I was listening to a podcast where the speaker was opining on the state of Haiku, and he said something that resonated to me. If Haiku did one important thing really well, it would gather a lot more attention and more support.
So, I was thinking to myself, what would that one thing be? An open-source software that people use to make a living? How about Blender?
If Haiku was the best setup to run Blender, would more money and attention be brought to Haiku?
Thoughts?
To be honest, I disagree with this podcaster. Hereâs why:
Haiku is an operating system. And an operating system is meant to do MANY things REALLY well. Operating systems were meant to do various tasks, not one. I think Haiku has a chance because itâs a fresh (maybe new) way of doing computation. Sure, thereâs Mac, Windows, Linux, but nothing else. Linux distroâs get boring over time. Windows isâŚWindows, and Mac doesnât have enough compatibility with Windows. Haiku will never be bigger than MacOS because, like MacOS, itâs limited to what the developers and hobbyists create. Thereâs just not enough developers to keep pace with major OSes.
When it comes down to the wire, Haikuâs ability to flourish will come down to peripheral support and application compatibility/ease-of-porting. Haiku already runs on the vast majority of CPUs and is VERY compatible with tons of hardware. However, itâs the small things like keyboards, mice, webcameras, network cards, USB-C devices, and of course, the elephant in the room, GPUs and 3D acceleration - which will come with time as long as there is a steady supply of donations to the project. Donât forget, weâre still in Beta 5, this isnât even a 1.0 kind of situation. Weâll get there.
I think Haiku has a bright bright future! Itâs clean. Simple. Built on some really ground-breaking and solid foundations - computationally and programmatically speaking. It blows me away how efficient Haiku is.
TLDR: I think Haiku is well on its way. It doesnât need to do any ONE thing well to succeed, it just needs to keep moving forward. App compatibility will be a key to its success - even if the developers donât see it that way.
And of course, more companies need to get onboard porting their software for Haiku.
Haiku is indeed a general purpose OS, so limiting it to one app would not make sense. Besides, Blender would benefit from 3D acceleration and thus would require drivers to do that (currently lacking for the most part). Hardware vendors arenât interested in open sourcing all their drivers and specs to Haiku.
Personally as an Haiku developer, I listen first to my own needs, and then to that of other people who already use Haiku. In my case, it is my main operating system on my laptop, and I wouldnât trade it for anything else. I use it to browse the web, write code for various projects, and so on. In this usage I donât have any need for 3D acceleration, but I would love to have a working soundcard, for example.
We already have severall things we do really well. Responsiveness in the UI, application startup times. A truely seamless update experience (If it involves the haiku package you update, otherwise you can just relaunch the application for the new version. No long waiting for logos on reboot needed.) Easy downgrading. You can just start a lower version. etc.
I highly doubt Haiku would ever be âthe best setup to run Blenderâ Single window applications that donât integrate with the rest of the OS donât fit that well with Haiku. Haiku has itâs strengths elsewhere, say in replicants, translators, working seamless drag and drop (even of more complex stuff, like color drops). etc.
Blender would likely be better suited on a linux OS that has good 3d rendering support. We do not. And even if we did on some computers, in order for it to be the best setup we would need to have it on most computers. And there is little chance of us getting a good performing 3d driver for nvidia cards in the future, even if we manage ones for amd and intel cards.
Like Pulkomandy I also like to check first what I, or my family, needs from Haiku.
Right now that is more âbasicâ stuff. Like discovering printers in the network without having to set them up. No way my family will know the correct IPP adress for that printer.
Yes and No. Just like a Porsche (i.e. put any car here), yatch, and/or wearing a Rolex watch does not fit everyoneâs daily driver desire. These products are perfect within their realm of purpose and for a certain clientele.
Another OS perspective, the real focus of most users is in discussing the available software for the OS - not the OS. The OS is mainly the canvas between the applications and the hardware. Kinda like understanding that people donât look at your carâs engine compartment moreso than the carâs exterior and seating area. But, it is important to have a good canvas - just like a engine compartment and engine (i.e. for those people interested in engine components and engine compartment designs - or just wanting a good canvas).
Personally, I donât know what to think. Without more interest, there wonât be more money, and without more money, Iâm thinking that weâll always be behind the curve. Youâre right, the apps are where itâs at, and we donât have anything that makes Haiku really attractive. What is or are the apps that will give Haiku market share? Windows has games, Linux/BSD as server apps, Mac has (or had) creative content apps. BeOS if I recall correctly was supposed to be Media creator focused, before they ran out of money. I was thinking if Haiku had a Haiku-specialized app that had a fan base already, that would be something.
Hi Haiku Friends! Hereâs my ten centsâ worth, from someone whoâs been trying Haiku out the last 4 days first on a âlive USB installationâ and then on a proper SSD installation. Still in my âtest drivingâ phase, trying to see what Haiku can and canât do for me!
For me, I think that the Unique Selling Point of Haiku is that itâs free and open source, but with less of a learning curve compared to Linux or FreeBSD (even desktop oriented varieties like FerenOS or GhostBSD!). On the downside, the fact that itâs got no sleep mode or hibernation mode to conserve battery use while having the ability to continue working where you left off, and the lack of webcam support are two big downsides for users like me who come from other OSâs who kind of expect those to be just there.
I think that for a lot of Windows 10 users who are using older hardware which are not compatible with Windows 11, an easy to use alternative OS like Haiku which is free and non-threatening could be a godsend when Microsoft stops providing security updates for Windows 10 in October 2025.
So if Haiku users and developers think about where there might be potential for an influx of new adopters for the OS, I think that âlight usersâ whoâd rather save money on buying new Windows 11 machines might be a good target user base to aim for. For that, getting a good web experience for Haiku with modern browsers that people recognise (like Firefox, which can sync with the browsers on their phones), and which can recognise the webcams on their Windows 10 PCs, and sleep when they close the lids on their laptops might be a good focus for the project.
Oh, nearly forgot. For it to be meaningful for people to switch from their E-O-L Windows 10 installations, Haiku needs to be demonstrably secure from malware attacks and phishing attacks that will no doubt increase even for minor OSâs now that the bad guys will be coming at us with A.I. enabled knowhow. Safety from obscurity might become something of the pastâŚ
Thatâs my ten centsâ worth anyway. From a newbie Haiku fan!
I think your information is out of date. My proverbial grandmother could install Linux Mint or Ubuntu. Click on âTake over whole diskâ and your UEFI installation is automated. Compare that to the hoops we have to jump through: create a Fat32 partition, make a directory, move a file, rename it (âwhy is it not named that in the first place?â). We require the new user to know that there are different disk partitioning formats, and which ones to use for what purpose. That is a steep learning curve.
They will update their hardware. They always do.
But even for those few who donât, Windows 10 isnât mysteriously going to stop working next October. There are still diehards running Windows 7 and Windows XP. And there are a thousand voices on YouTube screaming âChange to Linux!â
So we have a year to change Haiku from a developerâs playground to an OS for the masses? Nice to see such youthful optimism, but this is actually a very conservative community where even a proposal to move the leaf logo up a few pixels will trigger months of debate.
And Haiku was promised several times that users would not switch to the ânewâ Windows (Vista, 7, 8, âŚ,) that was so resource hungry and complicated. I donât think we got that many new users.
We did get a few ones coming from Mac OS at each Haiku release, and asking we make Haiku a bit rore like BeOS.
Anyways, the gerenal idea is still right: Haiku is focusing on being a good operating system for personal computers. This may sound boring at first glance, but there are other os that are not really doing it anymore
This discussion has always been the same since back in the BeOS days though, what is the main thing that will draw users, and the answers are always as different as the ways you can use a computer. 
I think Haiku has to do itâs own thing, but yet be competitive with the others, which will mean a thousand things it has to do at least as good as the others, while having something extra the others donât.
Unfortunately, the state of PC sleep and hibernation is tricky to say the least. Ever since S3 sleep was largely superseded by S0ix/S2Idle sleep, sleeping has been mostly getting worse for non-Windows OSes in general. Some laptops wonât actually go to sleep, others may wake up on their own after some time, etc. Hibernation has also been a long-running problem, even on Linux. On Windows and macOS, S0ix/S2Idle has also been causing power drain problems too.
At the moment, Haiku is at the interface level mostly single-user. However, the underlying system is multi-user (i.e. ssh runs as its own user). There have been efforts to have FDE (full disk encryption) from @axeld, the operating systemâs immutabilty preventing it from being maliciously broken, and password lock for the screensaver (which is prolly more secure than most X11 screen lockers already).
With all this being said, security is not a top priority for Haiku right now. This seems to be in part due to the overall goal of R1 sticking closely to what BeOS R5 was. As in, do not expect the security model and related changes (like multiuser in the UI) until post-R1. At least, unless the developers change their minds between now and later.
Also that bit about attackers targeting Haiku using âAI-enabled knowhowâ is a bit funny, since there have been attempts to generate Haiku-related images and theyâve all turned out horrendously due to models not knowing what Haiku as an OS even is. Doubt that the project needs to worry about that in the foreseeable future.
Thanks for the explanation about the S0ix/S2Idle sleep and the implication for non-Windows OSes. Sleep and hibernation had become so ânormalâ to me that I hadnât realised how it makes it difficult for non-Windows platforms. Having said that, when Iâve used Linux and FreeBSD on my 2012 vintage Thinkpad X230, they seem to go to sleep or hibernate OK but perhaps that is because the X230 was still able to go to S3 sleep, being a somewhat older PCâŚ
Looking on the bright side of life, I guess that the ability of Haiku to shut down and boot up within seconds does make it less onerous to shutdown in between sessions.
I wonder whether thereâs an easy way to get back to where I was before shutting down though? Perhaps a way to get particular apps to start up in a particular workspace so that I could perhaps do an âopen recentâŚâ document to get back to where I was. Do any of you Haiku gurus on laptops have a nice workflow for doing something like this? Thanks.
Haiku apps generally open where they were at last shutdown already, if they were not closed beforehand.
I would be interested what you base this assumtion on.
X11 screenlockers have a fatal design flaw, if the app crashes your sesion is visible again.
I donât think our screensaver app does anything diffferently.
Also, packagefs mounts beeing RO does not prevent malicious edits to the OS, you can just edit the packages on the disk, create new ones to mount over. etc. Hell you can even drop a driver somewhere in non-packaged and the OS will happily load it into kernel memory.
I agree that, for now, it mainly protects users to accidentally delete a system file. But, it will probably be an advantage when security will be implemented. For example, an integrity check will certainly be easier and faster to do on packages than on thousands of files. Sure, it will still have to check memory and files in other places but itâs a reasonable guess that you will gain in speed and so reactivity compared to what are doing some antivirus.
There are a couple reasons, some of which come from the many (many) blog posts of jwz (XScreenSaver author) about screen locking. But the most user-facing advantage is that Haiku has no TTY consoles by default, rendering attacks involving switching to one of them like this impossible:
All this being said, after skimming through some of their blog posts again someone should consider getting a cat to walk over their keyboard while Haiku is locked. Or kids, as theyâll work too if an issue with cinnamon-screensaver is anything to go by. 
Huh? we have the kernel debugger, that is basicslly the same thing. Just that there is no way to lock it, it is always available etc
We should instead simply tell people that this provides basically no security, and will continue to do so untill we have re-archtiected this in a way that the logon happens in a seperatr app-server, or app-server session/context. And the only way back to the logged in session is by the login window returning a sucess. (i.e deliberately switching, not just crashing)
KDL isnât quite the same thing, since it only allocates a rather small amount of memory and will take down the entire system when it does run out. This is in contrast with a virtual TTY console on Linux, which has access to the entire amount of RAM and therefore enables a lot more potential for malicious actions.
The kernel debugger can be disabled in the kernel settings file
Once you are in the kernel debugger, you donât need to perform an attack and exploit a bug to get privileges, since you have entire control of the machine already