle does, to accomplish that they have design a completle new vehicle.
This reminds me of the GGI Cube which did something similar. Sadly GGI is today rather irrelevant.
### 5) Linux now has hundreds of game titles, thanks to PS3 development
Aehm, no. That conclusion is way of base. Even if Sony uses some custom Linux, that has no implications on Linux in general, it doesn't even have much implications on the PS3 itself. Sony can and most likly will lock down that Linux as much they can, they for sure don't want to make it easy for everybody to pirate games with a one-click solution, so no ssh into your PS3 and 'dd' to get your image file. And neither will they use standard OpenGL, their hardware will have special features which the game will want to use, I kind of doubt that they will wait till the ARB has accepted their extensions instead they will make it so that it works for the PS3 and give little to nothing about compatibility. They for sure won't make much of their API puplically available either.
Last not least, even if their Linux will be open and even if you can dd your PS3 DVDs into an ISO file or export them via NFS, you are still stuck with some weird architecture binary which will be borth nothing on your standard PC.
If Sony uses Linux they use it to have a stable TCP/IP stack or whatever, the rest will most likly be 100% custom stuff for the PS3 complele incompatible with anything you know from the PC side and the games will most likly end up using custom PS3 APIs instead of being proper POSIX apps.
Linux on PS3, if it ever comes, will be no more then a tiny implementation detail, completly irrelevant to anybody then some geeks.
Just because something has a Linux sticker on it doesn't mean it will be open, hackable or anything. PS2 Linux didn't even allow to hack the memory cards, PS3 Linux for sure won't allow that either.
WineX/Cedega is as close to mainstream Linux gaming as you will come in the next years.
Those interested in zoomable interface might want to have a look at Flexlay, it is basically a collection of OpenGL based layers and objects that can be placed and edited on an almost unlimited large workspace (as much as a float can hold). Its currently mainly usefull as a simple editor for 2d games (SuperTux, netPanzer), but also comes with a drawing component that allows todo simple paint operations (like Gimps brushtool) on an unlimited and non-pixel based canvas. Beside zooming and panning it also has support for rotating the drawing area in realtime, which give it quite a natural paper-like feel. There is support for graphic tablets too.
Those interested in it should check out the latest SVN version, since the last release doesn't contain any of the more interesting features.
PS: This is blunt self-advertisment, hope you can forgive me, but it kind of fitted here and might be interesting for some people. And by the way its GPL.
### You simply cannot do this sort of thing in realtime, which is what gaming requires.
You missed the point. The fact is that you CAN do all these things in realtime, not at the same resolution, not with the funcky raytracer enabled to get caustics and such, but normal, paralax mapping and all those other things that you can do with todays GPU will get you extremly close to what a ten year old render farm put out. Effects in Terminator2, Abyss and the like are quite doable in realtime today.
The thing that you however still can't do is get the animation and camera angles right. Movies have a director, a bunch of artists, camera mans, maybe a crew to do motion capture and whatever. All those people tweak the hell out of a few second long sequence to make it look cool, these people do the major work to make it look good. Now what have you in a game? A crappy AI, a very limited number of motion captured animation and Joe Sixpack gamer who is bunny hopping around the final boss. No matter how much rendering power you throw at it, Joe in bunny hopping mode will always look like Joe in bunny hopping mode and not like Arnold doing a cool fight at the end of Terminator2. The simple reason is that there is nobody to do any directing, no actors to do the job and no way to change the camera angle without turning the game into an uncontrolable something.
To make the difference a bit more clear, just compare GTA or GrandTurrismo with some car chase in say Ronin, sure Ronin is a real movie with real actiors and such so CPU power is still a bit lacking to get that, but the point is that in games you always sit behind the steering wheel, while in a movie you get all kinds of interesting camera angles, you simply can't get that without interrupting the gameplay quite a bit. There are some attemps to bring movies and games closer together by adding something along the line of 'inline' cutscenes (NFS, Burnout3) where you get an instant replay of crashes or jumps, but they still are far far away from a proper directed car chase in a movie.
Last not least, movies are 2 hour long, everything that doesn't surve a purpose gets cut out, games are 10-30 hours long and nothing gets cut out, if the player walks a few minutes or hours in circle since he can't find the exist, so be it, no director there to change it.
Thats a difference that simply fades away over time, a ten your old renderfarm can't really do much more then a PC from today. Sure a PC can't do exactly what the renderfarm can, but thanks to better algorithms it can get results extremly close to that of the renderfarm with much less wasted CPU/GPU (no need to render hundreds of polygons when a single one with normal or parallaxmapping will do). PC also don't need to do renderings at cinema useable resolutions which reduces the work they need to do even more. If you compare the ten year old Toy Story with todays Doom3 there really is not much difference when it comes to the rendering quality, a little to less polys here and a unsharp texture there, but thats something that will quickly disappear in the next years. Beside from that, such issues are exactly my point, in a movie a blury texture doesn't matter, you simply don't move the camera close enough so that the viewer notices, in a game players will stick there noise into the wall, so even if the PC renders better then the movie, the results will be worse. Interactivity ruins the results today far more then the rendering power.
Its actually less bullshit than one might think at first. The biggest difference between Jurassic Park and a game with dinosaures is not the polygon count, but that the game is interactive, while the movie is not.
In a movie you have a fixed set of camera angles and actions to be performed, if you could throw all your polygons, artists and CPU power to render those, you would get results close enough to the movie. However in a game you end up having neither a fixed camera angle, nor fixed actions, most stuff is up to the player. You just don't have enough artists to tweak each and every situation. One time the player might have a bazooker, next time a MG and next time he might want to crash into the dino with his jeep. So since you can't prescript all actions you have to let a physic engine and AI handle it, which in turn burns down valueable CPU, which you no longer can use for pushing polys around, in addition to that you no longer have an artists involved who can fine tune the stuff that happens on screen, so you might run into clipping errors or silly looking situations.
Overall it is simply impossible to get an interactive situation look as good as a movie, even if you have all the CPU power you need at hand you still lack the artists for the fine tuning and often have zero control over the camera angle.
Beside from that we already are in a situation where yesterdays cutscenes are tomorrows gameplay scenes, yet, most gameplay looks for more borring then the cutscenes we saw before.
### If more people wanted them, they would have bought one.
Its a question of money, people buy what is cheap and gets the job done, not what is best. Apple computers tended to be extremly expensive compared to the PC counterparts (even so that is changing with the Mac mini), so they only sold to a minority, that doesn't mean that they are worse. Of course today there is also a bunch of vendor log-in into play, you can't just swap architecture and OS without losing all your apps, so people stay with what they are familar with.
When it comes to source code I basically always use version control, when it are single-file things I tend to use RCS, when it gets more I switch to CVS or SVN, but when it comes to binaries (.xcf, .blend, .png, etc.) I tend to not use version control systems or at least use them to a much lesser degree. Reasons for this are mostly ease of use related, with text editors you often have build-in support for checkin/checkout, with graphical editors you seldomly have them. Being able to do a 'diff' between different versions is also relativly useless with binary files, so I end up doing manual versioning instead (file-1.png, file-2.png, file-with-special-light-2.png, etc.), since that is much easier to access from the 'editor' (Gimp, Blender) and also much easier to work with. If i want to have a look at and older version or want to compare to different versions I just load them up, would they be stored in version control it would be much more tricky to get access to them. A filesystem plugin that allows to transparently access the version control system from withing the filesystem itself without falling back to svn or cvs command line tools would surly help to fix this issue, but so far I havn't seen any of them in the wild.
Last not least there is also the problem with newer version control systems that they are meant for larger projects, managing a single dotfile with svn or cvs is quite a pain, doing it with rcs on the other side is relativly easy. Would be nice if they would provide some 'instant' mode that doesn't require a seperate repository, but just works like RCS in a dot-directory or something like this.
### but for what most people do with their computers, Linux is not only adequate - it's better
No, its still worse. Linux might be better for a bit of mail and internet or some business uses with limited application requirements, but thats as far as it goes, people do a lot of other things on their computers beside mail and internet browsing (games, tax programms, house planing, custom software from @work, etc.). And simply getting a video from the net is already a pain in the ass under Linux, yes I can use mplayer which can play a hack of a lot formats, but tell some computer newbie how to compile and install it properly so that even inline-videos on webpages work out of the box, won't be fun.
### The sad part is, people just don't know it.
It doesn't matter if they know it or not, its simply not a good replacement for Windows at the moment and probally won't for quite a bit of time into the future. If for nothing else, gaming is still the killer feature that Windows can do and Linux can't and no Doom and Unreal alone don't make Linux 'ready for gaming'.
Rule of thumb, don't listen to Linux users when you want to know how good or bad it compares to Windows or other operating systems. I mean when I read:
..I refer to the area of e-mail viruses - they just don't make them for Linux like they do for Windows...."
"Michael Stutz has used Linux exclusively for over a decade."
and then:
"Q: Do you think that Linux has enough applications that people can completely leave Windows behind?
A: Absolutely"
and then
"Q: Is there anything you need to run Windows for?
A:
Its clear that this man has really no clue at all. His Windows knowledge seems to come directly from the yellow press and his Linux knowledge seems pretty biased, after all if you use it for ten years exclusivly you might actually think that some of its issue just have to be this way and couldn't be solved otherwise. And neither seems he have much clue about what people are actually doing with Windows today. After all I think he his quite right with his limited viewpoint, todays Linux doesn't compare that badly against a ten year old Windows, sad truth is that Windows and its application has moved a lot forward, while Linux is still 10 years behind.
Linux has its niches and areas where it can show its benefits, but simply claiming that its a perfect and complete replacment for Windows today couldn't further away from reality and is only damaging Linux fame. Lies don't help, be honest about what Linux can do and especially about what it can't do, then you might have a chance that people will continue to listen to you and not just turn back and think of you as some Linux-fanboy.
Sodipodis interface is one of the most ugly I have ever seen, an order of magnitude worse then Gimps ever was, I find it completly unusable. On the other side there is Inkscape, a fork of Sodipodi which has a quite lovly interface, nice to use an all. However from the feature point of view both are largly lacking, so is Sketch, yet another vector paint programm for Linux.
They are all ok for doing a bit of svg drawing and such, but for professional work they are pretty much unusable, worse then Gimp is compared to Photoshop. From the feature point of view they are somewhat inbetween CorelDraw3 and CorelDraw5, being somewhat better in some areas (anti-aliased rendering) and a lot worse in other (printing, colorseperation), those named CorelDraws have been released before 1995, world hasn't stand still till then. Another problem of Inkscape/Sodipodi seems to be that they are SVG editors and not ment as general purpose vector graphics editors, which means features not supported by SVG might have a harder turn to make it into the programm.
Its nothing special really, its more or less just syntactic sugar for passing anonymous functions/function pointers around and that is something that you can do and do in quite a few other languages too (Scheme[lambda], Lisp, C++ [STL], Python, ... and even C [qsort]). Ruby just does it in a less typing intensive way, but the basic idea is not so much different then what you do in other languages.
### Image quality is generally considered to be better on an LCD, as each pixel is generated by a specific set of transistors in the screen, which produces a crisp image.
But only when the screen resolution matches the LCDs resolution, which of course isn't much an issue for desktop work, but for those who play a game once in a while this is a huge issue. Most newer games can adopt to different resolutions, however both older games and games that force you to switch to a lower resolution for performance reason are a real problem, the image quality for them ends up being far worse then CRT due to the scaling of the image. This is making LCDs quite a bit less flexible then CRTs when it comes to choosing your resolution, since you are basically stuck with whatever your LCD provides and the multiples of that of course.
### Is DirectX a better game programming platform than OpenGL?
First of you have to seperate, OpenGL is not a 'game programming platform', its a graphic API and really doesn't do anything more beside graphic, no sound, no joystick handling, nada, just polygon and line drawing, you actually can't even OpenGL a window with pure OpenGL.
DirectX on the other side is a whole suit for gaming, contains everything from sound to input to network and graphics. The Direct3D part of it is the one that is comparable with OpenGL.
Comparing OpenGL to Direct3D today ends up more or less as a matter of taste, some people prefer OpenGL, some Direct3D, in the past OpenGL was a bit of ahead, but Direct3D has catched up a lot.
The non-graphic part of DirectX is surly a bit of an advantage, since people have one thing to concentrate on and don't have to cross their fingers that the way they access the soundcard is going to work on another distro (Alsa vs OSS vs some sound daemon), but I don't think it is that much of an issue. Linux isn't changing the way it handles stuff every second day, so it shouldn't be to difficult to adopt to Linux. The real throuble is really more the market share. Its not so much the lack of Linux users, but more those who are interested in gaming. Those that are interested, simply dual-boot, those that are not, well, they wouldn't buy a game anyway. To make it short, there are just very little Linux users out of which a company could make money and to serve those the hassle to support Linux is just to big.
MacOSX for example might have a smaller number of users than Linux, but a lot more people interested in actually buying games for MacOSX, thats why you see more games there and so little on Linux.
5) When there is a painless way to build LSB-conforming binaries and packages. lsbdev and lsbcc might sooner or later provide that, but last time I tried I wasn't really able to get much done at all.
You could use a LSB conforming RPM on both SuSE and Redhat, yes, but you could NOT just pick a random SuSE package and use that on Redhat.
LSB packages work basically completly independend from what the distro provides. If a distro is LSB conforming it only means that it can install LSB-rpms (by using alien on Debian for example), it does not mean that the distro itself consist of LSB-rpms.
Ok, well maybe not exactly doomed to fail, after all Live Isos are nice to play around with for a couple of minutes, but I really don't see all that much use for billions of live-isos popping out all over the place. Live isos are good for recovery purpose, to give a quick first look at Linux and maybe for some kiosk installations, but beside from that their are relativly useless.
Live-Isos are really not the solution to lots of Linux problems, instead of seeing multiple games live-isos pop-up all over the place I would MUCH prefer a repository of statically linked/self-containing free games packaged in a distribution independend way, since that would be something that I as a Linux user could make use of. Such a Linux-games iso on the other side is basically completly useless for me as a Linux users. And on the other side its equally useless for the Windows people, I mean which Windows user can you impress with SuperTux, TORCS and friends? They might not be the worst games ever, but the Windows world has tons and tons of better games to offer. Most Windows people will most likly just laugh a bit and turn back to Windows world where they can have real games.
### Why whould having one distro be better?
It would mean that one Linux would finally be compatible to another one, as it stands now you have a very hard time getting a binary build on distro X to run on distro Y.
### Would you like to run my cell-phone version of embedded Linux on your desktop?
Nope, but thats not the point. For different purposes, different Linux distros might make sense, however today we have multiple dozens of distros that serve *exactly* the same purpose.
Binary xml wouldn't be just about getting the files smaller, but also about making the parsers simpler. Parsing a xml file today is quite complex and slow, sure it doesn't matter much for a webpage or two, but if you have larger amounts of data its really no fun at all, a proper binary XML standard might speed that up by an order of magnitude or two.
### Can anyone think of any way to ease this translation?
Emacs does this quite well. If you find some key combo that doesn't behave the way you want it, you type 'C-h k ' and instantly you will be pointed to the function that is bound to that key and its documentation, from there on you can now either click your way through the documentation to find related function or click the source link and you end up in the exact location in the source where that function is defined. There is also 'M-x apropos' with which you can search for functions by name and a whole lot of other nice things that allow you to browse the source.
In addition to all that Emacs allows you do heavily customize it without even touching the source itself at all, there are lots of hooks and variables that you can tweak or which you can use to insert your own code.
Emacs of course has a lot of problems, but when it comes to introspection and modification of the source code its still one of the best apps around there.