I even afraid that if the versions is renamed before R1, it gives others negative impression of the community failing to achieve R1 standards and trying to hide this fact by renaming.
I would guess that these crashes only happen under specific circumstances, as it runs here for weeks 24/7 pretty much without any crashes. I typically reboot every few months, though. It basically works as a file server, so at least that usage is not a problem. The only issue I see is that the memory usage grows over time.
2 Likes
The only times I see the system crash is when I copy files off a USB drive, unmount the drive, then unplug the drive. The other time is if I try and boot off a USB SSD that’s plugged into a different USB hub.
I think your fears are misplaced. We would simply be stepping into line with what everyone else does.
Things have moved on from Be’s day when release meant buying a CD in a shrink wrapped box with manual. Today’s consumer is primed to expect periodical updates.
Whilst Haiku rightly rejects the current mindset that your desktop looks and works completely different one day to the next according to the whims of Microsoft or Apple, we should acknowledge that even a mature software will always be something of a work in progress.
Some people are cheating; they crash the system on purpose to solve tickets. 
1 Like
I think this is a very reasonable take, a stable-release could happen once these are tackled. There will be always an occasional kernel crash though.
You don’t just DO it. You do it along with a properly crafted press release.
“Our initial goal for Haiku R1 was to create a drop-in replacement for BeOS R5. We reached that point some time ago and indeed, we have surpassed it. But the software world has changed significantly since the project was founded. The idea of a single definitive Release version now seems oddly antiquated, a habit from the old days of software on CD-ROM with a printed manual that just does not work in the world of Free and Open-Source software. Accordingly, we are are moving to the following system …”
2 Likes
https://dev.haiku-os.org/ticket/17331 probably needs done before R1/beta6, definitely before R1
1 Like
This true, but Haiku Inc. would be missing a fundraising opportunity if it didn’t sell a limited-edition “Haiku R1”-branded installer USB stick to celebrate the R1 release.
I know I’d buy one.
1 Like
You can rename it R1 if you want, but on my machine I still get a system that:
- Has no sound
- Has no touchpad
- Crashes at least once a week (it may be a bug in the wifi driver I use)
- Has no proper power management, so no way to put the computer to sleep
- Has a web browser that cannot play youtube videos.
So, my guess is, if we would remove the “beta” label, people will just grab another stick and continue beating Haiku, because that’s not where the problem was all along. In fact, even worse, the beta label sets the expectations, and people trying Haiku will be more tolerant to these problems because it’s a beta version. If you pretend it’s a final release, their expectations will be higher (that they can immediately switch from Windows or Mac OS) and they will be disappointed. If you switch to some kind of yearly/rolling release, that means you give up on the idea of having a roadmap and a vision.
If you want to get people to think of Haiku positively, this is done with good public relations, something we are doing a terrible job at right now. Where’s the roadmap? Where’s the praise for a project that takes time to do things right, affords to have a long-term vision, and also makes it evolve over time but without constantly chasing new things? Where’s the communication on a stable API and ABI that allows old applications to still run unchanged after years, long before their authors have moved on to other things? Where are the announcements of progress towards POSIX compatibility?
Doing all those things would allow to control the narrative a bit, tell the press where we’re headed and what’s happening. If you don’t do that, people will make their own opinion with the info they can easily get at a surface level.
So, we have some options:
- Continue the work as it goes now, with a general focus towards R1 (which is a lot of bugfixes as set in Trac) and with developers still free to work on whatever they think is useful. Seems good to me.
- Define a new roadmap, probably abandon R1, and pick another goal for the next few years. I have no idea what that goal would be if it’s not continuing whatever we’re doing now already. Define a new release scheme to get there.
- Decide that we’re not going anywhere in particular, that bugfixes don’t matter, and that we don’t need a roadmap at all. Releases every year with whatever was completed at that point. No specific stability goals.
If we change version numbers, this is what we’re doing. It’s a deeper change than just how we name versions, or at least I think that’s what people are really asking for. And I would rather not hide it under “it’s just a version number”, because if it is, then it doesn’t matter and there would be no reason to change it.
6 Likes
Any other day I’d tell you that Iceweasel covers that just fine, but, funnily enough, today (hrev59314) Iceweasel crashes on startup with an illegal instruction.
Also, Falkon is so broken atm it can’t even render this forum right and WebPositive is crashing every 5 minutes.
Maybe Gnome Web still works?
(And yes, I know you’re referring specifically to WebPositive. I agree it is in a sorry state, but I wouldn’t say it’s fair to expect it to be up to par with the mainstream browsers, not yet at least.)
About the actual topic:
I think it’s funny Haiku is in a much better state than a lot of OSes that were commercially released (I’m thinking of all the Win16/Win32 line of OSes) but that’s a moot point.
Computers have grown much more complex over the decades and an OS is now supposed to handle so many components in the background without hiccups and crashes.
Now, you can point and laugh even at recent successful commercial OSes that fail spectacularly at that but that doesn’t mean Haiku is free to do the same.
I’d rather have Haiku in a beta limbo until it’s 100% done than have it be released today and be rightfully bashed for not being actually ready.
(Which is why I think dropping the “Beta” status from releases would be a mistake, as “Beta” is all you need to say to acknowledge Haiku’s current status. Removing it would just hide that fact from potential users)
And if you think it’s actually ready, sorry, it’s simply not. The fact alone that most web browsers on it are broken atm is proof enough. I say that while having used it as a daily driver for months now.
In the end, if I wanted an half-arsed sorry attempt at a modern OS I’d just use Windows 11, and that’s what a lot of people would say about Haiku if R1 released today.
Tl;dr: I very much prefer the “it’s done when it’s done” attitude over releasing an unfinished product just because of peer/market pressure.
2 Likes
I can’t speak to any of the other browsers, but as of yesterday’s hrev at least the LibreWolf port works quite well; there are some interface oddities, like not being able to use the mouse with plugin submenus but keyboard nav works there so they’re still usable.
1 Like
Iceweasel available is enough for preparing Beta 6 to be significant as “First version pf stable build with Iceweasel”. If webcam or Nvidia driver is also completely available when Beta6 is released, that’d even be better in promotion.
My fault; seems I didn’t test yesterday’s changes thoroughly enough. Fixed in hrev59315.
2 Likes
Well, that’s what we’re discussing, isn’t it? Where is that 100% mark? What would it take? I’m not seeing a lot of agreement here.
And by the time we get there, will the goalposts have shifted once again, with new hardware creating new requirements and we find ourselves once more at 85%? Or should we just accept that all software is now in a state of perpetual beta?
It’s the bugtracker list linked above… that is an already agreed upon goal by the devs..
Will Haiku have a major update every year? I don’t think we should base versioning off of a number that changes annually.
Major changes R1, R2. Minor feature updates R1.1, R7.2. Small updates and bug fixes R2.1.7, etc. It doesn’t arbitrarily change version numbers based on date, but makes changes based on update. Generally people will get an idea of what they are getting into with an update based on this scheme.
Or we could go the Apple or Windows route and change versioning once in a while. 
Probably not though.
3 Likes
Regarding the release of R1, the before mentioned idea of compatible replacement for BeOS makes sense. but also something BeOS had that Haiku doesn’t seem to have quite yet it hardware and web support on the same level as BeOS. I don’t know what BeOS’s web support was like in its day, but it should be generally functional on Haiku for most use cases. Hardware support wasn’t great with BeOS, but if we can have a similar setup with a good assortment of easy to find computers manufactured between 2010 and 2020/2025, and an official list of supported components then Haiku can probably qualify for R1.
I’m in no rush to have Haiku non-beta. Having good hardware support and web browsers would be great though.