Firstly, I’ll get back to finishing the test Haiku Wiki on my return to New York - I’ve made no progress because I’m back in my native UK at the moment.

Secondly, I was thinking - one day - there could be a market for a HaikuStick, in the same way you’ve got these ‘sticks’ emerging that run Android and Windows (i.e. https://www.amazon.co.uk/Intel-Compute-Computer-processor-BOXSTK2m3W64CC/dp/B01AZC4IKK) . Such a product could bring in some vital cash to fund improvements to the Haiku stack.

Thinking about it - all you need is someone to port Calligri to Haiku and you’ve pretty much got an OS for office work and Internet use, without the constant spying of Microsoft and Google or the lack of consistency of Linux on the desktop (this, I believe, is Haiku’s USP)

Any thoughts?

I’m not sure there’s a market for such quite expensive a device for Haiku in its current state. Probably not even for R1. I have no idea about “the markets” etc., but I imagine Haiku’s target audience currently in the computer savvy enthusiasts that would rather look for an inexpensive way to try Haiku without major investments.

Also, developing support for such an all-in-one computer would be quite expensive, I guess. Maybe a nice, fast ordinary USB3 stick to plug into an existing computer would be more interesting for the Beta phase and R1.

The concept of a “SmartStick” is simple - anywhere you have a HDMI-capable monitor or TV, Bluetooth keyboard, and Bluetooth mouse, you then have a highly portable personal and private computer by plugging in the stick. The concept has been implemented for Android, Chrome OS, and Windows. However, the underlying CPU/GPU/RAM/FlashHD is really at the bottom in terms of capability. This may change if Apple eventually comes up with an iStick or even a MacStick! :slight_smile:

USB3 capable systems are generally booting via UEFI. While I have heard of much progress in both areas, it is unclear if universal approaches have been implemented in Haiku. Also, hardening the boot sequence to allow running a wide range of hardware would divert resources from the more immediate goal of reaching R1.

One thought - I believe that one can run QUEMU and/or VBOX relatively easily as a portable application from a USB drive from a Windows based system. If so, it might be possible to consider a USB Flash Drive, bearing the Haiku logo, with a portable pre-configured virtual machine and market this (with a small price premium) as a way to spread the word and raise funds?


Find the idea with a USB stick also very interesting, unfortunately it will probably fail due to the lack of hardware support.

Could Haiku also well as a data recovery system imagine. Use USB to boot, retrieve and backup data, or restore it. Or if you could make an antivirus manufacturer to port its software for Haiku, then this would be a good application.

USB sticks are something we are likely to do, maybe for beta1, or maybe for R1. They are better than CDs because one can easily update the OS image on them.

We could sell them ready for UEFI, which works well, but is not fully integrated in our build system yet (but we can manually craft an UEFI compatible image).

As for hardware compatibility, it is our goal that Haiku should run, somewhat. Even if it is VESA and maybe no network and sound, at least you have something to try.


Like this? http://www.tunetrackersystems.com/discoverhaiku.html


DiscoverHaiku is nice but still basing on the alpha 4.1 or?

Also seen on http://www.discoverhaiku.com/ it has HaikuDepot, so it’s more recent than Alpha 4.1.

Thanks. I see it, great!


So, there is no need to re-invent the wheel. The price appears about right for a 8 GB flash drive with value-added components being a bootable/installable recent version of Haiku, introduction video, documentations, and an example of a major commercial non-trivial application (TuneTracker).

Maybe having a note and a link to this on the Haiku main page would be all that is needed for those afraid of the command line and/or lacking the confidence/experience in going beyond simply using their OS X or Windows as existing on their system.

It is important to separate “Haiku on USB Stick” and “Haiku on SmartStick”. The latter is basically a tiny computer, I personally have seen only ARM-based ones with various Android versions (maybe one or two with RemixOS). They’re mostly used as SmartTV appliances for “empty” HDTVs (the ones without an internal Smart/OS functionality).

If there is decent x86-compatible SmartStick device out there (like the one from Intel mentioned above), it’s worth trying to install Haiku there to see what works and what doesn’t. Since it connects to displays via HDMI we need support for modern Intel Graphics, at least setting resolution, display detection and preferably full 2D acceleration. To be able to use wireless keyboard/mouse and wireless network we need support for Bluetooth devices in our stack and a driver for built-in WiFi chip.

I like the idea of fully supported Haiku SmartStick (or nettop or any small desktop computer), but not in current state of Haiku developer community. Ideally, such project needs it’s own (highly enthusiastic) dedicated team improving hardware support, developing and porting drivers, fixing bugs, etc. Unfortunately, chances of that happening anytime soon are even slimmer than celebrating the release of Beta 1, let alone R1 this year.

To dream the impossible dream…

I agree with the suggestion that Haiku needs a device, an application, or a spin of sorts to compete with Chrome OS, etc. and be a light Web centric system. Maybe even a custom Web+ could fit this in the future; a user could start up Haiku and go right to a BeIA styled home screen.

That said, I do not believe that a Chrome competing device is a stick like the Intel Compute; they’re a fad and lack the power to do anything useful. I have a Chromecast, Roku, and used to have an Apple TV – imho, those beat the compute sticks in every way. As for a real computer, I would recommend to connect a tiny SFF or “mini” desktop to the TV with a wireless mouse (and maybe a keyboard) – and I might add, they already run Haiku.

If Haiku had to focus on this area, I really would say it would be the Raspberry Pi. Unlike the other arm boards, it is stable enough to target; unfortunately, I don’t know how to port Haiku to it or I would try, because I see the Pi as being an important area.

In any case, go Haiku!

2D acceleration? Even if some of our drivers do that (and we’re talking 1990 hardware there), the app_server does not make any use of it. Did you ever notice? Did the rendering seem slow? I think not.
Back then, the 2D acceleration had been benchmarked, and it turned out that with our design (double buffer, with app_server rendering most things in system RAM and not video RAM), 2D acceleration turned out to be slower because it resulted in:

  1. sending data to the video card (a full frame)
  2. sending commands to do the accelerated thing
  3. downloading data (resulting frame) from the video card back into app_server work buffer

In particular, AGP was not designed for fast reads, and step 3 was very slow. Hence the idea of 2D acceleration was completely dropped, except, I think, for using the hardware cursor. But even then, most cards support only a 2-color cursor, and ours needs more because it is antialiased.

So, no 2D acceleration is certainly not required.

As for “compute sticks” or whatever you want to call them, yes, this could be a 3rd party project, but I don’t think it is something the Haiku team is interested in. My workflow with Haiku involves a lot of compiling stuff, and to do this, I use a full-size, full-power PC. I also have a smaller and less powerful laptop for more mobile uses. I made them both work with Haiku by fixing the drivers I needed. That’s all you can expect from me as an Haiku developer.

I did briefly play with one of these ARM-based sticks, and I was not really amazed. It was running Linux, and it felt rather slow and not very good quality (but then, it was a rather cheap one). But, if I need to pay the same price as for a full blown computer, I would expect a reasonable set of IOs so I can actually make use of it. A stick with just one or two USB ports is not going to make it (plug keyboard and mouse and you don’t have any USB left!).

The Raspberry is just a board, not a full computer. You need to add a power supply, a case, etc. I would rather go for something like the Utilite (http://www.compulab.co.il/utilite-computer/web/home), mainly because I happen to have 3 of those around sitting unused. But, we need Haiku to run on ARM devices for this, first. And this is something I don’t really want to dig into now. Not until we get beta1 out.

Companies doing that using DosBox for example!
They use DosBox to distribute their old games not running on modern hardware anymore…
like Ultima 7 or 8 (sorry cannot remember the exact part).
If you start Ultima… DosBox will be run and directly starting the game. So for the user it looks like Ultima is just running on their system…

1 Like

Who is talking about games and emulators? I thought it was about Haiku on a stick and how to use Haiku to find a place in the market?

Today, you can run on any SmartPhone over emulators old DOS games and others (My loved ones are the AMIGA games).

I do not want you anything bad, but somehow the degree of the goal passes by.

1 Like

your answer is the question!

The pinebook could be a good laptop to target. It’s even cheaper that a USB stick. It has a good review. Maybe an intel developer board could fit in it.


But Hiaku is not Linux. Have you test the Pinebook with Haiku?

If you have test it, please give me informations about support of the hardware.

Greetings Lelldorin

hello friend thankyou

1 Like

The Pinebook is an ARM based notebook.

As far as I know, ARM support on Haiku is still work mostly in progress.

Also, all packages are currently for X86 and any old BeOS application would be either X86 and PPC.

Good idea for later down the road.