Slashdot Mirror

For most of those you seem to be missing the point. Haiku is designed to be a successor to BeOS. Many BeOS users (current and past) have used or now use other OSes, but haven't seen the things they liked about the BeOS present in them. That's why there's a need for Haiku. There is the Cosmoe Project that seeks to make Linux more like the BeOS, but personally I'm placing my hopes with Haiku (although I expect Cosmoe & Haiku can improve each other). That said, there are a few of your comments that I'd like to address.

* Focused on desktop, don't want to be a wristwatch
** So they believe in the future of desktops and not having a system that's built up of components but having a system designed for the desktop. Apple have their desktop OS running on a phone, Linux has been ported to just about anything under the sun, Microsoft have a stripped down version of their OS for phones and PDA but Haiku think they are better by focusing on desktop only -- mistake!

Haiku is designed to be a desktop OS. Not a server OS, or an embedded OS. Other OSes do that, and they do it well enough that there isn't a whole lot of need to compete with them in those areas. There are also trade-offs, so Haiku is aiming to be specialized toward the Desktop, not a Jack-of-all-trades.

* Compatible with Beos R5
** As he said in the presentation, why focus so much on being compatible with a 6 year old OS? - Maybe an emulator for the sake of it, but this is a priority?

Because much of the potential userbase is currently running BeOS R5.0.3 or R5.1 or Zeta. Many parts of Haiku have even been ported (if necessary) to run in BeOS R5. Parts like the Mail Daemon and ShowImage. Binary compatibility also ensures a fairly easy transition from BeOS/Zeta to Haiku, and that means that Haiku will have a fair amount of software available immediately. Many older BeOS apps are also closed-source (with the authors incommunicado), so binary compatibility is kind-of a big deal. Haiku R1 aims to be very similar to BeOS R5, Haiku R2 - Glass Elevator may break application compatibility.

* Kernel designed for responsiveness
** Low latency means lower performance and that dig at Linux he made in the presentation is inaccurate - firstly, responsiveness depends on options chosen in the kernel, having the option of better performance over lower latency is a GOOD thing.

I assume you haven't used the BeOS. Its speed will spoil you. Whenever I have to use Linux and Windows the speed difference becomes obvious. It isn't major, just annoying. Things like folders taking one second to open instead of a tenth of a second. Or applications that start essentially instantaneously. As a desktop user I don't really care that such responsiveness causes a small hit in performance, I'm not running a MySQL database or serving webpages in x milliseconds.

* Small footprint, fast boot
** 60MB uncompressed is not small, Linux even with X can be as small as 6MB uncompressed if that

True, Linux can be small, but the desktop distros usually aren't. They also tend to have a boot time on the order of minutes, while BeOS (and I assume Haiku) has it on the order of seconds. I.e. on the same machine it's not uncommon for BeOS to boot in ~7 seconds (ready to launch Firefox), and Linux to take 2 minutes. I'm sure you could reduce the Linux boot time, but even on my GP2X it still takes 14 seconds, and that's not loading network stuff or other things.

* Less Debug - no need to test with FreeBSD, OpenBSD, NetBSD, Multiple Linux kernels, Windows
** Absolute bullshit, he seems to encourage software to be made for their OS only, making it compatible with Linux, FreeBSD or Windows is a sin!

Yet another way BeOS is different. BeOS isn't entirely posix compatible. Virtually all non-trivial programs are multi-threaded, and BeOS uses it's own version of threads. It's related to the responsiveness of the B

In that case you need to wait for BlueEyedOS or Cosmoe to come along. Those are BeOS API ports on top of the Linux kernel. Meaning a recompile would suffice to get the BeOS apps together with Linux driver support. A worthwhile effort IMO.

Also check out Cosmoe, which is a port of the AtheOS API (similar to BeOS's) to Linux. I think it's a really cool idea that is not getting the attention it deserves. It shares some code with HaikuOS, so you can help both by helping one.

Of course there is Cosmoe , another AtheOS fork to which uses the linux kernel. That really is aiming to clone BeOS tho.

Syllable is getting more press than the other AtheOS fork Cosmoe.

And, of course, there's Syllable, which works the way you want it to (a BeOS-like user-land on top of a Linux kernel.) So it's not as if OpenBeOS is preventing you from getting the OS you want.

Some people here seem to not know, or be a little confused about how this company named yellowTAB came to be the BeOS people, and what is happening in the BeOS world at the moment. Also, I'm not associated with yellowTAB at all, this is all my observations.

yellowTAB
Bernd Korz first appeared in the BeOS community when he released a magazine named InsideBeOS. Only one or two issues were translated into english from the native German of the magazine (I purchased the first english edition!). This magazine stopped publication sometime around when Be flopped and as far as I saw, Bernd sort of disappeared again. When Be filed for bankruptcy, there were many outcries from the community for the source to BeOS so that it's legacy could continue. There were several formal proposals from people to purchase the source as well. Then came the announcement that Palm had purchased Be's intellectual propert(IP)... and were not going to continue development of BeOS. Luckily (in my opinion), Bernd was smart enough apparently to ask for a license to the source and not to purchase it. So, before Be sold their IP to Palm, they gave some sort of full license to yellowTAB to release new products based on the source code. Presumably that license just transferred to Palm's ownership with everything else.
So, here we are today, yellowTAB is about to release their new OS named Zeta (sort of BeOS R6) which contains unreleased code from Be plus new stuff and bug fixes that they have added. The article does not make it very clear, but yellowTAB is actually selling RC1 (Release Candidate 1) of Zeta now to the world. The R1 (Release 1) will be available to people who purchase RC1 for 10 Euros.

BeOS rewrites
There are a number of projects that are aiming their sights on rewriting the BeOS. Some totally opensource from conception, others planning on releasing their source after they have a public binary release under their belts. The main contenders that are around at the moment are OpenBeOS (soon to have a name change), BlueEyedOS , and Cosmoe .

OpenBeOS is taking the approach of totally rewriting the entire system. New kernel, new appkit, new interface kit, new storage kit... everything. They're not changing the basic structure of the system, in fact, they're attempting to completely duplicate the system even down to binary compatibility. As a basis for their project, they have used Be's old header files, and written API documentation. They are hosted at source forge and are using the MIT license for all their source.

BlueEyedOS is taking the approach that they perceive the Linux kernel and XFree86 as suitable for building less traditional system on top of them. They are using the Linux kernel as their Be kernel, and X windows as their display. They've released a demo CD so far that is quite interesting and appears very promising. Their code is not available to anyone but their developers at the moment because they've chosen to start their project closed. I've believe that I've heard though that they are planning for a source release later on. The demo CD can still be downloaded from their website.

Cosmoe is a primarily one man show that has taken the Syllable (was Atheos) source code and altered it to run on top of Linux and X. It's author's primary goal seems to be to give the Unix, Linux community a new user interface and the Be API. Code for this system is occasionally released under GPL and LGPL.

Wasn't AtheOS the OS that was all being done by that one guy and had the amiga-like GUI with the nice c++ API?

Yes, the same. The main developer stopped working on the project (at least, stopped posting to mailing-lists and the such), and after some time, people decided to fork Atheos to make syllable. It seems most of the former atheos community has moved over to Syllable. Another similar project used the Linux kernel with a BeOS like interface on top, this is Cosmoe. I am not familiar with the latter, though.

I think this is a very good idea. I'm sure some of the former (current?) BeOS users will have a bond with this filesystem layout. Beos had a layout somewhat like this and I loved the way it was easy to navagate. Hey, it could even help out some opened BeOS projects like B.E.OS and Cosmoe

Just noticed today that it has been accessible for download from its website for a while now. Going to take it for a spin tonight...

Regards,

Michel

How many things in X will we need to fix?

Looking for this? I personally think X (or at least XFree86) is outdated at the core. This has been solved by adding numerous extensions, but that doesn't really solve the problem. Can an application programmer count on the presence or functionality of extensions on any system? What we need is a new design, based on the experiences with X.

I like network-transparency (it's a huge win in some situations), but X uses far too much bandwidth. What about a widget-based approach like PicoGUI? Add window overlapping and implement smoothed fonts on server-side, then write a rootless X-server for it, and I think many people might have a shot at running it on their desktops.

And Cosmoe seems to be working and coming along quite nicely too.

When are we going to see a new release of BeOS?

Never. BeOS is dead. Perhaps you'll see parts of the API in PalmOS 5.x, but there won't be any new release of BeOS. IF you want a BeOS like enviroment check out Cosmoe OS (it runs on top of Linux).

-adnans