OSNews had the privilege to play with the upcoming B.E.O.S. bootable CD recently. The test/demo CD is based on Knoppix but it runs a BeOS-like app_server and toolkit on top of XFree86 and it is carefully optimized for extra speed.While we found a few bugs in this test version, the fact remains that CPU utilization is pretty low and the window movements are smoother than any other X11 window manager/envinronment we have seen so far.
On the second part of the demo, you get to use a “real” test app (as opposed to a bitmap mockup window in the first part of the demo) and to play a Puyo-Puyo puzzle game, named BePuyo (both apps are based on the BeOS API and were recompiled for B.E.O.S under Linux).
The ISO will become publicly available some time in the very near future, for everyone to try it out. The ISO currently weighs 101 MB.
Beos on Linux? Doesn’t sound that great… But then I haven’t a clue and I like to remain ever open-minded. My interest is certainly sparked though, can’t wait to give the demo a shot and see what B.E.O.S has to offer. One thing is for certain though… I’d love to see beos with linux’s hardware support. That’d be great.
Btw – it almost sounds like a window manager with compatability layers for beos, at the source level anyway. Can anyone confirm/deny/elaborate?
No, it is not “BeOS” (as we know it) on Linux. It is just a toolkit and a backend that looks like BeOS and is built on top of Linux technologies, so BeOS developers can port their BeOS apps to this new system with minimal changes. More over, it will have a new desktop environment, based on the toolkit.
So in essense, it is nothing but what GNUstep and WindowMaker did in order to ressurect the “OpenSTEP experiene” under Linux. Same for B.E.O.S.
But this effort is still a long way off… Lots of work needs to be done. It is not even alpha yet…
It seems like the most realistic approach anyways, Linux is gaining huge ground, why not make use of all the attention it’s getting?
Well the website has improved a boatload.
So how complete are the kits?
Looks neat… but are they gonna change the name?
So BeOS is resurrected again, this time combining the best of 2 worlds. Linux Kernel and power and all that hardware support we dream of, with the grace of BeOS and look & feel.
Congratulations to B.E OS team, and congratulations to the world, finally a milestone in the BeOS history.
Welcome to BeLand Penguin dudes!!!
Now the website’s dead….strange no?
Are those things in there?
If B.E.O.S is not even at Alpha stage, why does their download page list it as “Beta 1”
right now it’s reiserfs.
but I guess if rw support for BFS was ported to linux, there would be no reason not to.
Note that when you’re testing blits, it’s much faster to copy a pixmap to a window rather than using XPutImage. On my setup (43.49 NVIDIA drivers on a GeForce4Go 440) copying a pixmap to a window is fast enough to drive that card nearly at over 2GB/sec, which is about a third of it’s full memory bandwidth. I don’t know anything about GUI programming, of course, so I don’t know if it’s practical to cache all these pixmaps server side. However, if you’re doing something like a 2D game or something, creating a server-side pixmap is the equivilent of using a DirectDraw or SDL surface.
> Combining the best of 2 worlds
To me OpenBeOS is a better solution, as Linux doesn’t have much in common with BeOs kernel-wise. Anyway I welcome this.
Though I think 100M is a bit big just to show a window and to play BePuyo
but I’m sure it will get better soon.
> wing
Problem is, the linux kernel knows nothing about attributes and queries, so even though other filesystems like XFS supports them, it can’t make use of them. Or you need ugly hack to bypass the kernel like some apps I saw on frehsmeat.
There’s for sure a better choice for Kernels out there…
MetBSD for example…
Its known for having a *GREAT* design!
And theres XFREE w. DRI available too
There’re emulation layers even for MacOSX (in progress)
-A
(LinuxLinuxLinux)
Upps. Sounds like “Mettwurst”…
I wanted to day NetBSD of course!
-A
Just getting a mirror for this setup at BeForever:
http://www.beforever.com/
I don’t know when the downloads will start (soon!), but our server is in *Omaha, NE*.
Since we’re posting about the mirrors.
I am waiting for the “go” sign from guillaume also.
http://signal.fearmuffs.net
and i do hope that we can let it rip soon ;D oh and since peter posted a geo-plug, this server is in Norrköping,Sweden
Sigh, could you guys please wait posting news like this until the mirrors and such are up and running?
Waiting isn’t fun, I’d rather not know about it than having to wait
but then where is the longing passion of just running it later on then richie ol’ boy ;D
cheers!
Hi Dave!
Because is our first BETA release of a boot CD, only it. Tomorrow, maybe, we will be in a RC hehe!
But of course, we are in a alpha stage.
Michael Vinícius de Oliveira
~ BlueEyedOS.com Webmaster ~
>Problem is, the linux kernel knows nothing about attributes
>and queries, so even though other filesystems like XFS
>supports them, it can’t make use of them. Or you need ugly
>hack to bypass the kernel like some apps I saw on frehsmeat.
That’s wierd…. The kernel I compiled just today has Ext3 Extended Attribute support. You may be right about the quieries, but I’m tentatively calling BS on the EAs…
May I point out a solution to the bandwidth problem? http://bitconjurer.org/BitTorrent/ works like a charm and eases the load on the servers.
so what’s this anxiety before the release, don’t worry you’ll have your torrents…
Use linux kernel is a great idea. The original BeOS is almost dead (by lack of new hardware support) and use a new microkernel like OpenBeOS is a bad idea because there will never exist sufficient hardware support for a desktop OS.
With some tricks like realtime patches the linux kernel is a viable solution for a multimedia desktop OS. B.E.O.S can be the true “linux for joe users” (users that cannot learn Unix philosophy).
OpenBeOS and the other projects will be restricted to a niche like the original BeOS, just a geek’s toy OS…
Eugenia, having had the privilege to play with the upcoming CD, I guess you have a good connection with the B.E.OS team.
Having in your background a lot of BEOS knowledge, how do you see this approach going?
What are your feelings about this project? Is this going to be a successful project?
Apart from that, I feel a bit disappointed about the progress in OpenBEOS. I usually check their webpage and I hardly read any interesting piece of new showing their progress. Perhaps they don’t upgrade the web, or perhaps the development is not going to be as rapid as they claimed.
I really enjoyed using BEOS in the past. It’s a shame but I’m not very confident of having anything similar soon.
In any case, thank to the developers for their effort.
So, only when all those blue bars on your ‘status’ page reach 100% will B.E.O.S be ready for Beta? I guess even when that happens there still won’t be a tracker, deskbar, etc. Is that right?
I love the idea of this project. One other question though, Will I be able to run other X programs (such as OpenOffice!) using your custom X server?
…will this run under BeOS. All you infadels will be sorry once the great OS rises out of the ashes and rules all yall!
http://www.freelists.org/archives/openbeos/04-2003/fullthread9.html
there is some update info here. I think, it is not as silent as it seems from the webpage.
here is what michael phipps had to say:
http://www.freelists.org/archives/openbeos/04-2003/fullthread10.htm…
B.E OS is probably on a short basis what it’s all about. Drivers is a problem which B.E OS has a solution to for sure. Very nice to see the job-
However in the long run OBOS will make it and will be the optimized solution. Like someone noted, updates on webpages is indeed slow, but under the hood a lot of stuff is cooking. Here is where it stands today I think…
App/Interface – Completed almost, will be alpha any week now
Filesystem – Completed with smaller bugfixes left
Gamekit – Not very important at the mo’ and not big task neither. Smaller progresses being made
Input server – Partially working, smaller implementations left.
Kernel – Rapid development, but a lot more to do
Media – Close to Alpha and very up to date technologywise
Midi – Completed with smaller bugfixes
Networking – Progress being made currently lacks devpower (Updated netstack alot better than Be original)
Prefs – Completed almost for guiapps. Small commandline apps not priority
Printing – Completed, bugfixes and driver support left
Screen Saver – Completed
Storage – Completed almost, currently passing featurewise beyond BeOS
Translation – Completed, add-ons under development
I’d say OBOS has come a VERY long way and is hardly dead. Hardcore kerneldevs are what’s missing… where are you?
While this baby finishes it’s cooking, enjoy B.E OS, ’cause it’s gonna be great too.
If I wanted a proprietary OS I would use Windows.
is it able to really reproduce the BeOS experience when it comes to the pervasive multhithreading aspect. That was the point that i thought made BeOS truly unique and is still unreplcable in other kernels? The super-responsive feeling is due to this and made even the slowest system seem like a speed demon.
Is that real-time patch able to do this?
I think that with all the talk about which Linux Distro would be best suited for the desktop, B.E.OS seems to be overlooked. I hope that changes though. With all the praise that the BeOS interface got and continues to get, I think it just makes sense that Linux gets an easy to use interface for Joe-Desktop user. B.E.OS also addresses one of the biggest complaints – that Linux is slow (I know it isn’t slow, but the interface to Joe-Desktop user does in fact feel very very sluggish). The biggest advantage to Joe-Desktop user with B.E.OS is the driver support. Joe-Desktop user can just turn on the computer and not have to worry about drivers or any of that technical BS (although if they purchase a comp that has B.E.OS pre-installed – which is the only way that I see to get market penetration – they wouldn’t have to worry about drivers anyway).
The role I see for openBeOS is exactly what BeOS was intended to be, a “geek OS”. The drivers will be hard to find and the system will likely not run much (precompiled stuff anyway), but anyone interested will be able to set it up, compile and app or two, and be off and running.
So I think there is room for both to play nicely together.
It would be interesting to compare this technically, idealistically and look-wise to Cosmoe. Any chance of a progress report on Cosmoe too some time soon?
… with this latent “Knoppix is better”-bla-bla…? Which part of Live-EVALUATION people don’t understand? Looks to me like it’s the evaluation-bit… The CD is meant to give you an impression on how the real think will look + act like — not more, not less. So finally please quit comparing it to a specifically HD-less distro. It has nothing in common with it altogether and it doesn’t want to.
This weekend, everybody here will be able to download the iso, so give us a break, and wait for it to see it for real.
The CD is not supposed to be an OS, just a demo. It contains 90% of drivers.
B.E.OS is not as much ‘pre alpha’ that said.
The app_server is working. To be able to launch big apps
like opentracker, irc client, email client etc… it’s just
a question of time, because we have to complete the storage kit, network kit and some specific things in the interface kit.
Regards,
Guillaume
because it’s the bridge between the BeOS world and linux coder. Just like BeOS made a bridge with Atari, Mac and Amiga user.
Even if that OS only interest a fraction of the linux community it will be already great as app for it could be ported to real beos. It’s curently like that with many terminal app but it would be nice to have more GUI app.
Then ,let say you have a music app that support X feature because the driver exist, it will be a big reason to do a BeOS driver if an app already support it! It can establish some kind of gravitational pump that fill the BeOS world over time. The BeOS always fighted to break the egg and chicken paradox.
How much memory will this take with the GUI, would a dreamcast port be possible?
Even if that OS only interest a fraction of the linux community it will be already great as app for it could be ported to real beos.
I think that it’ll be the other way around. This si a good project because it’ll allow for the production and design of Linux apps through BeOS API kits. To me this will make developing apps for Linux much easier. You get the to make apps in BeOS style but have all the hardware support of Linux.
If you have an original Be app written with nothing but the Be API Kits then porting it should be very easy need little or no modification.
why dont they just use bittorrent to spread the iso ? its the obvious solution
…when I’ve got a dedicated computer multi-booting BeOS, OpenBeOS, B.E. OS, Yellowtab, and maybe Cosmoe, just for the heck of it.
Woohoo!
linux for all…
The iso will be available on bittorent as soon as the iso is published on the website (ie tomorow).
Regards,
Guillaume
The iso will be available on bittorent as soon as the iso is published on the website (ie tomorow).
Sweet, I am pretty looking forward to play with a new toy this week and weekend. 😀
using the BeOS logo for news posts on “trying to be similar to BeOS, but not” products is misleading, IMO. It should be changed.
I think the project should change it’s name, I always seem to read it as ‘Blue Eye DOS’ 😉
I have just tried the demo and I have to admit that things feel more smooth..
But it is still not as fast/responsive as it is on Windows (yes, Windows, not BeOS).
(it is close though)
(oh, and my system is fast enough)
Still nice work though
Rich,
Do you speak about the first part or the second part?
The first part is from an old code, before the
merge of our native rendering in the app_server.
The new version is faster, and when compiled with
-02 or more, the cpu usage falls.
I personaly made tests with GeForce4 and Radeon based
systems, when XFree use the ‘radeon’ driver or ‘nv’ driver,
the result is faster than under WinXP. But I saw that
some card are not well autoconfigured and use ‘vesa’ driver.
Look at the autoconfiguration result.
Regards,
Guillaume
I was talking about the 2nd part and it used the nv driver (which is not really that fast anyway).
The video card I am using is a GeForce3 Ti 200.
“I was talking about the 2nd”, it’s slow because it use the X11 rendering, the first part use our native rendering.
The app_server is capable to provide the performance showed on the first part.
To sum up, the embedded app_server is running in the X11 mode, it means, that it uses a real X11 window manager. This app_server is able to render everything in our native windowing system, but due to a lack of time, we preffered to use the X11 rendering mode.
Regards,
Guillaume
Hi,
Guillaume, I just want to let you know that my Hercules Radeon 9000 gets recognized, but the Vesa driver gets loaded anyway …
‘BE the difference that makes a difference’ – JEWEL