Disappointed with Haiku


#61

The ones you need a patched-like-hell kernel to run real-time on Linux, you mean?
Audio and VFX seems to be the realm of MacOS, these days. At least that’s what I see used there most often.

In fact Linux still has no “killer app” (or “tractor app” as JLG prefers to say). And this is why it’s still getting <1% of desktop machines market. Ubuntu (and before it, Mandrake and Corel Linux) helped in making it at least reasonably possible to install by regular users, and other distros have made progress on that since then. But it still has the problem of wanting to package everything in a single repo, and giving you the choice of a stable system with outdated software, or up to date software on an unstable base full of bugs. When they solve this, we may get somewhere with Linux.


#62

Linux has a lot of good applications. Just look at all the applications coming to Haiku now because of the Qt5 port. They were originally made for Linux. There are also other ones that haven’t been ported, like GIMP, Libreoffice, Qtractor and a number of LV2 synths.

The problem with Linux is, and always will be, the OS itself. When the same application is available for both Windows and Linux, everyone will choose to use it on Windows, because nobody actually enjoys using Linux.

I still to this day can’t use any of Linux’s theoretically great audio applications, simply because I can’t figure out Jack and ALSA.


#63

yeah, those, and it was expensive as all hell.


#64

none of those are linux apps, they’re qt apps and they run on every system that can run qt.


#65

I actually enjoy Linux a lot more than Windows. But it took me years to find the right desktop environment and finetuning it to my needs; something I don’t think should be up to users to do.


#66

If only every comment here were a diff for a bugfix…


#67

You can seriously stop the Linux bashing now, especially with crap like that. Linux has always been about portability. No miracle certain applications got ported (often more bad than good) to other platforms. That’s part of the initial idea to begin with. Linux has much more “killer apps” as you call it most of which you most probably never heard since they keep the Internet and what not else running from the background. Besides the important thing on Linux is how the entire operating system works and flows not that it has platform exclusive applications. Believing this totally misses the point of having Linux in use in the first place.

You always claim Haiku to be about the “usage feeling” but you then go bashing around Linux for not having a “usage feeling” but this is as wrong as it possibly gets.

I’ve no problem with people preferring Haiku over Linux but this wrong bashing really gets annoying by now.


#68

You could perhaps answer the question above

Adrien says Ryan has left, but Haiku Inc’s pages still list him as Treasurer.


#69

Ok, let me fix my messages myself. Ryan is still the treasurer, however he is very busy with other things. Kallisti5 is currently working on the financial report for 2017 under supervision from Ryan.

The problem is that the inc is in quite a bad state with only some people knowing how things are run and what needs to be done. So there is a transition period where they still have to be there and make sure all the processes are documented, etc. Then a new batch of people will be able to keep things running.

Currently I don’t have much time to spare for being part of the Inc myself (already running another nonprofit for unrelated purposes), and I’m also quite unfamiliar with the requirements for an US-based one, so I’m not sure if my experience in these things would even be valuable.

In the long term, I think the inc should get bylaws (they are working on it - slowly), make sure the board is renewed from time to time (let’s say members get elected for 3 years or something), and ideally it would be run by people who are not also our most active devs and sysadmins, because this is one place where people without much technical background could be of great help and there is no reason to keep it in the hands of the devs (and it confuses things as to what the Inc is supposed to do).

If this doesn’t work out, we will have no choice but transfer the assets to another organization, either Haiku-related (that would be the HSA), or more general (SFC or the like?). I don’t know much about how these are run, so before we do this, we should carefully consider it and analyze what it would mean for the project. And even then, someone still has to make decisions on how the money is spent and probably do their part of reporting on it. So it doesn’t even solve the problem automagically - it only reduces the amount of work.


#70

i’m not linux bashing. firefox was always multiplatform. nuke, flare and smoke won oscars and were only offered on linux workstations till recently, so i’m not sure what you’re getting at re: bashing.

the rest of everything you say there is projection, i don’t talk about “usage feel” ever (gonna note here there also isn’t a single linux-exclusive desktop environment so ux conversations don’t belong in a linux chat anyway).


#71

I’m not going to continue this line of talk here but you should go back once and check again what you said since what you claim here right now is orthogonal to what you said earlier: Case Closed.


#72

happy to be of service, i hope you have a wonderful day.


#73

Wait. The linux realtime patches are for hard realtime. Something we don’t even see using a telescope.

I can’t believe how an expert dev like you says things like that.

For soft realtime (what BeOS/Haiku actually targets) the standard linux kernel is far more efficient than what we have.

Yet, we are technically far from JACK and ALSA.


#74

Yes, the Haiku Inc. is the major source of our problems. It was excusable years ago for the good of faith. But now, after so many years of inactivity we are locked down. There’s no way for Haiku to improve without financial support and paid contracts.


#75

Still, it is the kind of things you are expected to mess with when using Linux for realtime audio (wether actually useful or not). We are indeed quite far from it in terms of latency (for various reasons, including the fact that we still ship debug mode kernels and the fact that our audio drivers use insanely large buffers to provide glitch-free audio playback). What I see as a problem in the Linux side is that you need to mess with quite a lot of things. My point is that they still don’t have an “it just works” solution (not that we do any better currently, but at least we don’t pretend to).

No, it is not. If it was, people would have found other ways to achieve things. In the case of contracts, one could start a kickstarter campaign, create a liberapay account, use bountysource, …

In none of these cases Haiku, inc would get in the way. They may even contribute some money to the effort. So, we are not locked down in any way. We even already have another supporting association in the form of the HSA, which already has a sane management. The fact that Haiku inc is still the one advertised on haiku-os website and gets the most donations means that people are still sort-of happy enough with how things are. Otherwise we would just kick the inc out of the project. And frankly, I think the people trying to keep it running would be relieved if they didn’t need to anymore.

Remember that the inc is run by the very same people who write the code and run our servers, in the same conditions (people doing this in their free time, etc). As such I feel it isn’t appropriate to say they are locking things or getting in the way. They are doing their best, as is everyone else.


#76

that’s not what we were talking about.


#77

You must agree that it’s a different point than what you stated before. RT kernel patches has nothing to do with that. Yes, under linux it may be hard with problematic hardware to get things working.

But honestly with all computers I have, I can do an “apt-get install jack2 qjackctl ardour amsynth jack-keyboard” and once started I get everything work without problems. You just have to remember closing youtube tabs in the browser.

Let me state more clearly my vision. I think we are missing a fundamental gearwheel with Haiku Inc. being inactive. For example I would like to have a president of Haiku Inc. (which, I must remember owns the Haiku trademark and makes it a little more than a mere support organization) that acts like a chairman. Axel is one of the best programmers I know, but I don’t think in these years he filled this role very well. Yet, Haiku Inc. should at least do something to fund the developer vision in the project. I stopped to have an excusable view of Haiku Inc., I think they should do something.


#78

I don’t agree on this. It is not up to the inc to define the project vision, and even less so the role of its chairman. The lack of a clearly defined vision is a problem in the Haiku project (as a consequence of “let’s clone BeOS R5”, the initial project vision, being completely irrelevant today, let’s face it), and the inc cannot do anything about it.

In fact this is even by design in the current status of the organization. When it was created, people were worried that someone with enough money would get a right in deciding what should or should not be done. The inc was set up as a proxy: they get the money, but they try as much as possible to not interfere with development. The drawback is that they cannot make any decision by themselves, and requests for funding should come from the Haiku project, or at least, be validated by the project. Which in turn leads to the next problem: the Haiku project is not at all organized in a way that makes taking such decisions possible. It worked when the project was small, but not anymore. There are many contributors, each with different views on what is important and each with their personal interest.

I’ve already stated mine, but let’s do it again because it never hurts. I run Haiku as my main OS because I think it’s what works best for me. Therefore it is my personal interest that it works well with my hardware, and that I can get work done on other things. This is the main reason I keep contributing, and the reason it works is that there are other people also doing similar work. I also find it quite nice that we can help people get starterd with open source, and as a result I’m involved in GSoC and GCI. And finally, it is a great learning experience for me, both in technical aspects and in how a team like that is run and how things can keep running despite all the problems.

As you can probably see, this vision is very personal. I don’t have a grand world domination plan or something. I just do what I need. But I think my work is appreciated here.

Of course, some things would be easier if you could just turn out to some lead developer and ask him (or her), “is that the Haiku Way?”. But, we don’t have such a person around. No one is responsible for the Haiku Way. And this is what makes the project so interesting to work with, because everyone can make the idea his own, and by doing so, contribute to the shared vision. The drawback is a lot of inertia when we need to change the vision, because no one dares to push it too far in one direction or another. And this is how we end up still defending gcc2 and BeOS binary compatibility support in 2018.

I think we can manage without a leader, still. But we may need to stop for a bit and think together about where we want to do. Each of us devs should come up with a short description of his personal vision. We will probably be surprised by how much each of us has a different idea. But I think we can still agree on some shared things. And then we can tell the inc - and the devs - “this is what we’re trying to do”. Then people can contribute and help making the vision a reality.

Yes, it is harder than just nominating someone as responsible for the vision and letting him decide. But taking part in such a collective management and seeing it succeed is one of the reasons I enjoy working in the Haiku project so much. And from what I know about them, I think many of the team members agree on this.

There is also a reason for why we have a team like that: our lack of a well-defined vision means that potential contributors who expect a clear direction and instructions usually feel lost in the way Haiku is organized, and will not manage to contribute in useful ways (I also see this happen in other places where I tend to work in the same way). So they will not stick with the project and find something else where they can contribute. As a result, someone who would be more at ease with a more organized team is unlikely to start contributing to Haiku, and even less likely to manage to take the lead on anything. The only way this would happen would be someone forking the project and putting his vision on his fork, and managing to attract more traction than the current Haiku. But so far I have seen no one seriously attempting this, so it seems, once again, that the current model works just well enough to keep people working that way.


#79

It’s interesting that you bring up so many points that are my contentions. But my biggest issue is that Haiku has NEVER focused on being completed… for the sake of Haiku. Haiku is a project. But Haiku is also a vision. A vision of what BeOS R5 was.

In the way things were left to just form as each person (dev) chose to go, Haiku has suffered the “cart before the horse” and “two steps forward, one step back” type situations year after year. Is it any wonder we aren’t even to beta yet, lo 18 years later. Progress is made, then regressions pop up. What worked months ago gets broken. Etc.

I have houses I rent out. They are inanimate structures. They are not alive. They are not conscious. They serve only to house people. Yet, the money I invest into them, is to honor THEM as well as the people living IN them. A house should be cared for, just because of its potential AS a house, and in doing so you honor the tenants in them and thereby receive the potential of financial gain.

In honoring Haiku x86, by reaching the very goal it was created for, you honor its users and thereby receive respect and honor for that completion. Everyone benefits and everyone can be happy.

And HaikuOS (to differentiate from the non-profit Haiku Inc.) should be honored in the that way, by completing its original goal. By whatever means necessary. Finish Haiku R1 x86. Stop the tinkering. Stop the add-ons. Bring it to parity with BeOS R5 (at this point, just making it reliably stable should be sufficient). And call it done.

THEN let’s move on… and improve and tinker and add-on…


#80

Please not. Show respect for the project and call it using it’s name. Otherwise I’ll call you LuposianOS from now on, because why not?

Greetings,
PulkoMandyOS