What is the point of BNX?
As stated on OSNews, Sven Jorg is slapping all of Haiku’s hard work in the way of Kits onto the QNX kernel and XFS.
I’m wondering why we need a 5th version of BeOS (R5, Bone, Zeta/Dano, Haiku, and now BNX).
But besides the eventual difficulty developing for all of them, and the user confusion that will result, what does the QNX kernel do better than the NewOS kernel that Haiku uses? I thought that the kernel was the bit that loaded all of the drivers, and that replacing the kernel would necessitate using a completely different set of drivers. I also heard that QNX had even less desktop acceptance than BeOS, and thus less drivers, so BNX would have less drivers than all of the BeOS variants.
Also, doesn’t using a completely different kernel make the idea of binary compatibility, and thus using the library of already-existing BeOS applications, impossible?
I also heard that the smooth multithreading roots of everything BeOS related would conflict with the realtime aspect of QNX.
Wikipedia mentions nothing in the way of XFS having attributes, so a huge major aspect of BeOS, which no operating system has yet to fully emulate, would be lacking.
But we have a free, opensource, fully operation kernel that goes by the name of NewOS, and this group is using a licensed, proprietary kernel that goes by the name of QNX. Why?


  1. BNX would use a much smaller driver base than Haiku, because of using a different kernel
  2. BNX would not be usable with the entire library of BeOS applications
  3. The QNX kernel would conflict with the multithreading roots of everything BeOS related
  4. XFS removes any possibility of filesystem attributes, which removes one of the biggest features of BeOS
  5. Haiku already has a fully operational kernel, which does not suffer from many of the deficits that Sven cites
So what's the point of BNX?

More quirks about Sven Jorg’s interview:
He states that BeOS can not take advantage of multiple cpus, specifically by saying that Zeta must work to support “multiple cores”. If I know my history correctly, BeOS has supported “multiple cores” since 1995, according to the Be Timeline. Cores means CPUs, right?
He also states that Haiku will be exactly like BeOS, including its age, but several times, such as in app_server and the be filesystem, Haiku has fixed bugs in the original Be Inc. implementation. Also, our beloved Axel has added 64-bit cpu support, SMP, MTRR, write-combining, and >1g mem support, if I remember correctly.

So what advantages does QNX provide over what Haiku already has?

–Walter Huf–

snes_rocks wrote:
What is the point of BNX?

Probably no reason to come here and start an inflammatory discussion about it… but I tend to agree…

“Haiku” was mentioned 6 times by Sven during the interview, but I didn’t see any mention that they intended to further the Haiku development at all… only to use the code to accomplish their “1 year goal” of running BeOS apps (and there was no mention of whether this would be binary-compatible support - but I highly doubt it based on their approach).

I understand it’s sort of a hobby project - and therefore I could care less what they do.

I didn’t see mention of using XFS rather than BFS - where was that?

They mention XFS under part 9, where they say “We’re also looking at taking some parts from elsewhere, like the TCP/IP stack from FreeBSD and and the XFS file system from IRIX”.
I highly doubt BNX will amount to anything. I was just thinking that we could localize the discussion about it to one place, because I thought there would be a lot of outrage and controversy and such around it.
But oh well. Maybe nobody’s heard about it yet.
–Walter Huf–

I think a few of us decided to politely ignore it :twisted:

[Beta wrote:
"]I think a few of us decided to politely ignore it :twisted:

yup… although I still have this burning urge to know if they are simply going to branch Haiku code and not contribute anything back to it, or if they will “accidentally” become active Haiku contributors… there’s a moral issue I’m having there.

The progress of every linux-based BeOS projects was outstanding; wasnt it? There will always be only two words people say regarding future of BeOS: haiku and zeta. No other.

I wish more people instead participate in development of haiku and give greater momentum to the project. I cant code… :cry:

I completely agree with all of the posters who have been critical of BNX.

I thought it was very bad form of Sven Jorg to make critical comments about Haiku, and then (almost in the next breath!) he basically said that “oh, by the way, we’re using some Haiku code” !

As far as I’m concerned, it seems to be common-sense that if someone (external to a project) makes critical comments about "Project X"s code, they should not then be using "Project X’s code !
Especially when they are not contributing to it.

( Constructive criticism of code by Haiku devs themselves is ok - that tends to be “part and parcel” of the general dev process. As long as it’s constructive … :slight_smile: )

Even if this Sven guy did try to contribute some code, I would refuse it …

Haiku is making great progress, and I completely support the Haiku devs! You’re doing great work guys - just ignore this Sven guy … And most importantly - please - don’t be distracted by him! He just isn’t worth the energy… Just treat his comments as “water off a duck’s back” , as it were …

Oh, and finally - poster “umccullough” makes a very good point, mentioning a possible (even if unlikely) scenario of BNX “accidentally” (or otherwise) contributing code. Just a plea to the devs - be very, very careful there! The phrase “a wolf in sheep’s clothing” comes to mind … If you refuse anything from BNX, I’ll be very happy. :slight_smile:

One only needs to look at the Reactos project to see what effect some possibly dubious code can do (and that wasn’t even from outside the project). For any Reactos people passing by here - please, I’m not meaning to “slag off” Reactos - not at all! The devs there have done some amazing work, and I really hope it keeps going and does well. I’m just mentioning it as a caution…

Code from “outside” (as it were) needn’t even be legally dubious. It could simply be a “ruse” to steer things in a particular direction… Again, Haiku devs, please be cautious …

This subject was ignored for a reason — Haiku is Haiku. “Be yourself, don’t worry about others”. Well, Haiku needs to Be itself, and BNX is essentially irrelevant.

latte wrote:
Oh, and finally - poster "umccullough" makes a very good point, mentioning a possible (even if unlikely) scenario of BNX "accidentally" (or otherwise) contributing code. Just a plea to the devs - be **very, very careful** there! The phrase "a wolf in sheep's clothing" comes to mind .... If you refuse anything from BNX, I'll be very happy. :-)

I meant it slightly differently than you interpreted it. I was sort of predicting that anyone who starts using the Haiku code may actually end up developing it rather than branching it. This isn’t necessarily a bad thing - but may not be anticipated by the BNX devs, thus the “accidental” in quotes.

I’m not really sure why you have so much anger towards the BNX guys - I really didn’t get the impression that they criticized Haiku as severely as you make it sound. In fact, in the OSNews interview, Sven did actually say “Thanks to Haiku”.

I just wanted to know what their plans were with the code.

I’m not angry at em; but i wish they helped development of haiku instead which has already picked up so much of momentum…

Leaflord wrote:
I'm not angry at em; but i wish they helped development of haiku instead which has already picked up so much of momentum...

I doubt they will hurt Haiku, and beyond this it’s irrelevant. I personally don’t believe they’ll get the work done in just a year - but even if they do it’s still another non-linux, non-windows, non-mac OS which is always a bonus.

I’m sure that at least one patch will float backwards to the Haiku resp. Even if it doesn’t (which i doubt), Haiku can still share ideas and concepts - which will help pave the way for the true future OS - R2.

Although he might have worded his answers poorly, and may (or may not) have a somewhat eccentric view on Haiku, this is good.

The license for Haiku is made so that these kind of things should be possible. They have decided to use much of Haiku’s code even before it’s finished. Enjoy the fact that the Haiku is so good that others want to take advantage of it! Even now.

Who knows maybe BNX will rock or be a great companion to Haiku. Time will tell.

tqh wrote:
Who knows maybe BNX will rock or be a great companion to Haiku. Time will tell.

Speaking of which, can somebody put up a one year timer so I know how long to hold my breath waiting for it?? :shock:

My problem are these article comments-
“his team’s unique approach”
"mating of Haiku’s open source BeOS servers with a specially-licensed, real-time kernel from QNX Software Systems"
Source -

Taking other people’s work and merging them together doesn’t sound unique to me. It is most definitely not original.

I’m not necessarily complaining about the approach, I’d just like it to be presented properly:

He’s too impatient to help create a new original OS, so he’s going to hack together existing products to create BNX.

Now what really is his motivation? and will it keep the project moving?

On second thought, I’ve decided not to hold my breath.

Wouldn’t it take as much effort to make Haiku’s kits use the new kernel and filesystem as it would to fix whatever is left to fix in the current kernel?
And Haiku will be released sometime this year, judging by its current progress and rate of progress, so Haiku will be out before BNX, and then people will see Haiku in its full glory and will want to help and we won’t have enough things for the developers to do, and then BNX won’t have any advantage.
So he should’ve helped out with the Haiku project instead of starting another ill-fated BeOS-recreating project, and thus help bring that Day of Glorious Rejoicing closer.
But since he chose this way, we will have to wait more.
–Walter Huf–

Should be interesting to see how QNX’s real time scheduling interacts with media and UI responsiveness. Could have benefits, could have drawbacks, either case should be fun.

Additionally, couldn’t BNX be used as a stepping stone to embedded BeOS?

Just a few random thoughts from a bored uni student.

The point of BNX is to keep BeOS in plain eyesight of people, even as Haiku matures.
QNX kernel has more wide-spread use. BeOS ABI compatiblity is addressed by Sven Jorg as weaving Haiku onto QNX. The performance from such weaving will be questionable, but at least add to the visibility of Haiku to developer audience, so if Haiku keeps up great work then this should be long-term gain.
The advantage of BNX over Haiku is addressing short-term platform availability for devs to continue working while Haiku is still in alpha. As Haiku’s adaptation of NewOS will continue better responsiveness by multi-threading accordingly, than RTOS kernel targetting embedded devices resource consumption as QNX does, considering UX, BNX should never displace Haiku.

April fools! I hope so. This thread is really old. I think BNX is a cold project.

BNX project page (updated 2013-03-11) here:

BNX project page (updated 2013-03-11) here:[/quote]

Sourceforge are updating meta information about hosted projects as of late, and then are listing these projects as being then updated, even if the haven’t been updated since 2003 or whatever.

I think this may have been interesting, but considering QNX has always been closed-source, and is now owned by Blackberry, how was it supposed to ever happen?