Hi,
I’m new here, encouraged to install Haiku in a virtual machine by some interesting videos on youtube.
I know that one of Haiku goals is to be fully compatible with BeOS.
My question is: are there any further long-term plans? After this “compatibility thing” will this system evolve more or it just be a replacement for old-closed system?
My question is if Haiku is somehow like FreeDos for MS-Dos (or ReactOS for Windows Server 2003) that nothing more (probably) will happen with it after compatibility is reached?
That is an interesting thought, but I think there are plans afoot to keep it up to date, but with only a small team… It’s a nice desktop system now, if it works with your hardware, but still a long way to go also.
I have R1B4 installed on a spare thin client, but use Linux, or BSD, as my main system, & I can’t see that changing any time soon…
Haiku is so much more than only a BeOS clone already.
The BeAPI as a way to write native applications for Haiku is here to stay,but additionally you can use countless Qt and GTK applications that you couldn’t use on BeOS.
Driver support is already a lot better than on BeOS,which didn’t even have Wifi as far as I know.
The long term goal of Haiku is to make a open-source operating system targeted especially on desktop computers (not servers,like the Unixes…).
It won’t stop when full BeOS app compatibility is done (I think that one is done already?) and it also won’t stop when it’s so good that you never look back at Windows.
There’ll always be some innovation and Haiku will continue to get improvements.
I Take haiku as a Portable operating system.
haiku is beyond the compatible things now.
libreoffice, blender …
these app are working which is not appear in beOS time.
Exactly, Haiku was created to preserve the BeOS system, so care was taken to ensure that old programs can continue to be used. Over the years, various changes have been made that differ from the original BeOS. In addition, as already mentioned, languages and environments have already been ported to haiku, which enable the use of QT, KDE and java programs.
Over the years, due to the numerous changes, some compatibilities of the old BeOS have been lost, because at some point you have to look ahead.
Because Haiku was built from the ground up, all parts of the system were reprogrammed. This is also for reasons of avoiding rights relating to the owners of the beos sources. Thus, haiku is free and can be further developed in various directions with the contribution of people (the doors are open).
I think this is a bit of a misconception. BeOS-compatibility was the major goal at the beginning of the (then OpenBeOS) project. Simply to have a fixed goal. As it turns out, creating an OS from scratch took quite a bit longer than first naively anticipated…
The longer it took, the blurrier got the ErsatzBeOS goal.
At least since the first beta, I suspect in most people’s view, Haiku became its own thing. Over the years (decades really) Haiku gained many features that go well beyond BeOS (e.g. localization, layout management, package management, modern hardware support, modern compiler support,etc.).
I don’t think there are official concrete plans, but rest assured, Haiku will continue to grow and evolve beyond compatibility with BeOS.
It should probably be noted that the 32-bit version of Haiku is BeOS compatible (mostly, I guess). The 64-bit version is not and is not under any constraints to be, as far as I know (which is fairly little).
The 64bit version will be mostly similar to the 32bit version eventually once it’s 32bit loader and such works… at which point most people would run 64bit as default and install the 32bit binaries in the 64 bit version only if they needed compatability with old binaries.
I’d be interested to know which 32 bit applications are still considered desirable to have in 2024.
I do happen to own a couple of very old 32 bit laptops, but I don’t think Haiku should make it a goal to keep these old beasts running. They will die a physical death before long anyway.
I’m still regularly using the 32-bit version on my 32 bit laptop. It works great, and I don’t see any good reason to drop support. I’m also using it on a 64-bit desktop that is limited to 4GB of memory. Won’t bother with 64 bit as I can’t use more than 4GB anyway.
Why do you think Haiku shouldn’t support older machines? It’s already supported, you just have to not break it.
Edit to add:
I think hardware support should be improved before any crazy feature-adding compatibility-breaking updates. At the very minimum there should be a Rosetta-like compatibility layer before any major jumps, but again, I see no useful advances, especially none that should drop support.
Haiku 32 bits uses PAE, so it can access more than 4 GB of RAM (with the same limitations on “per process RAM”, IIRC, but you DO make use of the RAM above 4 GB).
I run both 32 and 64 bits versions of beta4, and at least on my Phenom II… the 64 bits version is notably faster (I’m talking about 20-30% faster when doing builds for Python packages, for example).
Just wanted to mention those things, as on 64 bits capable hardware, unless you want to run the 32 bits version for compatibility reasons… running the 64 bits version seems like an obvious choice, unless you are REALLY constrained in memory, I guess. For reference… my Atom N450, 2 GB DDR2, netbook is faster on 64 bits than on 32.
Currently, you can’t run BeOS software (32 bits, compiled with an old compiler GCC 2.9x) on Haiku 64 bits. It theoretically could, and I think that there are some unfinished patches on that regard, but doesn’t seems we’re getting that anytime soon.
Thus… only the 32 bits version can run old BeOS software.
Regarding the boot mode… I mostly use only MBR / “Legacy mode”. No big issues there regarding 32 vs 64 bits , as long as the CPU supports 64 bits.
Besides knowing that some (earlier) UEFI machines might be a bit picky regarding the bootloader code (32 vs 64 bits, even if the final installation of your OS is full 64 bits)… I have little experience with Haiku and UEFI (other than a pretty smooth experience on ONE machine… I just followed the instrructions and all went fine! :-D)
I do happen to own a couple of very old 32 bit laptops, but I don’t think Haiku should make it a goal to keep these old beasts running. They will die a physical death before long anyway.
From what I understand maintaining 32-bit support cost very little to the Haiku developers since pretty much all the work has already been done. Aslo machines of the past were built different and are generally much easier to repair that computers of today.
On x86-64 there are 16 general-purpose registers, on Pentium 2 there are 4 general-purpose registers. Pentiums are register-starved relics. By comparison, ARM 7 have 14 general-purpose registers and ARM 8 has 30 GPRs in its 64-bit instruction set. PowerPC has 30 GPRs in its G2 instruction set which gave it the illusion of being faster than a Pentium 2. It just didn’t load and store as often.
When making observations about architecture differences, keep in mind that compatibility isn’t the only reason for performance differences.
Not sure if that was mean for me (or a “speaking to the general public” thing). In any case… yes, we are in agreement. Maybe my poor English, or my desire to keep thing simple enough, got in the way of making clear that I was talking about compatibility with BeOS binaries, and CPU arch performance differences, as orthogonal entities/characteristics.
@BiPolar
I was speaking in general to the people in this thread. Nothing personal was intended. I’m glad we are in agreement.
@thread
At one point, Debian Linux attempted to make a version of 64 bit Linux that used only 32-bit pointers for a fast way to use 64 bit CPUs’ registers but still not waste storage on bigger pointers than were needed on early systems with 4 GiB addressing or less. They weren’t successful but the point is that there are many variables involved with architecture changes.
I didn’t even think of Haiku as a BeOS preservation system. It certainly can be used as such but, as others pointed out already, it’s far from being just that. Just have a look at the Depot/BeSly or the projects announced here, and tell me how much of BeOS you see.
From a software developer’s point of view, I see little (if any) differences in my daily work, other than Haiku being surprisingly snappy even on very old machines. I mainly write Numerical Analysis and Multimedia libraries and honestly, if the hardware is supported, it doesn’t really matter if I work using FreeBSD, GNU/Linux, or Haiku. Pretty much everything I need is already available and works just fine, especially since R1B4 and the nightly images after that. Implementing part of a project on BSD or GNU/Linux, then on Haiku, then back to BSD (depending on what machine I have with me when not at home) is actually a common thing for me.
Of course it’s not just roses, more hardware support is needed. But Haiku is already way more than just a BeOS modern replacement. I’m pretty sure there are many Haiku users that never used BeOS itself and came here for an alternative desktop operating system.
As for old 32-bit machines, well, they are not dead yet. In fact, one of my 64-bit laptops died while the two old 32-bit machines I own are still alive and kicking. Granted, many GNU/Linux distributions dropped 32-bit support already, and FreeBSD is at the last major release supporting 32-bit - next one won’t. Additionally, several old machines won’t run GNU/Linux anymore, even if you use a distro that supports 32-bit, because of glibc incompatibility (happens since glibc 2.36, does not affect all old laptops, but does affect many).
I see an opportunity there. The general trend to drop 32-bit might bring some people to Haiku, if it keeps supporting 32-bit. Their other option would be to let the old machine collecting dust in the attic, while it still can run Haiku pretty efficiently.
