After a number of requests from the community, Bash and Coreutils have been ported to SkyOS. With GCC and Make already ported, it will now be much easier to port/compile Unix-compatible applications. In other news, SkyOS 5.0 beta 7 will be available to beta testers on Monday, and will feature, among other things, multi-user support. Elsewhere, Syllable just got a Voodoo3/4/5 driver.
SkyOS Ports Bash; Syllable a Voodoo Driver
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
This is my favourite.
http://66.90.81.8/skyos.org/images/wallpaper/t13.png
Thanks for the bash
Also is it on the todo list to make font colour for the icons on the desktop to change according to the background image darkness or brightness.
eg. I think it is called drop shadows in Windows XP
I keep on hearing about hardware-accelerated OpenGL for SkyOS. Does the guy that is writing the majority of the OS have a NDA with either Nvidia or ATI? ATI isn’t giving out their 3d register-level specs anymore and Nvidia never has.
There is no way for a small OS to get 3D support without paying big bucks to ATi or nVidia and having special contracts. Yes, you can go buy by porting the Voodoo/i8xx/S3/Matrox DRI X11 drivers, but these cards are not as much used for 3D as ATi and nVidia are. And SkyOS doesn’t have that kind of money to pay AFAIK, and nVidia/ATi don’t care to support a such a small platform that doesn’t have a market.
Yep, that’s what I thought. I wonder why they go around touting 3d acceleration all the time when its pointless considering that Nvidia and ATI are really the only players in town these days.
It’s the same thing with BeOS(Haiku), B.E.O.S will be able to get 3d acceleration though.
Something that people don’t understand is that the desktop is going to be 3d accelerated with Longhorn, Linux, and MacOSX, and these alternative OS that don’t have it will look like something circa 1982 in comparison.
Why not use Scitech snap drivers ?
This is what i ‘hate’ about new os’s . You need drivers.
Lack of apps and drivers is what kills any new os.
Anyways new os project unonnium os.
http://www.scitech.com
Eric, please restrict your “hatred”. To port and use the Scitech drivers you need to pay a few thousand dollars AFAIK. Trust me, SkyOS doesn’t have that kind of money. Robert is just a student in Austria.
Let Robert figure it out.
I no they are working on the low level stuff, but I am curious about the GUI. What is the schedule for it, have any mockups begun, or are you guys further along in the development.
Hehe, “nothing to see here..move along…don’t ask questions”.
Seriously, its already “figured out”. Unless he can come up with the big bucks and sign a NDA, then no 3d acceleration for modern ATI and Nvidia cards.
Thanks for bash! In the old terminal I found it almost impossible to copy a file with a space in it’s name. That shouldn’t be a problem in bash 🙂 Can’t wait till monday.
The Syllable devs are working on making sure the internals of the OS are the way they want and making core applications atm. Not that working on a new(er) gui isn’t being worked on, but they want the stuff to work right first. Most of us in the Syllable crowd would rather it works good first, then work on making it look better than making it look pretty first then trying to make it work later.
Something that people don’t understand is that the desktop is going to be 3d accelerated with Longhorn, Linux, and MacOSX, and these alternative OS that don’t have it will look like something circa 1982 in comparison.
So you are saying that you can’t make a beautiful fast GUI without hardware 3D acceleration? Do you seriously believe that?
I think the biggest advantage of OGL GUIs is that they are scalable. That could be done in software anyway.
SkyOS isnt equale only ONE majors programmer………………marketing or bullsh*t ?
Anyway someon over there can give me a copy of this beta ?
I will try to make some reverse ingenering to find something interesting like security problems.
ONE PROGRAMMER ON AN OS == JUST HACK AS YOU WHICH
It depends what you consider beautiful. In 2010, we’re going to look back at all these desktops we have today and think they’re butt-ugly.
There’s only so far you’ll be able to go just using software without taxing the hell out of the CPU. Most desktops today are already hogs. The natural progression is to leverage those 40 million transistors in your Nvidia GPU to do something while you’re not playing Quake. Remember, cards these days are progressing towards 2d acceleration as a subset of 3d acceleration.
IThe point is that without 3d acceleration for Nvidia and ATI cards then the desktop will be seriously hobbled in the future.
But hey, use virtual consoles if you want. I don’t care.
Beauty is always subjective. It’s the fuctionality that really matters. And I don’t really see it changing _that_ much in six years. I mean it hasn’t changed much for the past 10 years.
Who sais that the looks of the GUIs has to be much more technically advanced in the future? I mean, what more is there to do besides pointless animations? Unless the whole concept changes that is.
So you are saying that you can’t make a beautiful fast GUI without hardware 3D acceleration? Do you seriously believe that?
I agree with Lumbergh. Lack of 3D acceleration is a potential killer for smaller hobby OSs. 3D acceleration means an order of magnitude more performance available to drive the GUI. This can be used to create some very rich interfaces. Hopefully, it’ll be a strong force in moving away from the pervasive windows/icons/menus/pointers paradigm into something more flexible and well-suited to each application.
To give an example, consider programs like Photoshop, Maya, or even some advanced IDEs. The traditional “window = document” model doesn’t work that well for them. They offer the user lots of highly dynamic views into a larger work, and so have a lot of extra tools to help manage those views. Eg, Photoshop’s toolbars have some more intelligence than regular application windows, Maya has dynamically selected viewports, etc. It is not at all hard to imagine taking this a step further and building a UI that really gives the user the power of dynamically changing views without the complexity of manually managing individual windows. To make that work really well, you need something like an accelerated GUI that allows the OS to spend some effort in managing “views” without bogging down the machine. Toolbar views could be animated as they move out of the way as the user works on an image. Code definition views code be made semi-transparent as the programmer needs to look at some other definition under it. Non-square windows could be used extensively without incurring undue OS overhead. The possibilities are very exciting.
Not all animations are pointless. Animations are good for visual cues so you can see what is going on (this makes a UI a lot more slick feeling/looking). Animations can, and usually do, however, go over the top… case in point, the genie effect.
“The possibilities are very exciting.”
true but its not very essential at the hobby os level. with the rich amount of open source apps that can get ported to any hobby os , bringing it up to the base functionality is much more easier now. just look at how much open source stuff that skyos has leveraged. imagine the problem of every hobby os writing a html engine and compiler. if we get more open specs from graphics device manufacturers we can easily solve this problem too. unfortunately many of them are tied with NDA’s and consider these super secret stuff. I think the situation would improve if Linux or any other Free operating system penetrates enough to make a difference at the desktop level. the rest of the alternatives could just leverage that
I just thought of another example. All compilers give a lot of feedback to the programmer, everything from error-checking to detailed optimizaton reports. The presentation of this information is very bound by existing UIs. In the WIMP paradigm, there is no good way to display that information. Windows are way too heavy-weight (there would be a dozen such windows for each page of code), and tooltips have very limited possibilities for interaction.
The best way to display that information would be to show it in-line. There is a Dylan IDE that takes a very small step in that direction. It color codes lines based on how well the optimizer was able to optimize them. However, there is a whole lot more information available than simple color coding can show.
Now, if the UI was extended to really support dynamic views, this information could be presented in a truly useful manner. You could have something lightweight and auto-managed like a tooltip, but interactive like a window. And an OpenGL-accelerated UI could allow you to have such a UI while making it look good and consume few resources.
true but its not very essential at the hobby os level.
Well, you have to qualify the term “hobby OS.” SkyOS and Haiku are OSs that actually are trying to be something that people can use on a day-to-day basis. If advanced OpenGL UIs become standard, and really show their productivity advantages, why would anyone want to use a hobby OS that didn’t have those features?
“Well, you have to qualify the term “hobby OS.” SkyOS and Haiku are OSs that actually are trying to be something that people can use on a day-to-day basis. If advanced OpenGL UIs become standard, and really show their productivity advantages, why would anyone want to use a hobby OS that didn’t have those features? ”
i dont think they should be worrying about this at this point anyway. its not *essential* for these OS’es
Of course there are places in the world where 3D cards (At least the modern ones) aren’t commonplace at all. There might not be too many places where you’ll find a VGA-only system in 1st world countries (Outside of servers etc), but there are places where they just got their first computer yesterday, and it’s your old 386.
That’s a niche to be sure, and not the one (AFAIK) that SkyOS is targetting, but there’s definitely a place for a small, efficient OS that runs on the hardware of yesteryear.
Is the kernel interface for free? Please correct me if I am wrong.
I downloaded a pdf from NVIDIA showing off about how a large percentage of their driver is the same across platforms, just the OS specific stuff is different.
If NVIDIA do provide the linux kernel interface in an opensource format, isn’t it possible for Robert from SkyOS to modify it work with his OS?
All of what you said is true, and I’m typing this on a p166-thinkpad right now. But I don’t want to have a crappy user experience on my p4 3.2 ghz, 1 gig, ATI 9600 pro machine because some guy in Cambodia is running a 386-33.
There are some distros out there that cator to minimal-power machines. Vector Linux maybe, I forget the name. I think it would be an interesting project for someone experienced in the linux kernel to study the various 2.2.x, 2.4.x, and 2.6.x kernels and come up with a kernel that runs fast on a slow desktop. Strip out anything that is meant for a server and concentrate on a fast user experience. Maybe package it with gtk+1.x with those xft libraries(something I’m going to play around with.
Now something like Haiku or SkyOS is catering to teh desktop user that wants a smooth UI experience without some of the mess of a linux distro. The problem is that in 2006 or so when these OSs get mature, there’s going to be interesting stuff going on in the 3d accelerated desktop world of linux, windows, and MacOSX. I wish Nvidia and ATI would just open up their register-level specs, but because of intellectual property licensing and other things it doesn’t look like that’s going to happen. So that is a problem for those desktop systems.
I still see SkyOS as the pet project of one developer-just like AtheOS was with Kurt Skauen. Syllable has a team but isn’t the OS under the GPL? (the website isn’t very clear) If so then that’s another obstacle for getting drivers or any other proprietary technologies such as codecs. And Haiku isn’t even complete. So I don’t see how any of these OS’s can get ATI or nvidia until these issues are addressed
I’ve been following the AROS and Syllable projects. Mind you I like them both and am a fan of new OS projects in general, I have been since I ran my first Mac emulator on the Amiga many years ago. Up until now, I’ve been very clear about what each new project has to offer that others don’t. For example, during BeOS times it was particularly exciting, BeOS offered ALL SORTS of new features not before seen in a desktop OS — smooth, fair scheduling, with very low latency and high interactivity, beautiful multimedia handling, large file systems, fast booting, and it integrated some of the best ideas from other projects such as datatypes and so on. AROS offers AmigaOS compatibility, it’s basically an upgrade path for AmigaOS users and lovers, and it, like Linux, offers an open-source license. What does Syllable offer? I suppose it’s meant to offer an open-source OS that’s built from the base up to be a desktop system — I can see how this is very desireable. It also offers very fast booting. But unlike for example comparing BeOS and Windows, or BeOS and Linux, the differences seem to be smaller (between say, BSD and Syllable or Linux and Syllable). At this point, wouldn’t it be easier to modify existing open-source projects to give the desired features — such as fast booting and high desktop interactivity, than creating a whole new OS? The advantage of doing it that way, over writing a new OS is one of application compability. It seems that a port of KHTML is avaliable for Syllable, but of course the vesion I’m running now is much more recent. Even if you had to replace big parts of Linux or BSD to get the desired effect (such as replacing X-Windows with Y-Windows, for high speed and desktop interactivity at the widget level) or modifying the Linux kernel (for fairer or lower-latency scheduling) it still strikes me as easier than rewriting the OS, and then porting (or, rewriting) all of the useful apps.
This has come up with me because I’ve been experimenting with some things to make Linux more desktoppy on my Gentoo system. I’m using Nick Piggin’s scheduler, which makes things smoother and far more interactive, a carefully customized boot sequence which boots faster (though not as fast as Syllable in fairness), and so on. I’ve done some experimentation with Y-Windows and other alternative windowing systems that integrate the widgets into the server. It seems that rewriting the OS only gives you some (fairly small) advantage that can not be had by modifying what exists, with an enormous cost of compatibility and work. As for work, well, it’s a hobby, work away. But compatability and apps are more important.
Erik
On this page click on the link that says “NVIDIA Linux Advantage PDF Presentation”.
http://www.nvidia.com/object/linux.html
On page 6 it says that “95% of all code base is shared between all operating systems”. That remaining 5% is a Windows driver interface, a Linux kernel interface.
Again as I asked above is the linux kernel interface opensource? If it is can’t SkyOS and Syllable developers use that and modify to work for their windowing systems?
Yeah, that’s something I’ve been wondering about. The question is how hard would it be to reverse-engineer the binary part and see how it interacts with the kernel and X. Or dissassemble the binary part and figure out the register-level specs. Hard again and something that probably something that the SkyOS or Haiku guys would have trouble doing…but maybe something they’ll have to do at some point.
I agree with your arguments. One such project is B.E.OS and it is LGPL, Linux based with a BeOS application layer.
The only disappoint thing I see is that no one is trying to help its development.
http://www.blueeyedos.com/news/index.html#12_21_2003-lgpl
Is there an email client for SkyOS yet? That would seem to me to be a fairly important basic application.
I was basically arguing all of your points on a Haiku thread. Why not use a linux or BSD kernel instead of using NewOS. The best response I got was that the Haiku developers were kernel newbies and NewOS was a lot easier to understand. I still think it would be more productive to get up to speed on a BSD or Linux or some other established kernel that has drivers for it. And that’s the real problem – drivers. It’s a pain to get drivers for some hardware for linux these days and it’ll be 50x worse for these “alternative” OSs. That’s why (even though it doesn’t have much mindshare), B.E.OS has a good possibility of giving a good user experience with beOS apis. Plus the main developer seems to have a good understanding of getting X to perform well.
If your in the USA, it takes more then 2 months for a check to be sent over, and then your lucky if it even makes it.
Unfortunately, they also accept RipMe, I mean PayPal, but that is just as dangerous, as there is always the chance that the money will still not make it.. bah!
>If your in the USA, it takes more then 2 months for a check
>to be sent over, and then your lucky if it even makes it.
I find it completly stupid that you have to pay for a beta/alpha OS wich cannot run software or has not even got decent drivers. I can understand that you want help but asking money for it? You should be glad some people are trying to help, right?
Can we drop this Haiku/NewOS/Linux kernel argument already? I agree with some of your points, but the Haiku team have already chosen a direction and have made good progress. As you’ve noted, B.E.O.S. is already doing what you’re suggesting, so your argument is redundant. And finally, without speaking directly to the Haiku developers, you don’t/can’t know the real motivation(s) behind using the NewOS kernel, so your argument that “the best response I got…” is based on hearsay and conjecture.
I’m quite sure that the Haiku team have a pretty good idea of what they’re doing, and they’ve explored using the Linux kernel. Let’s wish them luck (or even better contribute) and see which approach works out the best (Haiku vs. B.E.O.S.).
P.S. I realize that you were responding to Erik, but it’s tangential to the article and leaves the same impression as an out-of-place cross posting (annoying and redundant). You’ve made your point, so let’s move on.
“I find it completly stupid that you have to pay for a beta/alpha OS….I can understand that you want help but asking money for it? You should be glad some people are trying to help, right?”
A few hundred beta testers would probably disagree with you….
At any rate, beta testers are not just paying for one beta. They get every single beta that is released between when they sign up and when 5.0 Final is released. They also get access to the beta discussion forum, where they can make suggestions, get direct help, etc. Finally, when 5.0 Final is released, they get a CD copy of 5.0 Final. So they are not just paying for one beta copy.
“….wich cannot run software or has not even got decent drivers.”
SkyOS comes bundled with a large amount of software, including Abiword, GIMP, audio player, MP3 Player, Quake I/II Demos, Minesweeper, and about 20 other programs, so to say that SkyOS cannot run software is incorrect. As far as drivers, all the hardware in my system works fine, with the exception of networking, which will be re-enabled in two beta releases. Most people have similar experiences. If beta testers’ hardware is not supported, they are always welcome to suggest to us to try to find a way to support it.
MP3 Player should say video player. ^^
Yes, I was responding to Erik and not to you Mr. Anonymous. And I’ll drop the conversation whenever I want, not when you suggest it Mr. Anonymous. I guess you didn’t even read the entire thread because then you would’ve seen how it had progressed to where Erik made his comments and its totally relevant because the same issues that affect SkyOS and Syllable also affect Haiku.
I’ve done a little research onto the reasoning of using NewOS – not just anectodal evidence, and yes, NewOS being a simpler kernel is one of the reasons (they are kernel newbies), and also they always refer to linux as a “server kernel’, which is bogus. The bottom line is that they don’t feel comfortable with the linux kernel, not because of some technical reasons. Let’s face it, they’re not going to get any better interactiveness because of NewOS than had they used Linux. The 2.6.x kernel already has some very low latency and it’ll only get better and with NPTL thread creation is very, very fast. I guess some people don’t like linux because its Unix even though its only Unix because its POSIX which haiku doesn’t have to follow.
I wish the Haiku developers all the luck and hope they have great success in their endeavors, but I just feel that they have some unecessary hurdles to overcome. Obviously they can work on anything they want, and any comments i have about it can go into /dev/null.
While it is true that the GUI isn’t the most important part at the moment, it is being worked on. A new Dock is being made, there will be new default wallpapers and stuff.
Maybe they can learn about kernels with NewOS and eventually borrow and adapt ideas from the linux kernel?
😉
“I keep on hearing about hardware-accelerated OpenGL for SkyOS. Does the guy that is writing the majority of the OS have a NDA with either Nvidia or ATI?”
This means 2D acceleration, not 3D.
We have been in contact with both nVidia and ATI. Both, while expressing interest in our product, regretted that they would not be able to either write drivers to support their products in SkyOS, or give us the specs we need in order to write them. One of the companies informed us that because all of their development is done in-house, it has never been necessary to document much of what they have built, so they could not provide us with the specs we would need. The other never supplied an answer to that question, other than saying it was not a possible option. It really is a shame, because with the specs we need, it would not take us long to implement full 3D hardware support.
That said, we will make due. Our hope for SkyOS 5.0 is to create a good, solid desktop for the average user. The average user does not need 3D hardware acceleration. They need a stable, responsive OS. They need access to the web/e-mail. They need a word processor. They need multimedia capabilities. All of these things SkyOS already has (e-mail application forth-coming).
We are realists. It is definitely annoying knowing that there are only two companies that make video cards (for all practical measures anyway), and that neither are willing to help people that wish to support their products, do so. That said, we will make-do. When we are able to afford or convince ATI and nVidia to work with us, we will do so. Until that time, we will make do with what we have, which is still quite a lot.
If you don’t play games..or only simple ones than you don’t need 3D support. The 90% of desktop users don’t require 3D support. if the 2D support is decent than for the majority of PC users it will be enaugh..or let’s put it this way..it can have a big market without 3D support. I also think that if skyos will have 0,1 % from the desktop market than NVidia and ATI will write drivers for it. look at linux..it doesn’t have more than 1% from the desktop market but it is supported by this 2 companys.
I think the problem is that you’re not really dealing with “average” users. You’re dealing with hobbyist type that are wiling to put up with bugs and lack of applications for something “different”. And that’s the key, especially since you guys are selling this thing. What makes SkyOS that much different than windows or linux or mac that they’re willing to plunk down $30 for an unproven OS that has limited applications. Obviously the ATI and Nvidia issues is beyong your control and unfortunate because I’m sure Robert could do some very interesting things with it, but the reality of the situation still stands.
if they would make a contract with a big company like dell or so(to be preinstalled on new PCs) than I am sure they will sell it even for 10$.. so you see the retail version costs 30$ even retail linux costs at least 2 times as much
And like I said, we are realists, and will work with the reality of the situation.
3D hardware acceleration is not the be-all, end-all of the desktop. The fact is, the majority of the users do not use their computer for gaming, or at least gaming that would require 3D hardware acceleration. They just need to standard list of applications, and a stable, responsive, easy-to-use operating system that ties them all together.
That is not to say that we will not strive to get 3D hardware acceleration support. When we are able to do this, we will welcome it with open arms. But until that day comes, why waste time and cause hardships worrying about something that, in the end, is not that big of a deal?
That is our philosophy, anyway.
You pay for SkyOS because people put time and effort in it. And in this world, time and effort cost money. By paying you show that you support that effort, and that you think it has potential.
If you don’t think SkyOS has potential, then simply don’t buy it! It is that simple! But don’t start complaining about it, because it then looks like you’re just trying to get a free drink.
We software buyers got used to downloading Linux and OpenOffice.org for free, and to pirate Windows. It might come as a shock, but the world isn’t free. No matter how you twist and turn it, money is *everything* in this world. I propose we all just deal with it. For the use of products, we must pay. Whether it be bread, cds, cars or software.
Maybe Robert should apply for a job at Nvidia :B
Or run my ideas past him. The ones about if nvidia provides an opensource kernel interface.
Or maybe he could investigate how their drivers work with the linux kernel, and try and provide a hack for SkyOS.
I also think that if skyos will have 0,1 % from the desktop market than NVidia and ATI will write drivers for it.
Nope, never happen and you’re forgetting that SkyOS users aren’t or won’t be average “users”.
Syllable has a team but isn’t the OS under the GPL? (the website isn’t very clear) If so then that’s another obstacle for getting drivers or any other proprietary technologies such as codecs.
Most of the code in Syllable is licensed under the GPL or LGPL. That does not mean that everything must be licensed under the GPL. Our driver architecture means that there is a very clear, very well defined seperation between drivers and plugins and the rest of the system. There is no reason why drivers cannot be written and released for Syllable under different, non-GPL compatable licenses. There is no barrier.
What does Syllable offer? I suppose it’s meant to offer an open-source OS that’s built from the base up to be a desktop system — I can see how this is very desireable. It also offers very fast booting. But unlike for example comparing BeOS and Windows, or BeOS and Linux, the differences seem to be smaller (between say, BSD and Syllable or Linux and Syllable). At this point, wouldn’t it be easier to modify existing open-source projects to give the desired features — such as fast booting and high desktop interactivity, than creating a whole new OS?
No, it wouldn’t. What you lack if cohesion. All the parts you bring together to make such a system do not fit together perfectly, and the feature set and usability will suffer. There are also just some things that you can’t change. We couldn’t have re-written the Linux driver interface to work the way that Syllable does, for example. The ease of driver management is a major advantage for Syllable.
You also totally lose your identiy. A Linux kernel, Linux userland and X is Just Another Linux Distribution. Yippee..we’ve not seen one of those before. Syllable is something totally different (The kernel and the high level APIs are at any rate).
The real goal of Syllable is the same goal as any other project “Be the best it can possibly be” I think we’re on track to do that with Syllable, and I think we’re doing it better than anyone else. Even if I am biased.
While we’re on the subject of OpenGL desktops, I think they’re horribly over rated. What does effects such as animations or sheared windows add to the user interface? Nothing at all. The practical applications of OpenGL can be done without 3D acceleration quite easily. To create a GUI you only need four primitve operations; set a pixel, draw a line, fill a rectangle and blit a bitmap. None of those require OpenGL. Stuff such as transparancy (alpha blending) and non-rectangular windows can be done in software with low overhead. Syllable supports both, but they’re rare. How many non-rectangular or alpha-blended windows do you see on your desktop right now? Why optimise for edge cases? So what if scrolling text in an alpha-blended round window requires 10% of your CPU; when are you ever going to do such a thing?
LOL “Mr. Anonymous”… I think it’s funny that you make a point of this, when really we’re all pretty anonymous here (Is Lumbergh really even your name? Who knows? Where’s your email address?). Post your picture up here, then go ahead and make your silly jabs about posting anonymously.
And your second paragragh is pretty much a rehash of what you’ve already said in the other article. But since you insist, here are some things to think about.
If Haiku were a new project, I could see your point about 2.6.x. But when did it start? BEFORE 2.6.x was released. At the time, the Linux kernel didn’t have the same low latency, NPTL was just getting started, and the OBOS team made a decision based on what was available. I don’t know… what would it take to change kernels?
What about boot times? BeOS boots in under ten seconds; I don’t see that with Linux. What about RAM usage? BeOS was pretty svelte.
A goal of Haiku is binary compatibility. Could Haiku keep binary compatibility or just source compatibility?
You mentioned X. Maybe the problem isn’t just the kernel but also the graphical system usually associated with it (namely X). I’m not talking just about speed, but about the capabilities of X. Can X handle where the Haiku team is going with Glass Elevator? BeOS was pervasively multi-threaded… is X equipped to handle that? Or perhaps the Haiku team could collaborate with the Fresco/Berlin team?
There’s more to Haiku than mimicking the interactiveness of BeOS. I think you need to think about the intentions behind Haiku. I gather that the goal of Haiku is more than recreating R5… Haiku is about the future (Glass Elevator) and carrying on where Be left off. BeOS was created with the intention of producing a legacy-free OS. In that regard, I think NewOS is the right kernel. Not a technical issue I admit, but something to think about (the original vision for Be).
“any comments i have about it can go into /dev/null”… Yeah pretty much sums it up for me too.
If you want to continue, let’s pull this thread to the article where you first raised the subject.
if drivers don’t have to be under GPL then whew that’s a relief but this whole GPL business still makes me a little anxious because I would like an OS to have the ability to incorporate proprietary technologies yet still remain open source for the most part (and yes i want an open source OS-for the most part).
In OTHER NEWS: remember Adrian Z from BeNews? He’s getting married today. Mazel Tov!
Something that people don’t understand is that the desktop is going to be 3d accelerated with Longhorn, Linux, and MacOSX, and these alternative OS that don’t have it will look like something circa 1982 in comparison.
I wouldn’t bet on it being a success. A 100% 3D desktop is just not going to be usable with the peripherals we have today i.e. mouse and 2D screen. Whatever happens, the desktop will, in reality, still be 2D. Trying to pretend otherwise will not work.
Let Robert figure it out.
OpenGL needs to be hardware accelerated for anything realistic to be done.
..this whole GPL business still makes me a little anxious because I would like an OS to have the ability to incorporate proprietary technologies yet still remain open source for the most part (and yes i want an open source OS-for the most part).
That is exactly what Syllable is. There is no reason why non-GPL components could not be used alongside the rest of Syllable without violating the GPL. The whole question of problems with the GPL is moot.
None of them will be trully 3D. We don’t have the peripheral to work in a true 3D environment, nor to see it. But they will all use 3D to accelerate the gui : windows are going to be flat texture surface for example. All 3D effect( a windows that glow to indiquate a new message for example) will be possible without any overhead on the CPU – and more : the gui are not going to be anymore dealed with the cpu.
So the gui will be faster, the computer too, and thing may maybe become beautiful(erhm. Remember webpage and flash
) – so everyone will be happy 
So the gui will be faster, the computer too, and thing may maybe become beautiful(erhm. Remember webpage and flash
) – so everyone will be happy 
So now I’ll be able to get stupid pop-ups advertising porn sites directly on my desktop? :>
(Already can if you have XP and the messenger(?) service)
What I am curious about is what happens when you’re using a 3d app (Eg glxgears for a simple example) inside a window? There’s only so much GPU time to go around (As gamers know all too well) so wouldn’t it just slow the whole desktop down? Do we really need people complaining about the framerate of their desktop?
Programmers being programmers I think they’ll hit the limits of the GPU pretty quickly, locking us all into another upgrade cycle. Only difference this time is that it’ll be graphics cards (NVidia/ATI must just be jumping with joy) rather than processors.
>We software buyers got used to downloading Linux and
>OpenOffice.org for free, and to pirate Windows.
We? maybe you but not me, speak for yourself please.
>It might come as a shock, but the world isn’t free.
It might come as a shock to you but the world is free.
Products cost money i have no problem with that i buy tons of software every year and have never ever said that everything should be free or gratis. GPL software is FRee in the sense as the world is free to..but you can SELL gpl software.
>No matter how you twist and turn it, money is *everything*
>I propose we all just deal with it.
You deal with it, for me money is a thing that help me to get comfort, stay in life its not the GOD for me as you describe it.
>For the use of products, we must pay.
we MUST do nothing. You can do everything.
we MUST do nothing. You can do everything.
Grow wings and fly :>
Why not sue Nvidia to open up their drivers since they collaborated with a criminal monopoly ?
Does make sense doesn’t it.
We have a right to tweak something if something breaks that we paid for. Auto’s come with specs so you can repair them .
I bet microsoft is behind all of this.
Why wouldn’t Nvidia want drivers on all operating systems unless a secret deal is in place.
There is almost always is one with these manufacturers dealing with monoopolies.
one word.
“support”
It costs time and money to test a driver out on hardware before its released. Then when that thing that inevitably goes wrong, “OMG my X server keeps crashing” they have answers or have to come up with a new driver revision. which would be a pain in the ass if they started supporting every OS that boots. Point is, when things break people want answers and the want results yesterday. They are unreasonable, greedy, and impatient. I’m just glad that enough linux gimps are in the NVidia fold to get linux drivers.
I doubt very seriously that Microsoft is doing any bribing at all with them and drivers, its simple economics that do the work for them.
PS they aren’t going to open up their drivers because the 3d graphics biz is pretty tight these days between ATI and Nvidia, they don’t want to give the other guy the inside keys to their hardware… I mean why help the other team reverse engineer their product?
Capitalism at its best.
I concur that a 3d desktop is not necessarily going to be a success, but that’s not the idea. Notice I said 3d-accelerated – as in levereage those 40 million transistors in your GPU to do tranlucency, animations, etc…while freeing up some CPU cycles. You could get cool bump-mapping, shadows and stuff. For some reason when I look at windows with shadows it just makes a lot of sense. Modern card 2d functionality is just a subset of 3d functionality.
“I mean why help the other team reverse engineer their product?
Capitalism at its best.”
they could very well patent it and grant a right for gpl’e code base. you cannot reverse engineer when a patent is in place. the company would just sue the hell out of you
they could very well patent it and grant a right for gpl’e code base. you cannot reverse engineer when a patent is in place. the company would just sue the hell out of you
Nvidia’s reason is that the driver code gives away some of the hardware design. The GPL gives no protection whatsoever when it comes to preventing someone using information gained from the code to replicate Nvidia’s hardware features. Even RMS would have a hard time arguing that the resulting graphics card contains GPL code. If their driver really does show such information then it makes perfect sense for them to keep it closed, same goes for ATI.
“Nvidia’s reason is that the driver code gives away some of the hardware design”
i also mentioned patents. hardware patents can prevent people from using that information so its not a good argument. combined with gpl’ed code neither the code nor the specs are usable by competitors
How do you prove the other guy is doing it though without illegal insider information “ie corporate espionage”?
That’s the trick, do you think for one minute that intel isn’t reverse engineering AMD chips and vice-versa? It’s all in who can prove what. Lawsuits and legal battles can take a lot of money and time that neither of the said companies can afford, if they did they’d very swiftly lose their edge. They make money by keeping their respective companies lean, not with lawsuits similar to SCO.
Hardware patents don’t do a damn thing because they aren’t going to copy the board transistor for transistor, they’re going to see how your tech works and make theirs better.
“Judge, they stole our specs because their product does what ours does, better….” yeah that arguement would last a long time in court.
”
That’s the trick, do you think for one minute that intel isn’t reverse engineering AMD chips and vice-versa?”
sure they do which actually adds strength to my point that nvidia isnt preventing anything from binary drivers since everybody can reverse engineer anyway and the products are so different between nvidia and ati that its not so much useful
..do you think for one minute that intel isn’t reverse engineering AMD chips and vice-versa?
They don’t need too; AMD and Intel have cross licencing agreements so they can share each others IP. Although I get your point.
However it’s a bad argument. If your business is built around reverse engineering and cloning someone elses chips, you’re always going to be 1 to 2 years behind them. nVidia and ATi have to work on new designs all the time, and by the time the consumer gets their hands on them the design has been finalised for months and the R&D is going into the next version. How would it help ATi to figure out what is in the NV40 twelve months after it was released and with the NV45 about to come out? It wouldn’t.
I saw it mentioned in the first 15 comments, then it just sorta went away after Eugenia [incorrectly] mentioned that it costs a lot of money to port SNAP drivers to an OS. In fact, it costs nothing to port their drivers, and it’s not all that much work. Unununium, a hobby OS did it in a few days. The downside to SNAP drivers right now is that you only get VESA drivers for free (and VGA) unless your OS is Linux. They are currently working on this though, and I’m confident that the problem with be solved in not too long. By the looks of it, they’re also working on 3D stuff too.
“regretted that they would not be able to either write drivers to support their products in SkyOS, or give us the specs we need in order to write them..”
It’s just plain sad. Another example of hardware manufacturers being too busy brown-nosing Microsoft to deal with another OS. They’ll do ANYTHING Bill tells them to, but cooperate with a new OS on the scene… forget it.
when is the product going to be officially released