There’s a notion become meme/joke among a certain OS kernel community called the ‘year of the Linux desktop’. I think Linux certainly has a place in the computing world, but the aim of one OS to rule them all, even in the desktop space, may be foolish. Microsoft and Apple have done their best for their respective markets, but their OS programming isn’t exactly pretty, or lean, or open, and they have both been encumbered by DRM for decades so I won’t touch them. I think a sharper goal is ‘year of source-available operating systems’ - leaving open the idea that different OSes can be better suited to certain tasks and machines.
The default GUI of OpenBSD may not be the last word in user friendliness but I’ve made peace with it, partly for its reputation for security but also the simplicity and reliability of its TUIs. Its driver selection is wide enough for those in a position to select hardware to run it, stable releases are made every 6 months and the volume of GUI apps in packages is in the hundreds, so for me this makes a solid, general purpose, networked OS. But even I can’t live off of OpenBSD alone; Linux is still superior for music production as the kernel is easily configured for lower latency and more synths, DAWs, plugins etc. are targeted and better tested for it. I have a Slackware installation specifically for this purpose and I keep it air-gapped so I can setup the packages once and forget about it.
Engineering is always a compromise and with OpenBSD I forego some performance, partly due to security features but also to CPU features that the developers deliberately don’t use, like hyper-threading. Linux is faster but its developers still have to be concerned with security because of its prevalence in servers, not to mention the software that gets bundled with it to create bootable OSes. This is one area I think Haiku could excel, BeOS’s architecture and GUI were heavily optimised for SMP above all else, which Haiku has continued. Throughout its tool-chain it can afford to prioritise responsiveness and throughput at the expense of security because it’s not used anywhere important at the moment.
Perhaps Haiku’s closest functional equivalent would be Linux + Xorg + Plasma 5. Plasma 5 has fractional scaling and is much lighter on CPUs than GNOME, but Haiku is even lighter and doesn’t even need GPU hardware acceleration for a fluid experience. I imagine Haiku could be a great fit for the mid-2000s ThinkPads with their Flexview screens, once all the kinks in its DPI scaling are sorted, and I think these sorts of machines should be Haiku’s specialism.
I think Haiku is being aimed at the home user desktop in the main, whereas most other systems are trying to cover all bases, & therefore becoming extremely complex. Simplicity has its advantages…
True, but aiming for a general purpose desktop also means being usable online, and being usable online nowadays means, at least, a multi-user OS with a root account for administrative tasks, WebKit/Blink or Gecko-based browser, a mechanism to install patches to the base system and package management for everything the user installs on top.
Haiku needs a browser for help manuals and NetSurf would probably suffice, and I know from my Slackware installation that some offline programs depend on networking libraries. But as Haiku’s striving to remain single-user I think it would be wise to drop many other internet-related functions it does have in order to maximise benefits in areas where Haiku is competitive, and to make it easier to maintain.
“general purpose desktop”
“multimedia OS”
“one OS to rule them all”
Haiku is a resurrection of BeOS - which was part of the early exciting years of computing. And ppl on here are always reminiscing, and looking up old softwares. So why not scour the Internet archives for abandonware apps, and port still-useful ones to Haiku in an updated way? Haiku could then have its own niche of being a retro-mod OS.
Maybe it’s just me, but I’m still excited for computing today. I’m excited about RISC-V’s ascension. I’m excited that JavaScript features are being ported to HTML and CSS with the prospect of making web pages lighter. I’m excited about developments happening with the OSes and programs I currently use.
You can have your ‘retro OS’ if you want, but that alone isn’t what gets me up in the morning, unless it’s FreeDOS as that will run on just about any x86 machine. I think Haiku has a place in the world today, but trying to fulfil a vision from the 90s will prolong its development cycles longer than necessary.
Those are entirely sane minima. I agree, more or less.
Where the hell did you hear that? The technology simply isn’t 100% there, to have Haiku with multiple user accounts. But this is planned to be included, despite some people on here moaning about how “Haiku is meant for desktops”.
Every desktop OS, more or less, is multi-user, because almost every PC has multiple users. A lot of those people I mentioned are, unfortunately, clueless; they labour under the misapprehension that “multi-user” means multiple users who access the machine remotely, what’s called a “multi-seat” or a “time-share” setup. When I explain the difference, the usual reaction is “Oh! Right!”.
I mean, sure, some people on here want Windows 98 or Classic Mac, but this is the year 2023, and Haiku is an operating system for the year 2023.
I mean, sure, some people on here want Windows 98 or Classic Mac, but this is the year 2023, and Haiku is an operating system for the year 2023.
Any development on Haiku that doesn’t go towards capitalising on its strengths strikes me as reinventing the wheel and diluting the product. The incumbent Unix-like OSes already handle networking, security, GL and general purpose GUI programs very well on most hardware. We don’t stand to benefit from another OS that does more of the same, we stand to benefit from an OS that can do some tasks better on some hardware compared to the incumbents. The developers shouldn’t be afraid of scrapping code that gets in the way of the aspects of Haiku that are or could be outstanding.
From what I’ve read about multi-user support, it’s probably not something that’ll make the cut for R1, but maybe R2. So, see you in 20 years time?
I guess we move in different circles. 99% of desktop PCs I have ever seen, at home or in the workplace, have exactly two user accounts installed: The Superuser/Administrator and one regular user account. The days of having separate accounts for each family member are far behind us. If my kid needs a laptop today, I just buy him one.
If someone gained remote access to my OpenBSD system while I was logged in as an ordinary user, they would only be able to tamper with files owned by the user. Tampering with files owned by root would require the other party to have either the root or the user’s password (if doas(1) is used). The passwords are stored in hashed form in the /etc/master.passwd file which, you guessed it, is owned by the root user. So to summarise, the attack surface can be minimised in a multiuser OS.
Security? Well, I want the security of knowing that my fellow luser isn’t minutely aware of my preferences in music and porn, as well as the status of my lawsuits, because he spends every waking minute snooping in my ~ folder.
So the OS and applications, which I can reinstall in a day, are important and must be protected at all costs. But my tax returns for the last twenty years, the baby photos of my son, my wedding video, my LIFE, are vulnerable? That is the kind of sysop mentality Haiku needs to look beyond.
That’s why multiuser is only one piece of OpenBSD’s armour. Besides, there’s nothing stopping you from taking your sensitive docs offline and only giving them internet access when absolutely necessary, or even assigning them a different user.
Anyway, I didn’t come here specifically to spread propaganda about my favourite OSes. I came here to persuade that Haiku is best suited to being a companion to those OSes rather than a replacement, and it needn’t be concerned with security when there are already many OSes that are.
how about realism idea?
Haiku only have several software or hardware supported.
i find out many mini-pc which used for industral or businese are so suit for Haiku.
they only need vga touch-screen and one or two software.( just like Purchase, sale and storage system。
we develop some important software and hardware.
then, Haiku can live in these speical market.
i remember there is a os named qnx which only live in auto car.
As long as it is my laptop where I run Haiku, good idea. But I don’t want to replace my hardware.
Other developers think the same, and also write drivers for their own devices.
That is the realistic option: let people fix Haiku on whatever hardware they want to use it on. Otherwise, they will not use it and also not fix it for anyone else.
Other developers think
not everyone is developer.
how about this idea:
if one person want to run Haiku in their computer, he or she ,can mail this computer to Haiku living developer.
30 living developers,then,send 30 computer.
50 living developers,then send 50 computer.
take these computer as donate。 is the idea smart?!
I have my own pile of hardware I plan to work on already. But I have no time to do it
Because Haiku does not earn me enough money and I have to work a full time job alongside it. Otherwise, it would be possible.
Also, I chose my hardwqre carefully with my own needs. I wanted a 13’ laptop (not too large so I can carry it everywhere I go, not too small so I can work on it comfortably). But other developers have different needs, such as very powerful desktop workstations. So, you can’t make everyone happy with a single machine.
But you can do it the other way: have a look at the compatibility list from BeSly, and pick your hardware from the list of supported machines if you can. Or, otherwise, add new entries to the list
is the idea smart?!
or just send some computer to haiku as donate, then pray .