Back when it was becoming clear that the time of the BeOS had come and gone, enthusiasts immediately set up the OpenBeOS project, an attempt to recreate the Be operating system from scratch, using a MIT-like license. The project faced difficult odds, and numerous times progress seemed quite slow. Still, persistence pays off, and the first alpha release is drawing ever closer. We decided to take a look at where Haiku currently stands.
The Be operating system and the myths surrounding the scene have always intrigued me. I’ve used BeOS for a long time, and I still consider it to be the best operating system ever made – when put in perspective, of course. Windows was still crashing and burning every three seconds, and the Mac OS didn’t even have memory protection. All in all, the state of the mainstream operating system was abysmal. However, there came the BeOS, descending from heaven while riding a chariot of fire, pulled by magnificent stallions – and all was good. Sure, it didn’t do multiuser (but neither did Windows and the Mac), and it was terribly afraid of anything having to do with networking and the internet, but who cared?
It was lightning fast, very stable, easy to use, and for the first time, it was an operating system that made sense. Installing an application or a driver amounted to little more than dragging and dropping a file/directory in place. It had an instant-search system based on metadata that could be used to perform all sorts of tedious tasks other operating systems (still!) need separate programs for. But most of all, the BeOS had something that no other operating system ever had: personality. BeOS felt alive. If you weren’t part of the BeOS crowd back in the day (and I was pretty late to the game too, 2000 or 2001), then it’s pretty hard to understand what is meant by a piece of software having a personality.
We’re in 2009 now, and the Haiku project has come a long way. Many were sceptical about the project’s chances at success, but they developed onwards anyway. Now that the alpha release is getting ever closer, it also becomes more apparent that the Haiku team has some very big shoes to fill. BeOS fans are a picky, whiny bunch.
Does it work?
Haiku works. Actually, it works so well, it makes you wonder why they haven’t declared the project alpha already, and why they haven’t put out a nice .iso image for everyone to enjoy. There are a number of reasons for this, but first, let me focus on the things that Haiku already does well.
Haiku boots in an instant, just what we BeOS fans have come to expect from our little pet operating system. The same row of icons still serves as the boot screen for Haiku, and hardcore BeOS fans will know what the icons mean and what happens during each blinking icon in the boot process. Whether or not Haiku employs the same boot order is something I actually don’t know. Still, it’s a “feast of recognition” for BeOS fans.
The current pre-alpha builds boot into a desktop that is instantly recognisable by any BeOS user. Sure, the icon set looks more modern, and Haiku has anti-aliased fonts, but for the rest it’s pretty much r5 all over again. There is a big hint, however, that we’re dealing with an operating system that is in heavy development, and not by any means ready for general consumption: Haiku loads a terminal window by default. Welcome to the Haiku shell, it says.
You can choose to work with the ‘vanilla’ Haiku image or build, or opt for the ‘pre-alpha’ build which includes a number of applications within the disk image. With r5 compatibility being one of the pillars of the Haiku project, getting applications to run is paramount. In any case, Haiku runs important applications like Firefox without a hitch, and the same applies to BePDF, Wonderbrush, StyledEdit, VLC, Vision, and many more. Stability seems very good application-wise; I haven’t encountered an application crash yet.
That’s all fine and dandy, but how well is the operating system itself doing? Sadly, most of my testing is done in a virtual machine (more on that in a minute), so I have little to say about how well Haiku runs on real-world hardware – reports from developers and users in #haiku, however, indicate that stability is excellent, with a number of developers using Haiku full-time. Not only for development, but also for day-to-day usage. In any case, in my VMwae virtual machine Haiku has never crashed. I’ve had multiple applications running, and have yet to encounter any problem. My most memorable experience so far was clicking on a PDF link in Vision, which automatically launched Firefox, which downloaded the PDF file and launched BePDF. This just goes to show how solid Haiku already is.
Problems
Still, there are problems. The biggest problem I’ve encountered so far are the difficulties concerning installing Haiku on real hardware. Even though the installer is working just fine, the DriveSetup [the partitioner] is not. All it can do at this point is initialise an existing partition as BFS; it can’t create, delete, or move partitions. This means you will have to create one by other means first. The installer also does not make the partition bootable; you will have to run the makebootable command from within Haiku first.
This is where I ran into problems. I used dd to turn an USB stick into a bootable Haiku installation, but for some reason, getting it to boot was a process ruled by chance. Sometimes it worked, sometimes it didn’t. I had Haiku installed on a partition, but was never able to get as far as to run makebootable from the Haiku system running off the USB stick. This might very well be my own fault, since many Haiku developers manage just fine this way.
In order for Haiku to be proclaimed “alpha”, I strongly advise the team to make sure the installation procedure is fully functional, so that aspiring users do not run into any walls when trying to install Haiku. Back in the BeOS heydays, operating system enthusiasts weren’t put off by some hard labour – but in this day and age of virtual machines and ridiculously easy-to-install Linux distributions, alternative OS projects cannot afford any install problems.
There is also a list of bugs that the developers want to have fixed before they want to move Haiku to alpha status.
The other side of the spectrum
When KDE 4.0 hit the streets, many complained it wasn’t ready, and that it should have never been released as a final release. The reason KDE 4.0 was released as-is had to do with the KDE team wanting to attract enough users so more bugs could be found; they wanted to avoid being in eternal beta. Sometimes, I get the feeling the Haiku guys are in danger of reaching the other side of the spectrum KDE 4.0 was on: eternal alpha.
During my short tests and my time in #haiku it became clear that Haiku is pretty much ready to go alpha – installation problems aside. In fact, some Haiku developers personally believe Haiku should move to alpha as well, and I find it hard to disagree with them. Still, I understand their hesitance; the project is burdened by huge expectations, and they want to avoid failure at all costs. Still, at one point, you have to let your baby make its first steps without holding its hands. Sure, it’ll topple over some times, but it will only learn from that.
Haiku is ready to move to alpha. Sure, it’s not ready to take over the world and feed your cat, but it sure is ready to be labeled as alpha. Exciting times ahead, boys and girls: BeOS will be back.
GCC 4
As many Haiku community members know, one major hurdle that has been making it difficult to port new software to Haiku has been the lack of an up-to-date GCC4 compiler. While a GCC 4.1.2 cross-compiler has been available now for some time, cross-compiling software for a GCC4-built Haiku can be painful and frustrating. What Haiku really needed was a native GCC4 toolchain to run on a GCC4-built Haiku install. It has one now.
They just keep moving along, and it really does show the time thats being put in. all in all i too am rather impressed with development, but what it needs is better marketing. But that will likely come when the software is Beta, and that is still a ways off. I haven’t tested it on real hardware either but in VirtualBox it runs fine. only 1 crash so far.
I was running Haiku up to a year ago on real hardware with very few stability problems. Of course, the system became unstable under certain circumstances but that’s exactly what one would expect from pre-alpha builds.
The SATA issue is what stopped me becoming more active in testing back then and currently still prevents me from booting on my hardware.
I’m not entirely sure what happened to the driver architecture back then, or whether the changes to it introduced the problem in bug 1444 but I can’t wait until that bug gets squished.
Well done Haiku team and everyone involved in the project. Can’t wait to get my Haiku fix all over again…
Is this project meant to (eventually) compete with Windows, OS X and Linux-desktops or will be only something for geeks (forever)?
I’d love for desktop-os in general to be more competitive and more innovation happening.
From the screenshots, it looks like it is meant to compete with Windows 98 at best. Pitiful GUI compared to KDE, GNOME, XP, Vista, OS X.
Which means you have not tried using the GUI.
OpenTracker (the desktop) has features that I never see in a Windows/Linux/Mac desktop, and they are just a click away. Just because the desktop is not full of worthless flash and eyecandy does not mean the power is not available.
(emphasis mine)…
See, that’s the problem – looks aren’t everything.
To me, it looks simple and easy to use – unlike the modern UIs of Vista, KDE, GNOME, etc.
I’m that guy that puts my Windows machine back into “Classic” theme so I don’t feel awash in eye candy that makes me want to vomit when using a recent OS.
And OSX doesn’t even appeal to people like me – it just looks yucky. Linux can be made tolerable if necessary.
I’m that guy, too.
I’m also the guy who’ll try never to get a Microsoft OS beyond WindowsXP. Will I succeed? I don’t know, but to be fair, I have absolutely no need to Vista/Windows7.
Yes, as Thom pretty much accurately portrayed in his story, installing Haiku on native hardware for the first time tends to be a huge hurdle for the influx of noobs and curious people who want to try it out.
On the one hand, it’s a feeling of accomplishment once you get it done, but I often find a lot of people wandering off in disinterest when they fail to get it booted on their hardware within a short window of time (maybe ~1 day maximum is what people generally devote to the attempts).
Sometimes they’re lucky and the usual suggestions work, sometimes their hardware is a bit finicky, doesn’t want to boot from USB, requires a swap of the bus_manager from ide to the newer (incomplete) ata version and rebuild, or requires disabling a troublesome driver (often times it’s a freebsd-ported network driver causing interrupt sharing issues).
I always feel bad for those people, but some of them stick around, rise to the challenge, and figure out the right combination of tricks to get Haiku running on their hardware – the elation they exhibit in #haiku makes it pretty obvious when this occurs
I really hope installable Haiku CD with partitioning support appears on the scene soon – and following that, I hope the ata bus_manager is completed, and some of the more troublesome driver issues are resolved – then they can concentrate on apps and other luxuries (such as wifi!)
Fun times ahead!
I think the best option for projects like this is to focus on virtualbox which now even has 3D.
In the long run just on on top of Linux+X+OpenGL+Virtualbox.
That way you just have all the advancements in linux drivers for free and only have to care for the virtual hardware.
A native install could provide the same setup.
It would be very similar to Amithlon in a way.
I don’t understand these types of suggestions – someone recently suggested the same in the SkyOS forums.
IMO, that’s a sure way to demote an up-and-coming OS to permanent-hobby status. It will never make it out of the virtual machine.
Virtual machines are great for people who want to run other operating systems for compatibility reasons, but I don’t see the advantage of using them for OSes designed for daily desktop usage like Haiku. Citing that it “solves” the hardware support problem is just ignoring the actual problem.
Well, I wish all open source OS would support every piece of hardware there is.
But at the moment nearly all of them don’t have the manpower to do that. And for some of them that day will probably never come.
You have to face reality. If the performance hit from building Haiku ontop of a very minimal Linux system (like Amithlon) is not noticeable and nearly everything “just works”(tm) then it might be worth looking at.
Or you can remain in pre-alpha (for real hardware) for a another 10 years.
False dichotomy, especially as Haiku already works fine in virtual machines.
Or you can create a driver SDK common to various OSes, like BSD (I believe it is used by Haiku) to combine manpower. It works for CUPS. It works for Xorg.
It kinda used to work for X.org .. nowadays new X.org releases are Linux only and have to be ported to the BSDs, because that whole x-platform thing held Linux advancements back quite a bit.
And Cups is way better on OSX than on Linux.
I think it should absolutely be focused on virtual machines…. VMWare player, Xen, etc. Using a VM makes it EASY for people to try out without being bogged in hardware hell. I know I’m tired of fidgeting with hardware settings just to try something out. VMs are to the point they’re going mainstream. If you make your OS something that runs well in VM then you will get lots of users to start making apps and USING it. Hardware support will come when enough people are available to use it.
This somehow defies the purpose of an operating system, moreso of one which is known to be extraordinarily slim and fast. If you have to first boot Linux to boot Haiku, the whole exercise becomes somewhat pointless, especially if you consider that the host operation offers a much broader choice of applications than the guest. All that Haiku in VirtualBox would offer you this “it looks like BeOS” experience. But as soon as your curiosity has been satisfied, you will most likely return to your GNOME/KDE/Windows/XFCE/[Other desktop of choice] and continue with your serious work there.
Haiku keeps getting better and better, God willing it’ll be running full time on my Aspire One once it hits Alpha.
I might even have it single booting on my MBP one day, completely replacing Fedora. We can all dream eh?
The Be API is quite possibly the most beautiful API I’ve ever had the pleasure of coding for.
I can’t wait for this moment to come! I used to run BeOS R5 before they went out of business. I decided to get a Mac because there was no way I was going to run Windows.
I am a happy Fedora/Zenwalk/Slackware Linux user but Haiku/BeOS is the only OS that would make Linux dissapear from my laptop.
-2501
No need to wait; in spite of its pre-alpha status, Haiku runs quite well on my Aspire One, with native support for video, audio and ethernet.
You can try it booting from a USB to see if your specific model will boot/work.
From linux, you can use:
dd bs=1024 if=haiku.image of=/dev/sda
In windows, you can follow…
http://www.haiku-os.org/community/forum/how_to_install_haiku_to_usb…
…or simply use:
http://modeen.se/HaikuOnAstick.htm
Pre-alpha images with some apps (Firefox, Vision irc client, Wonderbrush image editor, Pe programming editor, VLC, etc.) are available from:
http://haiku-files.org/raw/
It boots but no wifi support and no sound either.
It flies though, even running off the usb flash drive.
Wifi is a ways off, but either the HDA driver or the OpenSound media node should support your audio chip
Try the HDA driver first – you must enable it manually for build in your haiku images:
http://dev.haiku-os.org/browser/haiku/trunk/build/jam/HaikuImage#L1…
OpenSound is an optional package that will conflict with any native drivers that are already installed, so if HDA doesn’t work, that would be the last resort.
I know sound works in Haiku on the AA1 – as I’ve seen it running on koki’s machine
Wow! I’m excited about all the positive comments in this treat. Can’t wait to give it a spin…
Thank you very much for the useful links Koki!
Cheers
I have been diving into Haiku big time for the last two weeks. The only major problem I presently have is that Haiku will not support my USB speakers.
I am looking at the code right now to see if I can fix the problem myself.
The other problem with handling the filesystem seems to go away if I breakup the key files into diffirent folders. If I can get a handle on why I will file a bug report.
However, while BeOS works fine on my sub-1GHz machines, Haiku really seems to what the faster to machines to fly. Not that it does not work on my 550 P-III machine, but you will notice the diffirence compared to running BeOS on the same hardware (Bootman is great).
I have had no problems installing the downloaded Haiku images using BeOS, but I get shivers down my spine trying to imagine what a Windows/Mac/Linux user goes thru moving said image to a disk partition. And VMware slows it down too much on the hardware I have to show the OS’s true power.
As soon as Haiku goes Alpha with a boot CDROM image I expect a lot of people to be impressed.
Anyway, I suggest that anyone who wants to try it out go to HaikuWare.com so they can get an image with lots of sample apps.
FWIW, I have run Haiku on many different PII/PIII boxes, include a couple of Pentium MMX machines. Most recently, I booted it up on an old P75 laptop with 40mb RAM – and it only failed because there’s no driver for cirrus logic chips, and the VESA compliance on that old machine is less than what the Haiku VESA driver likes to use
If you end up with a supported graphics chip on an old machine, it’s definitely quite usable. I will likely throw it on my trusty old P200 MMX fulltime soon for some benchmarking/timing tests I want to run.
It doesn’t run as fast as BeOS on older CPUs, but that’s partially because Haiku hasn’t really been optimized yet, there is a lot of debug chatter and other stuff like paranoid memory allocation checking in the code. It has also been built to support newer hardware/more RAM, etc. which makes it a little heavier overall than BeOS was. It’s still very very lightweight however compared to the mainstream OSes.
http://www.flickr.com/photos/umccullough/3270279452/sizes/o/
Note that this was actually using it day to day for browsing and such, as well as doing actual Haiku development/debugging, not just idling around.
Edited 2009-02-10 20:19 UTC
Build http://haiku-files.org/raw/haiku-pre-alpha-r29155-raw.zip reboots immediately after the post-screen using qemu-kvm version 0.9.1 (kvm-74) on F10 64bit kernel 2.6.27.12-170.2.5.fc10.x86_64 #1 SMP.
Works fine without KVM.
Haiku is the perfect lightweight OS for netbooks, I really hope they don’t miss the boat on that one.
that would be a good target. 80% of the hardware on all the various OEM versions is the same and almost all have linux drivers already available to “borrow” from. They have dual-core processors which BeOS can benefit from, and they have more limited resources to show what BeOS can do when other OSes slow the machine down.
Unfortunately, while the netbooks share “80%” of the same hardware, they also are all “20%” different. And there is a *lot* of models, with new models every month. In particular you want things like ACPI to work flawlessly, which is the kind of part to be typically in these 20% (depends on the motherboard and on the device buttons, and so on, I guess).
The other thing to change from one netbook to the other is the wireless card, which is the key to the “net”-book concept… And there is a lot of work to be done here in Haiku, no wireless stack at all for now. And I understood that the wireless stack wasn’t a priority for the Alpha release, so…
To sum up it would be quite difficult to support all of them, especially the newest ones that no Haiku developper already own.
Besides, linux drivers cannot be “borrowed” to be used in Haiku. mmu_man explains in details (among other things) why they cannot (easily) port Linux drivers to Haiku in this very interesting post about the need for “open hardware”:
http://www.haiku-os.org/blog/mmu_man/2008-11-03/say_what_you_want_f…
Basically, Haiku is not Linux so you cannot simply “reuse” the drivers; additionnaly, Linux drivers tends to be a big mess for several reasons, meaning that it’s hard to port them successfully to another OS.
So what’s sure is that hardware is still a big issue, as for any “alternative” OS. Which is sad because yes netbooks would be an interesting target for Haiku because of their limited ressources and because of their eagerness for quick boot and responsiveness.
But who knows… ? For now, my EeePC 701 boots quickly Haiku and has sound and Ethernet. Basically it’s wireless and ACPI that lacks. Which are both a big deal of work to implement, but let’s be patient.
Basic (read Intel CA implementation) ACPI is there but isn’t enabled by default yet.
Wireless support will comes, and the very few (wmb2200, iw2100, rt25xx, hfa384x aka Prism 2 & 3 …) that are/were available for BeOS R5 should works fine under Haiku, at the cost of having to edit by hand a config file, which is not very beos/haiku/user friendl, we all confess that.
Today there is far less windows-only hardware than 10 years ago, thanks to Linux and BSD communities big efforts to bring hardware makers to more openess. Atheros opened its HAL. Intel opened pretty much all its hardware devices, even their GMA. AMD/ATI do it now too, Realtek were since long already, and so on. Except for very niche hardware markets and, what a surprise, nVidia, the list of device supported by alternative OSes is increasing. And more and more auxiliary chips are made by asian makers, whose business model is less based on closed documentation, most propably because they’ve grown by… looking to what and how others makers were doing stuffs.
The end result is it’s less difficult to have access to some hardware documentation that before. And Haiku already benefits from it, which make its hardware support future less darker than it was in BeOS times…
Still, using an alternative operating system will never give you the everything is supported of the #1 OS.
That’s why they call choice freedom. But no choice comes without alternatives solutions…
Its always nice to see Haiku making progress. I wonder when they’ll hit alpha?
Nice to see that Stephan integrated the GUI patch, but where are the rest of the controls?
Is there a live CD from which Haiku can be installed? (BeOS is having problem booting on my large disk)
Someone said it, a minipc with Haiku and a good media center, fast browser would be really nice.
Edited 2009-02-10 21:27 UTC
Still being worked on.
I made a video a few months ago showing all the various features of Haiku. I am still amazed at how well it is coming together, it definitely beats the early days of 2002 when everything on the teams page was pre-alpha with the exception of OpenBFS, which was beta.
http://www.youtube.com/watch?v=jSMT8cM20m0
Very nice demo! – almost the same hardware I demo’d Haiku on at LinuxWorld in August – except you have a better video card (and it shows). I used an Optiplex GX270 also, but it was the SFF version and had integrated intel video.
Thanks, I planned to use an old TNT2 card when I built it out, but the Dell nVidia card worked out of the box, even with DVI and my Apple 20″ Cinema display.
From an OSNews article on the 3rd birthday of Haiku (http://www.osnews.com/story/8114/Three_Years_of_Haiku_The_Long_Road…):’And if Haiku is “ready” by 2010, no one will care.’
Well, it was at least a good prediction on the year of release
I’m well aware that all operating systems have their advantages and disadvantages, but when Be arrived in 1996 (PowerPC) it could hardly be compared with NeXT which had been available since the late 80’s (and which did multi user).
BeOS _was_ a good but not the best.
Dom.
It depends how one defines “the best” I suppose.
hey, there is only room for one obsessive NeXT user on OSN, and I’m not leaving! ;P
As operating systems go, they were quite different. BeOS was smaller and lighter/faster, while NeXT was more robust (especially in terms of RAD tools); BeOS stuck with the document-centric philosophies of the original MacOS, while NeXT focused more on application-centrism – etc, etc, etc.
There’s a big difference between not having a release and having one even if it’s an Alpha because to have a release, you need to define your goals and then classify the bugs accordingly.
It’s a big work but IMHO it’s necessary as it creates a kind of ‘common grounds’.
But one think that is necessary is to use a naming/numbering/description/release note of the release which is as descriptive as possible and *honest* for the user: a release name such as ‘Haiku Alpha 0.x Developer preview’ (adjust the number as needed but 1.0 means *stable*) would be fine..
The article talks about KDE, the reason why even now KDE4.0 is flamewar material is that the naming and the release note was dishonest, suggesting that it was ready for endusers when it wasn’t which is something that we don’t expect from Free Software developers.. (well that and that several distributions were assh… to select KDE4 as a default KDE desktop when it wasn’t ready)
Come on, Thom, if you’re going to make sweeping statements about no other OSes having multiuser and list some, you need to be more careful and accurate about your history!
“Windows” is more than just Win9x mutations: it also includes Windows NT and its descendants as well. Well, Windows NT has had multiuser (not merely silly passwords and user folders everyone else could access, but actual multiuser) since it was first released in 1993. And yes, it was released as a general purpose OS also meant for desktop use, and was in active mass production: even though it really didn’t gain much ground until around 2000-2001 numbers-wise in comparison to Win9x mutants, the truth is it still had more users than BeOS probably ever has had
I’m sort of surprised you didn’t mention that Haiku has now been booted natively on Macs (at least in the last 24 hours, someone going by the nick of bebop-haiku managed to boot it on his MacBook) and a few other things. Sure, as of the last information I had, bebop-haiku hadn’t managed to get the sound fully working, and Bluetooth… well, I don’t think he had anything to test that with, but he was booted and running Vision and FireFox (the biggest problems with FireFox are the common BeOS FireFox issues that may be had on BeOS as well). Who knows? Perhaps Apple will inadvertently find themselves once again being the maker of BeOS/Haiku-compatible machines
NT? Multiuser? How? Two users could *not* use the box concurrently.
Operating systems like Linux allow two people to sign into either a CLI or GUI (or both) at the same time. That’s multiuser.
Definition of multi-user can be discussed forever… but either way you’re wrong.
NT3.51 was already multi-user capable – enter a product called NTrigue (technology now commonly known as Terminal Server).
Once an operating system supports the notion of different owners per process with proper separation and permissions, it is pretty much multi-user.
In my opinion, BeOS is (still!) one of the most amazing operating systems ever written. I used it for a while. There wasn’t any desktop effects or technologies with bloomy names, but it was lightning fast. In 90% of all applications you hadn’t to wait a second to open it.
Back in 2000 it even looked better than KDE 2.0 (or 3.0) and Windows Me in my opinion.
I am now at MacOS 10.5 – which is great – but I never experienced such a amazing system like BeOS again.
Long live Haiku! =D
1. Live CD
2. Partitioner/Installer
3. Basic Wireless Networking Support
It’s kind of restrictive having to run the OS in a virtual machine, and I’ve got an old machine I can’t wait to install Haiku onto. The reason I’d like to see wireless networking working correctly is because I’d like to be able to connect the old machine to my WPA-encrypted network instead of using an ethernet cable (much cleaner).
However, given how much of a royal PITA it is in Linux and how little wireless manufacturers help open source software developers with specs and other information necessary to get their devices running, I don’t have very high hopes for that one. My experience with their crap devices has been almost 100% bad (even on Windows).
I occasionally run Haiku in VMWare player on a fairly fast dual core laptop, and it runs fine.
But a few months ago I managed to get Haiku running natively on my home machine (C2D E8400) and I was amazed by the difference! It ran fast and felt very much like R5. This isn’t a great surprise, since it’s quite fast hardware, but after some optimizations it could be this fast on older hardware too. At least I hope 🙂 Sure It has dropped me in KDL sometimes, but you can expect this of a pre-alpha OS.
The only problem I ran into was the lack of support for SATA; I had to put my BIOS IDE settings in legacy mode for a successful bootup.
So if you have the opportunity to test on real HW instead of inside a VM, do it. You won’t be disappointed.
For info about installing natively just search the haiku forums.
I did the installation from linux where I used dd to copy a RAW Haiku image to a new partition. After this I used a linux version of makebootable (http://stefanschramm.net/dev/makebootabletiny/makebootabletiny.c) to make the partition bootable and I made the partition selectable in grub. That’s all there is to it.
It should support most SATA. If there is a problem Haiku would like to know about it so we can get it fixed and you can run in SATA mode.
I used BeOS from R4 until about a couple of years ago.
Recently I toyed with the idea of buying a netbook but having a tiny 10″ Compaq m300 kicking around I gave Haiku a try. The only problem was the lack of wifi (so its back on FreeBSD for now…) but on the whole it was much better behaved than when I ran R5 + bone on it.
Nice work guys – keep it up!
Love it.
Been following it for a long time now
Looking at BeOS today, is almost like looking at a commodore 64, with regards to the development of other systems. I don’t think it will succeed much, and it’s flock of people with Munchausen-tendencies, isn’t for mainstream embrace.
BeOS R5 had antialiased fonts too, not subpixel AA but the standard shades of grey.
In today’s world, besides nostalgia, why would I want to use Haiku? I’m sure BeOS was a great OS, unfortunately I hand’t the opportunity to test it back then, but today we have MacOS X, Vista, Win7 and a plethora of (good) Linux distribution, plus various BSD and OpenSolaris operating systems.
So, again, besides nostalgia, what would haiku bring new in this world already full of choices? Will the Haiku developers be enough to sustain this project and make Haiku an OS for the masses?
Speed. I don’t understand why I still have to wait 5 seconds for an app to open.
Edited 2009-02-11 13:06 UTC
As very well highlighted by the article, BeOS was well ahead of its time back in its days.
And today… Well, there is still a lot of splendid ideas and design choices in BeOS/now Haiku that makes Haiku competitive towards other OSes.
For instance, give me an OS that combines at least these features:
* less than 10-sec boot time
* no lagging, no latency in user interactions
* small hardware requirements, small footprint on CPU and RAM usage, very light HDD usage (both OS and other software are very small)
* easy to use and efficient GUI yet still appealing
* never crashes, eventually part of the OS crashes but you just need to restart this part, not the whole OS — same for device drivers installation, just restart some part of the OS, do not reboot the whole computer
* easiest software distribution model (in my opinion): no enforced package manager, no unmaintanable registry base, just unzip where you want and it works
* virtual desktops with independent resolutions
* real shell (bash) with all the power of UNIX (including easy scripting)
* powerful and efficient file-system
* well documented and well defined stable API
Plus some other features which are quite unique to BeOS/Haiku, as far as I know, for instance the “translators”. This is a bit like codecs but for any type of files and which works with any application. Just install the “MP3” translator and voilà every application just knows how to play and record MP3 files. Now just install the “PNG” translator and you can display and record any graphics in PNG format with any graphic application. And so on with any data format. Do you see the beauty of this?
But I agree that it’s somewhat difficult to imagine what is so extraordinary here if you never tried BeOS/Haiku before. 🙂
However like you and others, I’m not sure that Haiku will succeed because of the smaller software base. But this may not be an issue… People just “need” to port all of the free open source software that made Linux famous and the software base should be there. So why not? 😛 It’s a bit like an ARM port of Ubuntu. If the software base for ARM Ubuntu is fine, then it should be for any OS thanks to the open source model. 🙂
* less than 10-sec boot time
No one really cares about this since no one really resets his computer 10-20 times a day to see the difference.
* no lagging, no latency in user interactions
Linux/Vista feel fast on Core2 Duo with 4GB of RAM… MacOS X is fast too.
* small hardware requirements, small footprint on CPU and RAM usage, very light HDD usage (both OS and other software are very small)
(Some) Linux distroes? (Slackware, Zenwalk, but there are more).
* easy to use and efficient GUI yet still appealing
Any OS with KDE?
* never crashes, eventually part of the OS crashes but you just need to restart this part, not the whole OS — same for device drivers installation, just restart some part of the OS, do not reboot the whole computer
Again, Linux?
* easiest software distribution model (in my opinion): no enforced package manager, no unmaintanable registry base, just unzip where you want and it works
MacOS, maybe Linux?
* virtual desktops with independent resolutions
Besides multiple monitors, who need independent resolutions for different virtual desktops?
Again, any KDE/GNOME based Linux has irtual dekstops. Even MacOS, nowadays.
* real shell (bash) with all the power of UNIX (including easy scripting)
Uhm, Linux?
* powerful and efficient file-system
Linux? OpenSolaris?
* well documented and well defined stable API
I’m not a programmer, so I won’t digg into this.
That translator stuff are really nice, I haven’t heard about them until now.
No one really cares about this since no one really resets his computer 10-20 times a day to see the difference.
Notebooks? Also, that 10-second boot time is great for netbooks.
Linux/Vista feel fast on Core2 Duo with 4GB of RAM… MacOS X is fast too.
With *4GB* of RAM, and they only *feel* fast? Try Haiku on a 7-year old machine with 256MB of RAM.
(Some) Linux distroes? (Slackware, Zenwalk, but there are more).
Until you run some heavy apps. All BeOS apps open quickly and run fast.
Any OS with KDE?
This is subjective. IMHO, KDE is not easy to use.
Again, Linux?
Wake me up when I can update or replace the whole networking system, media kit, desktop environment and still not logout/reboot.
MacOS, maybe Linux?
Linux? Package managers.
OS X does it exactly like BeOS, except using .zip files.
Besides multiple monitors, who need independent resolutions for different virtual desktops?
Web developers? I want to see how my site looks on different resolutions and color depths without going all the time into the Display preference pane.
Filesystem: I agree here, ZFS is great.
API: Cocoa & BeAPI are probabily the best APIs out there.
Translators: I can’t understand why noone has copied this already. We’re in 2009, and we’re still focusing on *programs that open formats*, instead of just file types.
No one really cares about this since no one really resets his computer 10-20 times a day to see the difference.
Sure it matters, I’d be pretty pissed if my TV took 2 minutes to start up, and I rarely use it. Different expectations and all. OTOH, people that reboot 10 times a day tend to be dual booting… which I think is the target audience for now (that and VMs).
Linux/Vista feel fast on Core2 Duo with 4GB of RAM… MacOS X is fast too.
I’m using the Windows 7 beta, and while nicer than Vista, it’s still an order of magnitude slower than the BeOS or Haiku. If you’re just comparing Linux, MacOS, and Windows then you lack a good point of comparison.
* easiest software distribution model…
MacOS, maybe Linux?
Meh, different strokes for different folks. Personally, I hate the way Linux does that… “ipkg install package-3.2452.123.1″… Ok, looks like it worked… now what just happened, and where did the executable go?
Besides multiple monitors, who need independent resolutions for different virtual desktops?
Web developers, people that prefer working at a lower than maximum resolution for most day-to-day stuff…
Uhm, Linux?
Ya know, just because Haiku exists doesn’t mean you have to stop using Linux. Also, OSes borrow features from each other all the time. And competition is a good thing.
* powerful and efficient file-system
Linux? OpenSolaris?
Having is not the same as using. Tracker queries have been a major feature of the BeOS for a long time.
Curious: whenever there’s a discussion about whatever improvement in boot time, usually the following discussion contains *a lot* of messages: people do care about fast boot time.
1) So this means that to have a fast desktop on a laptop you have a battery life measured in seconds?
2) No everybody is rich: think about the OLPC..
As for the rest, I agree with you, BeOS is no more superior to the competition on these points.
Linux and Solaris have some great filesystems (XFS, ZFS) but until user-space catches up with them and actually starts to make use of the advanced features they offer, you may as well be running ext3 for all the difference it makes to the user.
The interesting thing about BeFS (& AFS, which I have to mention of course) isn’t just that they support neat features such as arbitrary meta-data streams, it’s that the user space and applications have been designed to take advantage of that support.
By the way, there were some wonderful demos of BeOS that should be still amazing to watch today. It was something like that :
(1) Open a video in a media player. (opens quickly and plays flawlessly, of course)
(2) Drag the image of the video on the desktop: a shot of the video at that moment has been taken and recorded in a file stored in the desktop.
(3) Now select some portion of the video and drag it: this time you got some audio of the video, just the part you selected.
(4) Instead of dropping on the desktop, you can drop into another application and the other application directly works on the tidbit you selected.
(5) Of course all these applications runs together with no lag at all. Especially none in user interactions.
I remember playing simultaneously ten or twenty MP3 files at the same time on some Pentium 120MHz with 24Mo. No problem at all, while it was quite impossible with any other OS. (not to mention that at the time these other OS couldn’t even play simultaneously several wave sounds but hey…)
It’s just… So easy, so natural, so efficient. It’s not about something that other OSes cannot do, it’s about how things should be done.
Well, whatever. I’ll stop bragging for now and we’ll see in one year or two where Haiku will be then. 🙂
i agree with you on this, but…(cont.)
… i’m afraid you’re overlooking the fact that existing userland software written to run on linux or to be multiplatform, is implemented in a way that results from the relative scarcity of assured high level facilities in the main platform – resulting in the rise and use of complex, large, all- encompassing toolkit libraries, that in turn need further code layers (themes, wrappers or “engines”) to adapt to the specific platform they are deployed to, or appear adapted while not really being so
porting existing applications like KDE ones or openoffice would require porting their base framework for the most part
the problem is, merely doing so would result in applications that would exploit little to no native feature (distinguishing ones like the translators you mentioned, in particular), would feel evidently “alien”, and worse, would defy the goal (minimalism and efficiency) of the system they’re ported to (due to the unnecessary layers – own framework plus “native look” skin engine – they’d retain between the app’s main code and haikuOS native gui library)
in order to avoid the bloat resulting from this, differentiating (branching) the application main code, instantiating native classes instead of, say, OOo ones in platform specific code paths, would be a sensible approach from a SW design point of view
but it would effectively be the same as redesigning the application from the inside, i suspect it’s unlikely it will ever be done for large applications (paradoxically often the most useful ones)
in Ubuntu’s case is fine because you already have an all – linux software base, and problems would only come from architecture dependent software … with porting (better, converting) linux SW to other OS platforms (even on the same architecture, say X86) that may have their own special features or idiosincrasies, it’s the other way around
the open source model concerns distributed development and code sharing
it doesnt say anything about the elegance and cohesiveness that goes in the *design* of code and in the management of a project, or the skills of people designing code, or that people tasked with porting a package will choose the best overall approach *for the target platform* (and not just for the application, or for themselves)
Right, right, right, you’re completely right from the first to the last sentence.
However what’s sure is that only time will tell what kind of software support will have Haiku in some years. So I have the right to dream and drool a little… 🙂
By the way, I’m more concerned about hardware support than about software support. Drivers have always been an issue for any “alternative” OS and it’s definitely the case for Haiku, as it has been for BeOS. On the other side, there is a lot of wonderful software that does an amazing job already available for BeOS/Haiku and more software is to come everyday. You can still switch to another software and change your habits (and it happens that it’s for better not worse). While chances are high that any random computer does not have the proper drivers and it’s not so easy to switch the hardware parts… 🙁
BeOS had some very nice things going for it (and I still boot into an R5 desktop from time to time), but most of the things BeOS did were also being done by OS/2 at roughly the same time. And OS/2 did many things BeOS could not.
I think people are overstating BeOS’s capabilities and level of innovation here, but OSNews has always been very much biased towards that OS for reasons I don’t completely understand…
Why? Simple, the developers have found that none of the existing operating systems fulfill their vision of the os they want to use. It doesn’t have to be an OS for the masses. If it is then great, but for all intents and purposes it only has to satisfy it’s creators.
There are no shareholders lurking in the shadows waiting for a return of their investment. The investment here is the spare time (as in time with which you do what you please) of the developers and the return is the OS itself, hence it cannot fail unless the developers call it a day and abandon the project.
Now as for Haiku’s future, there’s a huge difference between what one would wish for and what one expects. What I wish is that Haiku does really well and claims the throne as the open source desktop OS. What I expect is that Haiku will gain enough of a following to ensure a healthy community and also attract enough developers for vital ports and native applications. I also believe there will be some commercial possibilities for Haiku that hopefully will involve opportunities for atleast some developers to make money while working on Haiku.
People really need to stop being so anal about success and failure, particularly when we are talking about spare time projects. Like other projects of it’s kind, Haiku doesn’t need to make excuses for it’s existance.
It’s free to use, so take it for a spin and make up your own mind.
Simply put: an OS experience that none of the other OSes you mention has.
Unfortunately, no one can be told what the BeOS/Haiku is… you have to see it for yourself.
And regarding the question
…
– Why the hell no?
– Because they can.
So true, so true… just like LOST: if you tell someone about an invisible island in Pacific on which a plane crashed and has polar bears, a smoke monster, a group of hippie scientists that did time travel experiments called DHARMA and a group called Others that kill everyone who wants to harm the island, they’ll think you’re crazy for liking the show. Only when they see it, they’re … uhm, converted.
For one thing, there is no wait cursor!
Beos/Haiku believes that the user should control the OS, not the other way around.
The only one who could really answer that question is you. Hypothetically, it would probably be because aspects of the OS appeal to you. If that doesn’t apply, then it’s probably a case of “move along, nothing to see here.”
As you’ve pointed out in other posts on this topic, there is a lot of overlap between Haiku and Linux – and in terms of raw functionality, I can’t think of anything that can be done in Haiku/BeOS that cannot be done in Linux or other modern OSes. But, speaking for myself at least, the appeal isn’t the raw functionality – it’s the particulars of how that functionality works.
Personally, I see the main appeal being that Haiku is the only (relatively) modern OS I’ve used that extends the whole “UNIX philosophy” to GUI. By that I mean the concept of “do one thing, but do it well” – small, apps that rely on standardized means of communication so that they are separate, but can work together. And that’s in contrast to the more application-centric model, where most applications are essentially walled gardens, so they need to be complete end-to-end solutions.
They have done a fairly good job of sustaining the project up until now – and if anything, I would say that the development has been steadily picking up speed (and new developers), especially in the past year or so.
“Personally, I see the main appeal being that Haiku is the only (relatively) modern OS I’ve used that extends the whole “UNIX philosophy” to GUI.”
actually I think the appeal is the opposite: that Haiku is one unified system and not a collection of different parts cobbled together from different places with different agendas (kernel from here, X from there, Gnome from somewhere else, a zillion distros…) to make up a system. There is one bug tree, one vision, one point of contact for developers and users.
But it’s open source so you can still do anything you want with it. Haiku is special because it is an OS that is open-source AND unified.
I think we’re talking about slightly different things. I’m primarily talking about the way that Haiku / BeOS lets individual apps work in concert, as if they were a single unified piece of software.
While it seems that you’re referring more to the higher-level system architecture.
While on the surface both Windows and Linux look quite nice – under the hood we have one big mess. Layers upon layers of APIs, libraries, services, frameworks and subsystems.
On Haiku – you have a beautiful and consistent API, well defined kits (app/interface, filesystem, input, kernel, media, network, translation, etc). Everything is modular, integrated and consistent – it’s so much easier to understand what’s going under the hood. Even browsing the source code – it’s sooo much less intimidating than looking at the Linux Kernel or X.Org sources.
This results in greater stability, responsiveness and speed. Extending Haiku and adding new features will be much easier and faster, even with less developers.
Does a version compiled against GCC4 still break binary compatibility with existing BeOS apps?
Yes, unless you create a hybrid version which also contains gcc2 libraries. If you do this, the runtime_loader analyzes the elf header at startup and figures out what version of the libraries to load.
I remember purchasing a copy of BeOS R4 when it was releases, received 4.5 free of charge (free upgrade) and purchased the 5.0 pack which included the Gobe office suite.
The big buzz will not only be after the R1 release of Haiku but how it develops and moves forward. Integrating webkit into the operating system would be a great first step, multi-user capabilities.
With that being said, hardware support will also be crucial to developing mindshare and addressing short comings in the existing driver API given the changes in what is provided in the hardware world.
The great thing I do like about BeOS is that the GUI is already organised and well thought out – so whilst the ‘others’ are still bickering, Haiku developers can focus on expanding hardware support etc.
All it needs to do is boot in ten seconds, support a ton of emulators, and run wahcade/mamewah.
How about mmap? Are we gonna be able to get a WINE port someday?
http://dev.haiku-os.org/browser/haiku/trunk/headers/posix/sys/mman….
This was one of the major improvements over BeOS – along with many other POSIX compliance fixes
looks like your article made it to slashdot
I am mostly interested in cross platform development between BeOS-Haiku, Windows, Linux, OSX so the issue for me is cross platform IDEs and which APIs to base apps and tools on. The rest of the features that make BeOS so elegant are great bonus points but a low end IDE is a big drag on development.
My IDE vendor of choice was and still is Metrowerks which I fell in love with from Sys8 MacOS days. At one time Metrowerks gave Codewarriors umpteen platforms to choose from including MacOS, BeOS, Windows, and a preview for RH Linux. Although the company is largely gone (Nokia bleh), I still use CW 8 on Windows and I now hate to use the BeIDE (Be’s Metrowerks version) because of the lame debugger. If Windows didn’t have this Metrowerks IDE, I’d be forced to go back to VS6 which is okay but hardly cross platform. I have not yet found which IDE on Linux I might yet like.
If Haiku is to get a fresh set of modern apps, it is going to need something much better than BeIDE and since Metrowerks is gone, then really all those IDE tools are fossils with no sources.
Now on Ubuntu I see some interesting cross platform tools that are Linux and Windows friendly and open source like CodeBlocks which I am still trying out on both OSes. I can’t help but wonder if that might not be a great candidate to retarget to Haiku as it is natively supporting several widget toolkits already. But of course it wouldn’t be a Be app, but really neither was BeIDE either.
Today cross platform usually means choosing between GTK, Qt, Mono, WX, Java and plenty more besides.
With Haiku API sources now available I wonder if a relatively small portable library for Windows, Linux and OSx could be released that would allow cross platform apps with BeOS APIs instead of those above?
That would also mean having another platform to build Be apps from with much better development tools and bring them back to BeIDE for final tweaks.
That would also allow many Be apps to be recompiled for other platforms and eventually persuade some of those users to tryout the Haiku platform.
I am still gobsmacked that Process Controller got ported to Windows and it works great. I missed that story till the other day. I would love to see many more Be apps come over too but having the Be APIs on many platforms would make that a piece of cake.
Now several comments have mentioned the problem of bringing Linux apps to Haiku esp the big ones, that is they bring their own basic GUI plumbing with them every time and are over sized as a result. If the Be API kit was out there free of Haiku, then you get a marvelous way of creating apps for many platforms but with Haiku as the home target.
I would appreciate any of the developers here to comment on cross OS supporting Be APIs, and what could replace BeIDE and what IDEs on Linux are useful to the Haiku devs.
As a casual BeOS programmer I actually find BeIDE’s simplicity somehow refreshing in a “back to the roots kind of a way”. It is an effective and flexible tool, but it needs a successor. Right no there is some community effort in replicating the BeIDE functionality with the project Paladin (listed on BeBits), but IDEs being such a personal preference there is *always* room for one more, especially if it can handle make/jam a bit (which BeIDE doesn’t). I think CodeBlocks using the native Be API would be awesome. It seems just about the rigt mix of functionality.
As for Portable BeAPI .. It is a big project, but I as well think this may be Haiku’s secret weapon! I am not one of those claiming BeAPI is the best, it definitely is not. It is however a very balanced, ingenious framework. Very basic overall so it can be used by a beginner (it is in fact one of the simplest C++ frameworks there is) but at the same time quite powerful.
This is looking awesome!!! Is BeOS about to make a comeback? I sure hope so – it was and still is the best desktop OS I have ever used.
Maybe even a live CD – boy that would ROCK!
You might want to have a look at this:
http://www.haikuware.com/view-details/development/app-installation/…
That’s nice, but I don’t have a computer that can boot from a USB device.
Heh, Amiga’s Workbench and Amiga OS later on sure had personality…