I have nothing against BeOS or Haiku, but I think their time has come and gone. I wouldn't mind seeing their ideas applied to a Linux distro.

1 thing that bugged me was this: "It stands apart from other projects like AtheOS, Syllable, QNX RTOS or MenuetOS because of it’s long history..."

QNX pre dates BeOS, by over 9 years...

I wouldn't mind seeing their ideas applied to a Linux distro.

Take a look at the composited&accelerated&animated graphic systems that you can find in every modern OS these days (hint: Haiku isn't included in the list), take a look at the declarative UI APIs that are emerging these days. I'll take Wayland + QML over Haiku any day. Same for Tracker vs Dock. Touch based phones/tablets? Forget about them. Pervasive multithreading? Apple's Grand Central Dispatch beats it. BFS? ZFS/Btrfs are much better. BFS attributes? Nepomuk can do what the BFS atrributes did and more (it may even possible to implement BFS attributes in Btrfs)

1. The benefits of a database-like file system are not obvious to end users unless the operating system leverages it extensively. BeOS (and Haiku) were built to be run on such a file system, [...]

2. Object-Oriented API. On Be/Haiku, C++ is the primary low level interface with the OS. This is simply not applicable to Linux and never will be for a multitude of technical and political reasons.

3. Focus on multi-threading/responsiveness. [...] Linux concerning the balance between performance vs responsiveness makes it obvious to me that it is difficult to do both well. [...]

They seem too intent on bringing back Be rather than build an OS with a certain set of capabilities.

I am suspicious of hobby OSes claiming that they can change the World.

Sorry, but after the (negative) SkyOS experience (they were going to overtake Windows and conquer the World. Look where they are now) I am suspicious of hobby OSes claiming that they can change the World.
Even if they have great, innovative ideas, it takes tons (say million dollars) of money and a large team of devs in order to get anywhere.
If they tell me that they can create a nice hobby OS, I can believe it.

Please ignore all of the negative comments posted to sites such as this because these users are not your audience.

I am suspicious of hobby OSes claiming that they can change the World.

If they tell me that they can create a nice hobby OS, I can believe it.

"Linux on the Desktop" seems to be struggling to hold even a single digit in user share:

I search for the biggest files to delete them and not for the smallest one.

If Haiku don't close in. then it can not be faster then Linux.

Touch based phones/tablets? Forget about them.

Pervasive multithreading? Apple's Grand Central Dispatch beats it.

BFS? ZFS/Btrfs are much better.

BFS attributes? Nepomuk can do what the BFS atrributes did and more (it may even possible to implement BFS attributes in Btrfs)

IMHO, adding every feature imaginable isn't the way to succeed in today's technological world. Computers have long been able to do much more than what 95% of people need them to do. A ridiculous number of people still use XP, which was first released back in 2001. It wouldn't surprise me if more than half of people could suffice with nothing but a webbrowser.

I think the way to succeed is to do the vital stuff well. Eye candy might be nice, but I prefer instant actions (e.g. opening a folder on the desktop, Haiku takes under a tenth of a second while other OSes average ten times that). With Haiku you can boot, do your work, and shutdown in the least amount of time. The GUIs are multi-threaded, rarely modal, and have a speed that make using any other OS downright frustrating.

I can display and edit my scripts which use 'find','grep',etc does-it make my filesystem a database??

Actually yes. A filesystem is a database.

In regard to ext2, ext3, ext4 and NTFS the functionality have largely been left unused for reasons I don't understand.

"GCD is "just" an abstraction layer to pervasive multi threading.

Programming multithreading in BeOS is hard

"ZFS is not a lightweight desktop filesystem, and Btrfs has a long way to go before it is stable.

"You've confused an API with an application. The equivalent of Haiku (BFS) attributes are POSIX Extended Attributed (xattrs)

1. Somebody on the team gets a clue about marketing.

Besides privilege elevation, daemons which run with other UIDs, a desktop envi does not need multiuser mode.

(complaints about Linux distros)

the R1 survey they recently did looked like a nice marketing exercise; but strangely, they ended up putting funding in a place from different from where the survey results were telling them).

[And while there may be no official statement of how Haiku is being positioned, the dynamics of the project do seem to point in the direction of a hobby OS.

So you are fine with desktops always booting to root?

They should have forked FreeBSD and added a Haiku like interface on top. I understand coding for fun but....a single user, 32 bit system? Why not build upon all the work that has gone into improving *Nix security models?

No, they're not. GCD still uses threads. Code written with GCD is still multi threaded. GCD just abstracts the details away. It's nice, but it's not like thread abstraction models haven't ever been done before.

X) No multi-user

X) No real existing application base

X) The OS is barely used on the desktop

let alone the fact that there this thing is not running on any mobile platform in any capacity

X) It has no server functionality

X) Its limited to the Desktop only currently

Agree totally with that. The desktop is now a mature market with relatively well-defined user expectations, what modern desktop OS developers should focus on IMO is the quality of the implementation. Making things bug-proof, responsive, reliable, lightweight, etc...

When was the last time you actually USED this on your desktop box?

Chicken/egg. Not Haiku's failing.

Nor is Windows NT.

X) It has no server functionality

And?

E-mail clients are a thing of the user-space, not a core part of an operating system.

About semantic filesystems, well... I still wait for them to become actually easier and more fun to use than a good file hierarchy, but that could indeed be a small revolution in a few dozens of years, once it has matured a bit.

"Asynchronous message passing and high level APIs make multi threading very very easy.

"Do you disagree that ZFS is not a lightweight OS designed for desktop users? I'd love to see your arguments to support that particular claim.

It can be used in desktop and in fact many people is using it daily. It would be funny if you were trying to tell me that there is some kind of resource usage impediment to use ZFS on desktops.

[q]No, they're the same concept. Both allow you to store arbitary key:value pairs against an inode.

Everything is part of user space these days, but point taken

I was thinking "more userspace" again, these semantic systems store their data in normal databases (sqlite, virtuoso ...). Perhaps it could benefit from some fs support, but it doesn't appear to be needed yet.

There are other security models than multiple users. In my opinion, an capability-based and app-centric security model would be a better fit for desktop use.

They wanted BeOS compatibility, and it'd have been hard to provide that starting from a FreeBSD fork. Besides

current Unices are not necessarily perfect for desktop use (monolithic kernels, scheduling not targeting responsiveness, user/admin security model...),

If you have multiple users accessing the same computer then you need file separation.

I've been able to buy servers for work for years from various major OEMs with no OS shipped, so why not desktops (and laptops)?

"X) No multi-user

"X) No real existing application base

"X) The OS is barely used on the desktop

"let alone the fact that there this thing is not running on any mobile platform in any capacity

"X) It has no server functionality

"X) Its limited to the Desktop only currently

But wouldn't that mean that applications see the hard drive in a different way than users do ?

I use that _all_ the time.

The argument "I don't use it, so it isn't useful to anyone." was NEVER a valid argument!

operating system which doesn't support multi-user in 2010 is useless. Period.

I wouldn't call the enterprise a fringe use case

I can't find the damn link, but I believe the results from Microsoft's Customer Experience Improvement Program showed like 90% of machines having only one user (the owner).

Does it? GCD is "just" an abstraction layer to pervasive multi threading. If anything it appears it's everyone else who is finally waking up to the idea of pervasive multi threading.

Sigh. Then why don't you read the APIs before talking about what you don't know? In BeOS there are APIs similar to xattrs and then there are more things. BeOS APIs like open_indexdir or open_query are not realistically implementable in a POSIX xattrs filesystem.

Nepomuk can do what the BFS atrributes did and more (it may even possible to implement BFS attributes in Btrfs)

GCD makes it easy for mortals to do mutli-threading, while simultaneously making it easier for the Operating system to provision threads more efficiently.

I also vehemently disagree that there is anything particularly difficult about writing threaded code. I am certainly not a superhuman and find it easy enough.

Valhalla,

I understand what you say and even agree with most of it. The point that I was trying to make is simply that Haiku is a developer centric project, not a marketing driven one, and that this is not likely to change (see the statement that I quoted).

1. The benefits of a database-like file system
Also imagine computer-viruses and so on, which extends every file with lots of big nonsense-attributes. Or viruses, which writes itself as attribute to files, etc.

"Valhalla,

I understand what you say and even agree with most of it. The point that I was trying to make is simply that Haiku is a developer centric project, not a marketing driven one, and that this is not likely to change (see the statement that I quoted).

We have the year of Linux on the desktop for that, there's no need to bring more guests to that party.

"We have the year of Linux on the desktop for that, there's no need to bring more guests to that party.

>i think that linux has more than "a percent" of market.

You may think that Linux is take 90%, but in real world it less than 1%! Linux complete failed!

>third the fact that you dont see any perspective in linux, its not a linux problem, its your problem.
Really? It's now my problem, I stayed Linux few years ago and pretty happy that I do not spend my time to it more.

>and finally, i read a few topics of your linux-faq.

It's bad that you do not speak russian language, some sense was lost during translation.

"X) No multi-user

Linux 1, Beos 0, I rest my case.

I think you completely missed the point of my post. Let me make it more clear: unrealistic hype based on wishful thinking is bad. We've had enough of that with Linux.

"BFS? ZFS/Btrfs are much better.

People who use ZFS on their desktops have traded speed for a bunch of features which they will very likely never use. Why bother?

OK. I suggest you go try and use whatever desktop OS takes your fancy, on top of the ZFS file-system with a 1Ghz CPU and 512mb of ram. Then while your waiting for it to actually "DO" something, dig out another equivalent system, install BeOs/Haiku on it and set it to your liking. You'll have plenty of time before the first 1 finishes, honest!