What about bounty? 3D acceleration bounty?


Correct me if I’m wrong, but most 3d development and other development benefitting from acceleration is handled through an abstraction layer, such as OpenGL. There really isn’t much stopping developers from working on stuff that benefits from acceleration. It just won’t be accelerated until we get the driver support for it. If Haiku can show a relatively quick release schedule, it would show some seriousness to getting a stable platform that can get acceleration drivers in the near future. I wish I had kept a log of these discussions, but I see more potential devs turned off because they see a lack of releases as a dead system. Nobody wants to work on a dead system. Very few mentioned lack of features such as acceleration. Developers generally don’t see lack of features as a turnoff. They generally see lack of features as an opportunity. They’re more interested in a solid system in active development.


I think a lot of potential devs aren’t waiting for hardware acceleration or R1 stable. I think Beta2 and a release candidate would be sufficient to attract some much needed help. Even if beta 2 is just a few bug fixes better than Beta1. I think that delivering the next release in a timely manner would show some meaningfulness about Haiku to the world. Of course this would mean a lot less if beta 2 took years to release.


Private opinion:
I am not a dev and i don’t care about things like 3D.
You should maybe understand, there is at least somebody who don’t give a flying fuck about your ideas, only want to use his computer, nothing less, nothing more, without theater.
You can tell, devs aren’t coming because this or that, but i don’t care. Why should i?
Tell what you want and work on it, because telling: you guys have no $Y, thus i cant do $Z is like telling: there is no O2 at the surface of the moon, our astronauts will die for sure.
I try to help where i can, but your “everything, right nao” mindset remembers to something, but not to entrepreneur mindset.
If they doesn’t comes, thats bad, but not a big deal, so don’t try to sell it as a general management problem. Show somebody or at least comments where they says: Haiku is no go because missing $Y, and try to do not reformat that as a general expectation.
Thats a private comment, BTW.


And I will copy and paste my same arguments as to why it doesn’t make much sense from the last topic about that. Will you actually respond to them this time?


What does this sentence even mean?!

GLX is for … X11. So it’s irrelevant.

KHR comes with hardware backends I believe. So we don’t really need to worry about that for now.

Vulkan drivers basically come for free once you get Gallium-based hardware acceleration in the first place. The kernel hardware drivers are the tough part and they are shared across OpenGL, Vulkan, Mesa’s “Nine”, etc.

This is correct. The APIs are pretty much there, they just aren’t hardware-accelerated.


Sorry about the missing word ‘apps’ :unamused:


I think the 3d is the more important lack in haiku actually… some bug fix too but… ok the second important lack. all the importan things was made already.


You never made a point on why the Binder isn’t appropriate. You just said random stuff, that it was different, unless I pointed out that the Binder is designed around the same concepts of Haiku but just a bit more evolved.


Stop dreaming. We got promised this for every release we made so far, and it didn’t really work.

These people are just making excuses for not contributing now. And a system without devs is dead, not the opposite. If people just wait for other to do the work before they join, whatever we do on our side won’t move them. So we work with the people who are already here to try to bring things forward. And too bad for the others, they’re missing most of the fun!


I must agree with PulkoMandy here. It would have worked a few years ago, when Haiku had momentum, as well as the general Open Source momentum, that’s more or less around when Alpha4 was released. At this point I think we are left with what we have.


I don’t think I’m dreaming here. The lack of releases up to Beta1 made the appearance of a dead system. The release of Beta1 just showed the potential for revival. A second Beta in the next year would actually show a renewed momentum. Different devs are attracted to different stages of development. Some perhaps are more willing to work within Beta and RC stages. What’s wrong with wanting to attract those devs?


I agree with PulkoMandy to an extent here as well. Especially since he came in around the alpha stages. He and others been working hard since then and from the beginning. Perhaps he’s a bit skeptical of the commitment a newer dev would demonstrate. He has that right.

I think it’s a matter of showing a renewed momentum. Let’s show the world that Beta1 wasn’t just a fluke.


My reasoning is that either people are really interested in an OS like Haiku, in that case they would spend more than 30 seconds examining it and would find the forum or the git logs and notice how active the project is. Or, they are not that interested, and a release (or multiple releases) are not going to change that. They may reluctantly start porting their apps when a lot of users start asking for it.

Anyone sufficiently interested in Haiku knows that we are not dead. From other people I don’t expect much.

We will still try to have shorter release times for the next iterations (probably yearly or so), but as you say, the beta did not have the impact one would have expected so far. I did not expect much from it, anyway. And I don’t expect much more from a beta 2. I think however a working web browser and sound card driver would help a lot in getting people to use Haiku, and that’s what I’m focusing on currently.


Which I think would constitute a big chunk of Beta2. I’m not meaning to argue with you specifically. Your personal focus that you’ve stated is right in line with a beta 2 release. Thank you for what you’ve done so far and the work you continue to do. Same goes to the rest of the team. Web+ is my everyday browser as it currently stands. As Net+ was before it. It’s already a huge step up.


Why was 3D slated for glass elevator?


Because it’s something that would delay R1 and isn’t exactly needed for the OS to function. It is pretty much just something nice to have for gaming and other 3d rendering tasks. From the beginning, it was realized that building an OS from scratch would be a huge task alone. Over 15 years later, this has been proven.


The original goal of Haiku was to replace BeOS R5, feature-for-feature. Anything not fitting this goal was not priorized for Haiku R1.

Back in 2010 we did a review of our goals and extended the scope of R1, including for example wifi, and the ability to update the system more easily (which evolved into the package manager we have today). We already made quite ambitious choices there, which probably delayed R1 by a few years already. It’s a good thing that R1 does not need to wait on 3D acceleration.

This does not mean we prevent anyone from working on 3D acceleration, in fact there was already an attempt at a GSoC project 2 years ago for it, which eventually failed for lack of communication, a disappearing mentor, and various other problems. If someone does the work we’ll be happy, but we think there are more important things to do first, and we don’t require it for R1, which has quite a lot of things to be done already.


Thanks for that short explanation both of you. I’d counter argue that hardware acceleration is now a critical and necessary part of any OS because computing habits have changed, and we need such power for web browsing. The web is unfortunately a bloated mess nowadays.

Rather than argue about that though, what would be the way to get acceleration that allows us to port existing drivers? I’ve checked the documentation in the Haiku handbook but it’s difficult for me to understand. The same goes for the Linux graphics stack; it’s a complicated mess to me. I was trying to learn how acceleration is done in another OS so I could try to copy it in Haiku.

What even is a driver? I know it’s a “low level” part of the graphics stack, but what is it doing? Is it telling the card what voltage to apply where to get a framebuffer, calculate some vertex math, or move data to another area? What’s an accelerant?!

What does the driver communicate with on the other side? Does it talk to OpenGL? Then OpenGL talks to the graphics API in the OS?

Sorry, I’m rambling because this is so confusing to me. I don’t expect answers to those latter questions.


I know I’ve said alot about focus. I’m not trying to discourage attracting devs and funding to get hardware acceleration. I want to see it as much as anyone. What I’m trying to discourage is shifting the focus of the current devs and funding away from their current endeavors. I’m all for a code bounty to attract devs to get more done as long as it isn’t much of a hassle for what’s currently being accomplished.


I think this focus you are talking about is largely imagined, pulkomandy is focusing on web+ and audio, Barret on improvements to media codecs, waddlesplash is all over the place but largely involved in Ethernet/wifi drivers… None of the developers are specifically targeting R1 stability etc…

That said I think the new release branching has potential to improve things as it gives people a stable release to work from while everyone.else can continue with improvements on master or feature branches.