Haiku (OpenBeOS)’s third birthday was a few days ago. While some BeOS parts have been successfully re-implemented so far, these were mostly the ‘trivial’ parts: screensaver kit, printing kit etc. Read more for a mini-editorial.The meat of the matter, the app_server, the interface kit etc, are still pre-pre-alpha. Even the kernel used is NewOS’ kernel, it was not created from scratch by the dev team. However, the filesystem, OpenBFS, was created from scratch, mostly by one of the 2-3 very active and very knowledgable Haiku devs, Axel Dörfler. But the rest of the team seems to lark, and in three years they still haven’t reached a user-usable operating system status (despite the fact that they had to re-implement a documented API instead of inventing it, and despite the fact that the kernel was pretty much ready-made). AtheOS, SkyOS had something usable by a three-years time, and remember, they only employed 1 developer each.
The reason I write all this is because I had enough of people being unhappy everytime I would write that “it took 60 full time Be engineers 10 years to bring BeOS to the BeOS 5 level, do you think that a team of 4-5 devs (the rest volunteers can’t help much, not enough system software knowledge) can do it in 2-3 years in their spare time?” But the fanatics, prefer to only think what they want to think. Whatever suits their beliefs. Instead of embracing YellowTAB’s Zeta OS (which is 100% BeOS, and it IS fully working today — not in 10 years from now), they make it a Zeta Vs Haiku war on each opportunity. This is just silly. Yes, I don’t like Zeta’s questionable legal status either, but when it comes to “whatever works to do my daily job”, Zeta works today, Haiku doesn’t and won’t, for many more years. Make no mistake, I don’t take YellowTAB’s side, as I don’t believe that YellowTAB has a bright market future either, but at least they have something to show, today!
I find it absolutely laughable when BeOS users say “oh, well, Zeta looks ok, but I will wait for Haiku”. So, they prefer to financially crucify a company that offers them today a solution (not a perfect solution by any means, but a usable BeOS solution nontheless), instead of showing their support for *BeOS*. To me, these people are not BeOS users. They are simply “OS-curious”. Once, they had some nice experience with BeOS 5 Free edition, but since then, they have moved on to other OSes. And when a company is serving them the next generation of BeOS, they simply, don’t wanna pay, but they “prefer to wait” years after years for a “free” solution (that might never come). That’s not what I call a “BeOS users, who care about BeOS”. I call it a “user who have been impressed by BeOS once, but he doesn’t care enough anymore”.
And that’s the real death of BeOS. Most people don’t care anymore (and why should they? No one wants to support an OS that sees official OS updates). And if Haiku is “ready” by 2010, no one will care. Too little, too late. What’s the point re-implementing BeOS in a way to even have binary compatibility and targetting the functionality of BeOS 5, when by 2010 Longhorn 2, Mac OS X 12 and a more mature Linux will be available offering out-of-this-world features?
You could always argue that Haiku is just like AtheOS and SkyOS or MenuetOS: a hobby OS, created out of love of studying OS technologies. I am sorry again, but Haiku never had such goals. Haiku is not an experiment, like AtheOS was. Haiku was created with the direct target of replacing BeOS 5 by creating an exact clone of it in order to accomodate the thousands of fleed BeOS users after the demise of Be.
If Haiku was able to release this exact clone one year ago (fully stable that is), they would have a good chance of me calling them “successful” to their goal (and then set their roadmap on catching up with OSX or Longhorn). But even if they release that Haiku 1.0 tomorrow morning, it’s already too late. Haiku does not have the luxury of time anymore (realistically speaking) to achieve its own goals. To my mind, it’s a failure.
Haiku needs to move on, it needs to re-set its goals, simply because its current goal, has already failed through market irrelevance. Timing was important for that goal, and now it’s just too late trying to “sell” a BeOS 5-alike OS to the world. I would suggest creating an OS that tries to innovate and competes with future/modern OSes, while keeping its BeOS roots and code, but not by copying Be’s mistakes and the irrelevant, right now, overall BeOS experience one could get out of a BeOS 5+.
— Eugenia
Hi folks!
Well, keep in mind these things:
1) License: Haiku DON’T need of a new license. MIT license is ok for me and for whole BeOS comunity.
2) app_server vs X: I stay with app_server. I never like of X. To cooler heads: Show me one Linux distro that Linux+X surpass BeOS Kernel+app_server speed to convince me.
3) The kernel choice: NewOS is the most wise thing that Haiku team had taken. And I loved the Axeld’s kernel implementation for Haiku. Go ahead. I trust in you, brother!
4) “Developers, developers, developers”: Axeld, of course, is a demon. But looncaz and others BeOS experienced developers could take a hand (sorry guys, but I’m very bad with C/C++. Maybe Java…).
5) Eugenia, maybe in 50 years I take your level in computer knowledge, but you already was in this community. You know all our hopes in BeOS. Take easy. I know that the market will not, but Haiku is our hope to someday get a thing usable to you review.
6) Finally, OSX and Longhorn: I don’t know in the whole world, but in Brazil Haiku CAN TAKE ALL MARKET in five years! Windows XP is used in less of 20% of the computers here. The mass still uses win98 or win95 or DOS still to run your Clipper/Cobol based apps!!!!. These peoples are afraid with Linux. My boss are afraid to get Linux in our server against Win2000. Do you understand?
Ok folks?
This is my two cents.
Michael Vinícius de Oliveira
~ BlueEyedOS.com Webmaster ~
Heh, want a server BeOS?
Just wait a little bit. One may just kinda appear from “thin air” even before Haiku gets to R1.
There is no sarcasm in my comment 🙂
–The loon
————–
Currently just researching ever more wider aspects of system security, and global electronic security.
You need a solid foundation to build on.
No… not really. BeOS, Windows/DOS, and MacOS (classic) did not have a solid foundation to build on. They started basicly from scratch. Other hobby OSes didn’t have a foundation to build on (AtheOS/Syllalbe, SKyOS). They just took the ideas they liked and worked from there. I mean no disresepct, but they have done a lot more in three years than OpenBeOS/Haiku has.
[…] They gather other interested developers, do a lot of talking but not a lot of developing, and a few months later the project is dead. Too many opinions, too few decisions.
Sure, cloning BeOS provideds a concrete goal to work towards. But what happens once you reach that goal? Then you will have to start making decisions about where to take the OS. Then you may have “Too many opinions, too few decisions.”.
you can’t build anything on top of nothing.
You can. It’s called inventing. Try it sometime.
Heh, want a server BeOS? Just wait a little bit. One may just kinda appear from “thin air” even before Haiku gets to R1.
Umm… that doesn’t make a whole lot of sense. BeOS is not multi user os, not seucre, not all that great at SMP (in 2004 terms, not 1996) and not really that fast…
Wow, I can’t believe that this ugliness is still going on. Let me add to it. First, I have nothing against niche or hobby OSes. They’re fun to mess around with. I think that Syllable and SkyOS are fantastic, and show a lot of promise. Allthough I doubt they’ll ever garner enough regular users to make it worthwhile for people to code/port many applications and drivers to them, I think they still have promise. (Of course Robert Szeleney might be able to code or port all of the apps and driver they need. That man is an absolute coding machine! A freak of nature in the very best sense of the term.) Anyway, getting back on topic, the problem that I have with Haiku is their underlying philosophy. Coding binary compatibility into the OS is time consuming, and as many others have posted on this topic, recreates many of Be’s mistakes. IMNSHO the Haiku team would be spending their time in a more productive manner if they nixed R5 compatibility, and worked on re-implementing the Be APIs and design philosophy into a more feature complete OS. And yes, I know, they’ve stated that Haiku will be “more than just R5”, but binary compatibility is a huge millstone around their necks. It’s the Haiku’s time to spend as they see fit, but I wish that they re-examine their goals and methods for extending the BeOS spirit. As for Eugenia, she has every right to say what she wants. First, although this website is called OSNews, if you look over on the left side of the page, you can clearly see that there are separate sections for News, Interviews, Features, and EDITORIALS. The piece that she wrote on Haiku was clearly marked as an editorial. All of you people whining about “this site is called OSNews, I can’t believe that she wrote such a biased piece!” should really sod off. Or learn to read. Editorial is good, because it invites debate. People who have differing opinions about a subject can have a nice exchange over the merits of their arguments. Editorial brings things into focus so you can evaluate the pros and cons of an issue, whether you happen to agree with the views expressed in the editorial or not. And as for Eugenia’s YT bias, I wouldn’t say that she has been exactly solicitous of YT or given Zeta much of an endorsement. The bulk of her comments on YT seem to be that Zeta is pretty much a dead-end…but at least they have something. You can actually install Zeta and it actually has some improvements over R5. Haiku isn’t even pre-alpha.
Oh. And the most import topic that I forgot:
7) Binary compatibility: Stay with it. BeOS commmunity *REALLY* needs it.
Michael Vinícius de Oliveira
~ BlueEyedOS.com Webmaster ~
Yeah, I bought an OS X machine (laptop) and as soon as my wife buys another x86 and hopefully someone writes a audio driver, I will be running BeR5 Pro on her old laptop.
Yeah, it’s missing some apps and Yeah, it’s missing quite bit of nifty things I get with OS X, but I’ll tell you… What Be does, it does better. I miss writing paper with Productive (I may be the only one), I miss the “no non-sense” speed and UI.
“Keep it simple stupid”
> So, they prefer to financially crucify a company that offers them today a solution … *BeOS*.<
I have been a BeOS user since R3.2, and have financially supported Be, Inc fully. I have never used the PE version to this day. I will not buy a beta of a leaked version of BeOS.
> To me, these people are not BeOS users. They are simply “OS-curious”. Once, they had some nice experience with BeOS 5 Free edition, but since then, they have moved on to other OSes. <
Again, I am a BeOS user, and very active in the BeOS community. Are you now resorting to “racism” of OS-users?
> And when a company is serving them the next generation of BeOS, they simply, don’t wanna pay, but they “prefer to wait” years after years for a “free” solution (that might never come). <
Free and legit are 2 different things. I have gladly paid for all of my BeOS software that has ever been available for purchase. I hope Haiku isn’t free. I’ll gladly pay for the fruits of Haiku team’s labor. YellowTab has simply made a few nice scripts and installers and tried to make Dano0 pull profit. It’s not even a final release yet. Not to mention, it has stability issues and hardwre problems that R5 doesn’t seem to have on any systems here.
> That’s not what I call a “BeOS users, who care about BeOS”. I call it a “user who have been impressed by BeOS once, but he doesn’t care enough anymore”. <
I am still and will always be impressed with BeOS. Sure, it may get surpassed, but as with OS/2, it had a time and nichend the way it interjected itself into OS history is nothing short of phenominal.
Think it’s great to see this much zealotry for BeOS. 200+ posts. It’s good that it not only NOT be forgotten, but kept alive with pretty much the same level as fanticism as Linux/BSD. Kept alive as best it can be, at least. Some people think that zealotry is a bad thing. Sometimes it gets out of hand, but for the most part it’s what this whole OS/computer hobby is all about. If you’re not enthuistic about your machine/OS, then why at are you even at this site? Somebody described BeOS as a ghost town. What a perfect description. For a history buff like me, it’s what makes BeOS so much fun to use. It’s a step back into history. My only regret is jumping on the bandwagon this late (BeOS 5 PE Max Edition V3), but better late than never.
They aren’t showcasing anything because they don’t intend to till its all done. The media kit *IS* functioning completely on the decode side, everything works here fine. You don’t see any screenshots of it around, or a beta pack on BeBits.
The network kit is stable enough to hold cvs, Firefox, Vision, etc. Waldermar Kornewald from the net team has come onto IRC using the Haiku net kit over PPPoE, without any problems. Again, no screenshots, no beta pack
There is a hell of a lot done, working, finished. They aren’t showcasing it because it would raise a sudden expectation that the kernel/app_server would be available very soon afterwards.
> Umm… that doesn’t make a whole lot of sense. BeOS is not multi user os, not seucre, not all that great at SMP (in 2004 terms, not 1996) and not really that fast… <
Is there something somewhere that sais it has to be multi-user to be a great server? OS/2 is still an excellent server, has excellent SMP capability, is fast, and has no multi-user in it’s core. It can be added at the filesystem level, but it’s not neccessary for security.
While I agree with you that people shouldn’t expect everything to be free, don’t you think Zeta is just a tad expensive for what it is?
Yes, Zeta is expensive compared to the free Linux, and the much more powerful OSX (for only $30 more).
But then again, Be engineers estimated once that BeOS 5 should have been sold for $400 for each copy, in order for Be to pay the bills of developing it. So, depends how you see the situation: as a businessman, or as a consumer.
Will BeOSMAX or PhosphurOS or Zeta work on a IBM T41 2373-1FM
Probably. Don’t expect the integrated wireless/bluetooth to work. And the NIC/graphics/sound all need recent drivers (assuming its the same machine I’m thinking of).
why is it that whenever BeOS and X Windows are mentioned in the same sentence people automatically think that you want to replace the BeOS GUI with x Windows?
X is a protocol and just happens to also work on BeOS.
Why use it? Well it allows you to port over a whole swag of GUI apps which you may not have had running under BeOS.
The BeOS GUI is nice and snappy. Yes it has its flaws. X Windows works fine here and has always worked fine on my hardware going all the way back to 1995 when i was using slackware on my Petium 90.
Download the native BeOS version here http://www.bebits.com/app/3606 and give it a try.
You’ll be impressed.
peter
I completely agree with that statement. People saying “it’s not even pre-alpha yet”, maybe it’s because they want you to think that it’s not pre-alpha, when in fact, it could well be close to beta or final release or whatever.
There was a post about “Haiku just springing out of the ground”. That’s probably what’s going to happen in any case.
I’ll be holding my breath.
Chris
It seem like BeOS / Zeta / Haiku is a hot topic that many want to fight to the blood….
BeOS was a real good OS, still is in many aspect. But the fact remains, it’s getting old fast. Even if Haiku can come out with something a bit better than R5/Dano in a couple of years… It will have to compete against Longhorn or MAC OS X Tiger.
Who needs Longhorn or pricy OS X to do some surfing or e-mail? Nobody… BeOS / Haiku can do the job. But do you see BeOS pre-installed in modern new PC? NO… Longhorn will, likw ZP today….
SO… It was good time running BeOS but I guess it’s time to move on. Linux may be a better alternative. It’s Free so much better than Zeta!
No… not really. BeOS, Windows/DOS, and MacOS (classic) did not have a solid foundation to build on. They started basicly from scratch. Other hobby OSes didn’t have a foundation to build on (AtheOS/Syllalbe, SKyOS). They just took the ideas they liked and worked from there. I mean no disresepct, but they have done a lot more in three years than OpenBeOS/Haiku has.
From looking at their old websites, I’d say that three years ago both AtheOS and SkyOS were quite far in their development. Sure they didn’t support the latest of the latest, but that’s not really my point. Have they really gone far in those last three years? And how did they get where they were in 2001? SkyOS began somewhere around 1996. I don’t know about AtheOS, but I’m quite certain it didn’t just pop into existance somewhere around 2000. In fact:
“Q: How long have AtheOS been in development?
A: It seems to be a lot of confusion about how long AtheOS have been in
development. I have worked on it for about 4 years. I set up the web server
about 6 months ago, and the first installable release was put on the server
about 2 months ago.”
And that was in August of the year 2000.
Certainly those projects are great and amazing, but don’t forget they had a history and a foundation on which they could build whatever great things they’ve done in the past three years.
Sure, cloning BeOS provideds a concrete goal to work towards. But what happens once you reach that goal? Then you will have to start making decisions about where to take the OS. Then you may have “Too many opinions, too few decisions.”.
But you should agree, that by then they have completed a set of goals already and won’t have dwindled away before getting on their merry way. There will also be experience with developing an OS and their various subsystems and how to go about designing them. Subsystems are available, waiting to be explored, improved or replaced. And while plodding away on R1, there is a separate project to discuss these new ideas and opinions for beyond R1, namely
http://glasselevator.sourceforge.net/
with special interest the mailinglist
http://www.bug-br.org.br/mailman/listinfo/glasselevator-talk
and hopefully reach some sort of conclusion/concenses before it’s time to decide on a set of goals for a next release.
mahlzeit said: “you can’t build anything on top of nothing.”
On which you replied: You can. It’s called inventing. Try it sometime.
I’d say, although it might be a question of symantics, that if there was nothing before and there is now, you can have done several things:
(1) discovered it accidentally or purposely by setting out to discover something, but having the foggiest about what you will discover.
(2) created it, like some kind of deity.
(3) not worked in a total void, but purposely set out to design and build or, if you prefer the term, invent this thing equipped with tools, knowledge of your enviroment (where, for example, the laws of physics can be of great help or an API or platform(x86/ppc/..)), knowledge of your raw materials (ideas, pieces of equipment you’ve seen before (dynamo/clock/radio/..)).
“If people like Eugenia, Jean-Baptiste Queru, and Scott Hacker (author of the BeOS Bible, people who were so heavily invested in BeOS, have given up on BeOS and have moved on to OS X, Windows XP, and Linux, I really can’t think of a reason to keep using BeOS.”
Now there’s a sheep if I ever saw one.
As for the people wanting Haiku to drop its kernel and use Linux/BSD and X server – you’re on crack right ?!?
Eugenia gave an opinion in this piece. The article was articulate and to the point.
More than a few of us got help, answers we needed or at least got steered in the right direction from a post answered by Eugenia.
I don’t know if I agree or not, but when I consider the source, I will be certainly thinking about it all over the next few days.
No… not really. BeOS, Windows/DOS, and MacOS (classic) did not have a solid foundation to build on. They started basicly from scratch.
MSDOS was based on CP/M IIRC. GNU and Linux were from scratch, but with references. Stallman replaced apps hand by hand. It’s not hard to imagine such a process is similar as Haiku’s. Stallman started in 1984 and it took him some years whereas Linux started in 1991. By that time, GNU was basically complete (usable) except for the kernel. The Linux kernel took some time too, and its hard to say when it was “ready” (for what purpose is also a debate worth) so i tend to agree with the person who says pre-alpha code is merely a perception, an imagination, an opinion. Not some hard undoubtable fact.
But as a general statement, yours is still far off. How do you define “solid”? And, science is based on science (proven logic). _V_ addressed this already.
First, although this website is called OSNews, if you look over on the left side of the page, you can clearly see that there are separate sections for News, Interviews, Features, and EDITORIALS. The piece that she wrote on Haiku was clearly marked as an editorial. All of you people whining about “this site is called OSNews, I can’t believe that she wrote such a biased piece!” should really sod off. Or learn to read. Editorial is good, because it invites debate. People who have differing opinions about a subject can have a nice exchange over the merits of their arguments. Editorial brings things into focus so you can evaluate the pros and cons of an issue, whether you happen to agree with the views expressed in the editorial or not.
Ok, so a good column or editorial is actually a controversial opinion to which the line between that and a troll is thin for some. The obvious side benefit is ofcourse more hits hence more revenue from ads.
Sorry, but your point of view is IMO completely bull, because it argues we ought to justify any editorial for the sake of being one. Just like we have to accept an opinion being promoted for the sake of it being an opinion. Well, think about it, and where one opinion is made another one ain’t read cause it takes up space and takes time to read.
I’ve read a damn good number of columns and editorials, and this one ain’t among them. Instead, a good editorial would provide references which support the points the author tries to make. It should try not to be. Finally, some mix of humor or procative reasoning doesn’t hurt. This piece was saltless, had only 1 reference, and was filled by unproven ranting. Sorry, no go, at least not for me.
“It should try not to be.”
Scrap that.
“procative”
Provoking.
<<Umm… that doesn’t make a whole lot of sense. BeOS is not multi user os, not seucre, not all that great at SMP (in 2004 terms, not
1996) and not really that fast..>>
BeOS actually does.. well.. ONE distro.
I use http://www.grc.com to help in my quests for port-stealthing, and have even launched attacks over my NETWORK to try and crash server apps, the system, or whatever. The fruits of my labor may be available as early as Beta 6, or likely as an extra application.. called a firewall.. of course, anyone running a server will also very likely have a hardware firewall 🙂 Just makes sense 🙂
Multi-user is already implemented, and I have been discovering more and more of the integrated user permissions et al that BeOS already has. Including the kernel, which does respect ownership.
I may not seem the type, but I have a small amount of experience with servers and server environments. I know what needs to be added to make the cut for low-end serving purposes, but I don’t really believe that a BeOS derivitive will ever trully be a great huge success in the server arena. Access to Haiku’s code however could make it simple, in comparison, to make a server distro.
And SMP performance sucks, eh? Hmm… care to show proof? Last I checked on my Dual-CPU server which dual boots three OSes (Linux-Mandrake, Windows 2003 Enterprise, and PhosphurOS Server Exp/2). Of those three, I really cannot actually find a way to make a good comparison that I can trust.
In terms of smoothness, there is no competition.. PhosphurOS is certainly the smoothest in terms of performance. In stability..well, only Windows has crashed on me (and it was Windows, not an application.. format reinstall fixed in only a couple hours of downtime).
I guess I will have to write a cross-platform SMP benchmarking suite to try and determine exact performance of say… 500 threads doing a loop of very different tasks.
100 threads accessing files,creating, writing, reading, then removing them.
100 Adding 64-bit integers together and then doing an error check.
100 writing blocks of data into memory (8 to 20MB blocks) and removing them.
100 threads doing floating point math
100 threads monitoring the status of the other threads
Set all threads to the same priority, except the threads the watcher threads. Set priority to the lowest priority each system offers. Eleminate other running processes. And then run the program using time $(my_crazy_benchmark_smp) to get the exact time spent for the tasks.
And of cf course the monitoring threads will log and calculate average performance for each of the four sets of threads, and display this after all processes are complete giving times and average.
I would write the application seperate for each OS, using eachs own preferred methods for all my tasks. So pthreads for Linux, Be threads for PhOS, and whatever it is for Windows, and so on.
Run the app on the EXACT computer, only restarting to the other OS between runs. Well, actually… Linux won’t start after Windows unloads, so I have to cold-boot each time.
Run the program in concession three times, and then compile the results and see which can handle the tasks at hand the best. This could be fun 🙂
And, the version of Linux I am not going to take into account, so don’t even go there. The PhOS version was made while I was at my early days at yellowTAB. I finished before Zeta had a name and it had been agreed to make a distro based on Dano. I have not modified it in anyway, and really couldn’t do much for performance anyway. SO three out of the BOX OSes, let us see how they work. (The Linux kernel is 2.4 something AFAIK).
Someone want to make some code up for the ptreads and the Windows parts of this app? I can detail exactly what number should be added, how many times, and what should be fed into memory how often. (the memory threads will run one after another, one creating the memory, the next seeing it is there, then the last freeing that memory).
I can code this in a few hours here for PhOS, but that is just me 🙂
–The loon
Oh and Kevin… try PhOS 🙂
Is there something somewhere that sais it has to be multi-user to be a great server? OS/2 is still an excellent server, has excellent SMP capability, is fast, and has no multi-user in it’s core. It can be added at the filesystem level, but it’s not neccessary for security.
No, your right. You don’t have to have multi user to be a good server. However, in most cases it doesn’t help security wise, which is an area BeOS is not very strong in.
Also, OS/2 has qualitys (SMP and speed) that make it a good server. BeOS just doesn’t have thoese qualities. As I understaqnd it, BeOS;s SMP is better than Win98 but no where near linux or win2k/XP/03server. Also, BeOS actually isn’t that fast, it’s GUI is just really responsive.
Basicly, BeOS is not a good choice to turn into a server OS. I’m not saying it’s not possible, but it’s not very logical. If someone is going to spend time working on BeOS, why not contribute to Haiku instead?
why do you think beos isnt fast? I’m not flaming, I just want ot know why you think that.
He probably means fast as in database access times, and such. In which case, he use to be right 🙂 That was before we got database access working.. you know.. SQL…. Of course, we only have SQLLite, so in that type of environment, we really can’t contend yet.
–The loon
What I don’t really understand, is why someone doesn’t develop a linux distribution that supports the same goals as BeOS.
The 2.6 kernel can be configured as a highly desktop-specific piece of software; the scheduler can be tinkered to suit, a single on-disk file system type could be supported, ensuring very low latency results. The remainder of the system should be POSIX-compliant to allow the use of any appropriate libraries, MTAs, services etc, that might be required.
Then, a single GUI toolkit could be selected as the only valid GUI system for the new distro – for argument’s sake, say Gtk+. A theme engine could be tuned to ensure that this GUI was as close to the original BeOS GUI as necessary, and all configurability compiled out. Using Gtk+, you could then determine whether to use a X11.org rendering subsystem, or direct to frame-buffer.
Choose a single blessed set of applications, and maintain a patch set for those which integrates them perfectly into the streamlined GUI system that you’re supporting. Each app should be modified to have complete support for extended attributes, and these should be as essential to system operation as they were with BeOS.
Now, I must admit to never have used BeOS much – but what I’ve suggested seems to maintain the aspects of the system which made it desirable in the first place.
> Coding binary compatibility into the OS is time consuming, and as many others have
> posted on this topic, recreates many of Be’s mistakes. IMNSHO the Haiku team would
> be spending their time in a more productive manner if they nixed R5 compatibility,
> and worked on re-implementing the Be APIs and design philosophy into a more feature
> complete OS
And as many others who has posted on this topic you have no idea what Haiku’s goal are. Release 1 isnt the end, it’s the beginning. R1 is the stepping stone towards future improvements. This concept is usually referred to as “building on a solid foundation”. Binary compatibilty will most likely get dropped on later releases but has been deemed necessary for r1 in order for it to be more usable.
> The bulk of her comments on YT seem to be that Zeta is pretty much a
> dead-end…but at least they have something.
Somehow investing in a “dead end” with questionable legality doesnt seem like a good idea to me.
> You can actually install Zeta and it actually has some improvements over R5.
I can install a good portion of Haiku on r5 and get improvements. Sort of like how Zeta uses portions of Haiku on Dano…
> Haiku isn’t even pre-alpha.
I guess you know that from your deep involvement in said project, right?
Still holding out even if others are loosing faith. I know I have not come across a better User orientated OS. I have serious doubts Microsoft/Apple/Linux will in the next few years be able to deliver something that is elegant, powerful and easy to use.
As always it’s a let see and evaluate technology when it is released approach that’s best in the computing world. That goes for Windows as much as it does any otehr piece of code. Anyone buying into hype and marketing needs thier head read as we all know from experience that most products hyped usually fail to hold up in the real world.
“Now there’s a sheep if I ever saw one.”
There is a huge difference between having a stong opinion on a subject, and being an extremely rude person who resorts to insulting people because you are incapable of using reason and logic when trying to argue a position. In all of my posts on this topic, I’ve tried to keep any criticisms I have confined only to the implementation the Haiku project, not to the people who are coding it. If anyone took any of my criticisms to heart, let me offer my most sincere apologies. I admire the dedication of the Haiku team, I just think they should make adjustments to the project goals. As far as me being sheep like, I currently run Mac OS X, Debian, FreeBSD, Windows XP, Solaris, Irix, and BeOS at home. I like to experiment with many different OSes and am not that interested what everyone else is using. My point was simply to illustrate that a lot of the people who were deeply involved with BeOS,people who made a living either writing about, or actually coding the OS, no longer find value in it as everyday OS. Their love for the BeOS doesn’t blind them to the fact that there are massive deficiencies in the BeOS that need to be addressed. Whenever I read an interview with an ex Be engineer on BeOS Journal (sorry, Haiku Journal), ZetaNews, OSNews, whatever, they always seem to be amused that people are still using BeOS as their main OS. BeOS was very cool, and there were some excellent features that it had that would be nice to see incorporated into a viable OS, but the strengths of the BeOS no longer outweigh its weaknesses.
> My point was simply to illustrate that a lot of the people who were deeply involved
> with BeOS,people who made a living either writing about, or actually
> coding the OS, no longer find value in it as everyday OS.
But this is a rather weak point, wouldnt you agree? I’m sure we can find people who have been deeply involved in Windows who doesnt find any value it in any more. Or people who previously used Linux but have changed to Windows because Linux have no value in their everyday use. Or from FreeBSD to Linux, or vice versa or..etc etc, you get the point.
An OS isnt a mean unto itself, it’s a means to accomplish tasks. The merit of an OS should be valued on it’s ability to perform a task and not on wether random person Y uses it or likes or or have some other opinion on it.
Is BeOS today suitable for every possible desktop task? Most certianly not. Will it ever be? Hopefully not. If you try to please everyone all the time all you’ll end up with is a bland product that noone really likes. I guess that’s fine if all you’re interested in is making loads of money but it’s not suitable if you actually care about the quality and reputation of your product.
> Their love for the BeOS doesn’t blind them to the fact that there are
> massive deficiencies in the BeOS that need to be addressed.
Naturally there are deficiences, that’s what Haiku is fixing on a small scale now and will do on a larger scale *later*.
Linux, for example, didnt become what it is today overnight. If we look back I am certain we can find almost the exact same arguments against Linux. Things take time, not everything can or should be done overnight.
Tqh wrote:
> I do not have a problem with GPL itself. Just how some suggests it should be used.
It’ s because as far as I see, that “train”, generally speaking , runs faster that BSD|MIT.
I have nothing against BSD|MIT, only that psicology, is also important when you have to do something big.
Maybe , for someone, GPL could be seen as a fashion, but it just works.
Now, there is the problem of viral GPL, but with a wise use of LPGL, it can be resolved.
For example, if I want to develop a commercial application in Haiku (guessing it is completed and all MIT), I can without problems.
If I want to develop a closed application for Linux with Ncurses libraries , I can’t because, afaik, they are GPLed.
If they were LPGL, I could do that!
So, MIT vs GPL-LPGL (wise), I think the second is better
for fashion and so fly on that wave 😉
In addition, there wouldn’ t be problems for third party commercial application (or driver) developers, and that thanks to LPGL.
All IMHO.
Ever noticed how all those “big” projects, ie Apache, XFree, Perl, Python, Sendmail etc, isnt under the GPL or LGPL?
The Haiku license is decided, it’s BSD/MIT, it wont change. There’s no point in arguing for or against this choice. If this is such a big issue then tough luck, I guess the Haiku project isn’t for you.
BeOS has PostrgreSQL as well as SQLite, and Postrgre is at or above MySQL (apparently, don’t start another flame war people). So that could be used to test.
Well, mine was only a thought about how useful could have been LPGL;
sure it wasn’ t meant to be the ultimate Truth…
as you wrote:
> The Haiku license is decided, it’s BSD/MIT, it wont change
Les jeux sont faits
so I do my best wishes to Haiku team to achieve something important 🙂
Who ported PostgreSQL to BeOS ?
Eugenia Loli Queru of course
http://www.bebits.com/app/1723
Thankyou Eugenia
error…
Who ported PostgreSQL to BeOS ?
“This port was not made by me.
Cyril Velter has ported it in 1999”
“Ever noticed how all those “big” projects, ie Apache, XFree, Perl, Python, Sendmail etc, isnt under the GPL or LGPL?”
A handful of projects. Ever noticed that there are far more GPLed apps? Ever noticed the license ratio browsing Freshmeat or SourceForge?
“The Haiku license is decided, it’s BSD/MIT, it wont change. There’s no point in arguing for or against this choice.”
Yeah, no point in arguing. No point in intelligently considering the pros-and-cons to a license; let’s all just do like you, put our fingers in our ears, and pretend everything is fine.
Go look at the state of Haiku development, and tell me there’s not a problem. They’re severely short of developers, and they’ve not chosen a popular license — indeed, they’ve chosen one in which contributor’s code can be taken by companies pushing closed products.
A license change could bring about many new developers. But no, you think it’s not worth talking about, and would rather let Haiku stagnate. Wow…
Changing to the GPL would lose some of the current developers. Although my contribution has been measly, I would completely stop contributing to the project if it was GPL. I don’t know about the other developers, but I’m sure some of them have the same feelings about it.
> A handful of projects. Ever noticed that there are far more GPLed apps?
> Ever noticed the license ratio browsing Freshmeat or SourceForge?
Quantity doesnt equal quality but I’ll be sure to inform FreeBSD, OpenBSD and every other non-GPL project that their efforts are in vain. They should no longer care about doing what they want or what they feel is right because, hey, there are more GPL project on Freshmeat and if they switch to the GPL more people will like them.
> No point in intelligently considering the pros-and-cons to a license; let’s
> all just do like you, put our fingers in our ears, and pretend everything
> is fine.
Do you seriously think this hasnt already been considered over and over again a long time ago? This wasnt decided on a whim over a lunchbreak or on the subway on the way home from work. It’s a concious decision so that the code will be completely free and can be used by as many peple and companies as possible.
> Go look at the state of Haiku development, and tell me there’s not a problem.
Thank you, I am well aware of the state of the Haiku project and what problems there are.
However, I, and obviously the Haiku Team, dont think those problems can be best fixed by changing license. GPL isnt a magic bullet. When the going gets tough the tough keeps on going…and the rest changes license in desperate attempts at gaining attention.
> indeed, they’ve chosen one in which contributor’s code can be taken by
> companies pushing closed products.
I’ve never understood this obsession with “companies pushing closed products”. You do realize that this is possible also with the LGPL, right?
I, for one, dont care if company X is modifying my code and make a closed product of it as long as they dont claim copyright and original authorship of the code. The BSD/MIT license “guarantees” just this, that copyright is maintained and credit given.
Different folks, different strokes. Not everyone subscribes to the views of the GNU foundation.
Obviously these are views shared by the core Haiku team otherwise it would probably have been using GPL or LPGL.
“I’ll be sure to inform FreeBSD, OpenBSD and every other non-GPL project that their efforts are in vain. They should no longer care about doing what they want or what they feel is right because, hey, there are more GPL project on Freshmeat and if they switch to the GPL more people will like them.”
Meaningless argument. Those projects already have developers; they’re not in desperate need of any more, so the license is working fine for them. And “like them”? It’s about attracting developers, not making buddies.
“I, for one, dont care if company X is modifying my code and make a closed product of it as long as they dont claim copyright and original authorship of the code.”
That’s fine; I never said one license was better than the other. But that’s just your feeling on this matter — and you can’t speak for the thousands of open source coders out there. The simple fact is, many more open source programmers prefer the GPL than BSD-style licensing. By choosing a less popular license, the Haiku project has limited its possibility to attract new developers. That’s just a fact!
(The other big issue is forking. Should Haiku really take off, having a codebase that can be incorporated into closed products will not be productive — it’ll just lead to incompatibilities and fragmentation akin to UNIX in the early 90s.)
(The other big issue is forking. Should Haiku really take off, having a codebase that can be incorporated into closed products will not be productive — it’ll just lead to incompatibilities and fragmentation akin to UNIX in the early 90s.)
And how does one explain the incompatibilities between various Linux ditributions?
So it seems, MIT or GPL, both create the possibility of forking/spin offs/incompatibilities.
> Meaningless argument. Those projects already have developers; they’re not in
> desperate need of any more, so the license is working fine for them.
Keyword being “any more”. Instead of chickening out and get new developers at the expense of their convictions they stuck to their guns and made it thru the rough times. What’s to say Haiku wont?
> The simple fact is, many more open source programmers prefer the GPL than
> BSD-style licensing. By choosing a less popular license, the Haiku project
> has limited its possibility to attract new developers. That’s just a fact!
Getting GPL-preferring developers is lower on the importance list than keeping Haiku free and BSD/MIT licensed.
I would have hoped that prefering the GPL license didnt make one so narrowminded that one cant contribute to OSS projects using other OSS licenses.
> (The other big issue is forking. Should Haiku really take off, having a
> codebase that can be incorporated into closed products will not be
> productive — it’ll just lead to incompatibilities and fragmentation akin
> to UNIX in the early 90s.)
GPL’ing the Linux kernel sure havent prevented it from being diversed into all too many distributions incompatible with eachother in more or less creative ways.
“they stuck to their guns and made it thru the rough times. What’s to say Haiku wont?”
That’s true, and I wish them all the best! But, equally, we could say those projects could’ve progressed faster had they been under a different license…
“Getting GPL-preferring developers is lower on the importance list than keeping Haiku free and BSD/MIT licensed.”
OK, but how high on the list is attracting new developers? That’s my main point — right now Haiku doesn’t so much need money, or apps, or marketing; it really needs a lot more developers. So doing something to achieve that should be a high priority, and I’m just suggesting a license change could do a lot of good.
“I would have hoped that prefering the GPL license didnt make one so narrowminded that one cant contribute to OSS projects using other OSS licenses.”
Yeah, I’m not saying Haiku should have total GPL zealots on board — not at all. And yep, many hackers do contribute to projects under different licenses. But at the end of the day, the majority of coders prefer contributing to GPLed projects; they want their work always kept open, with community success as the remuneration. To many the terms of the BSD license are less of an incentive.
There are loads of talented coders out there who support the GPL, programmers who could give a lot to Haiku, but are likely to avoid it with the current licensing.
“GPL’ing the Linux kernel sure havent prevented it from being diversed into all too many distributions incompatible with eachother in more or less creative ways.”
Well, you name me three open source apps that only run under one Linux distro (proper apps and utilities, not installers etc.). I’m struggling to think of just one! There’s no incompatibility issue mainly because of the GPL — it means that one distro vendor’s enhancements can be added into others.
I’m not saying BSD-licensed projects are massively prone to forks, but the successful BSD projects have been server-side and geek-targeted. If Haiku takes off, being an end-user desktop OS it’ll have more commercial interest (and therefore incentive for companies to add their own closed and incompatible features).
Yeah, that’s a long way off and I’m just speculating. And I respect your opinion — I just believe licensing should be considered particularly given the projects current status, thinking about how it can help, and make any moves early on before it’s too difficult.
If for some reason they go insane and move to the GPL, I will be requesting any code in the tree I wrote (OK, its the modifications to the Terminal application at the moment, as my sounddrivers aren’t good enough to go in yet) be removed.
GPL will attrat oh, lots of GNU zealots. Maybe. But then again, they’re all stuck to Linux anyway.
“There are loads of talented coders out there who support the GPL, programmers who could give a lot to Haiku, but are likely to avoid it with the current licensing.”
Hm, do you really think that we just need to switch to the GPL and we’ll have a load of developers who jump on board ?
Seriously, grow up. It’s not that we get loads of e-mails like “yeah, I’d really like to contribute but I’m a GPL zealot and will never do it”.
Kian:
“If for some reason they go insane and move to the GPL, I will be requesting any code in the tree I wrote (OK, its the modifications to the Terminal application at the moment, as my sounddrivers aren’t good enough to go in yet) be removed.”
I’m not sure you could do that. You have accepted your code to be licensed under the BSD/MIT license, and you can’t do anything, anymore about it.
The code you’ve “donated” will always be under that license. You can choose not to contribute anymore, though.
“If for some reason they go insane and move to the GPL, I will be requesting any code in the tree I wrote (OK, its the modifications to the Terminal application at the moment, as my sounddrivers aren’t good enough to go in yet) be removed.”
Well, that’s the gamble. Haiku could lose a couple of Terminal patches, but gain some talented kernel and appserver developers who prefer the GPL…
“GPL will attrat oh, lots of GNU zealots.”
What an utterly ridiculous thing to say. Linus licensed his kernel under the GPL — is he a “GNU zealot”? The thousands and thousands of coders who’ve released code under the GPL — are they all “GNU zealots” too? Do you have any idea about how the community works at all?
Most OSS coders like the GPL because they don’t want their work incorporated in closed products. Simple as that. 99% of them aren’t members of, associated with, or interested in the GNU project.
You know, moving to the GPL might attract a bunch of talented, dedicated coders who want to see their work always open, and could really give Haiku some steam. How is that a bad thing?
Or would you rather things stayed as they are, with very few developers, slow progress and some guy doing two-line patches mouthing-off on subjects he doesn’t understand, tarring the vast majority of the open source community with the same brush?
I really hope you don’t represent the Haiku community, because your attitude and lack of understanding isn’t doing it any good.
If gpl is so great why don’t Sun,Apple,and Microsoft use it for their OS ?
Please, if GPL is the solution to all worries, and that many developers are attracted by it, why isn’t there a fork under GPL license already?
If you want a GPL licensed Haiku, you will need to do it yourself. Haiku will stay MIT/GPL.
Wow. An interested Haiku follower notes that the majority of open source developers prefer the GPL, and suggests that a license switch might bring on board new developers and coding talent.
Then a load of you angrily respond with bitter flames, labelling me as some kind of “GNU zealot” and ignoring the issue at hand — that of getting more developers.
(Respect to Lars Hansson for actually thinking about the idea and debating it, rather than ignoring what could potentially be a boost to the project.)
I’ll say no more. It’s your choice. But looking at the responses, it’s no surprise that Haiku is pretty much stagnant and barely making glacial progress…
my post was totally out of line.
If for some reason they go insane and move to the GPL, I will be requesting any code in the tree I wrote […] be removed
So much for the spirit of the MIT/BSD licence…
If you want your code to be free for everyone *except* people happening to like the GPL, you should consider writing your own licensing terms. The BSD licence is designed to allow anyone to do anything with the code as long as credits are given where due. Not to play petty games with the guys you don’t like.
I contributed the code under the assumption that it would be free, and totally free whilst being used by the Haiku project. Moving to GPL stops it being free. I don’t care what third parties do, I just want the code to be under a proper, free licence whilst in the Haiku projects hands. If they want to change it, well, they can either stop using it, or lose a user.
Just to note, I have contributed code to GPL’ed projects. I’ve had patches accepted to Open Transport Tycoon (GPL), and I’ve offered patches to Nano and UnrealIRCD – both not accepted/ignored/whatever – both GPL. I just want and hope Haiku to stay under a real free licence – BSD or MIT.
Actually, i prefer GPL too.
I conisdered several times if i should dig Haiku sources, read some additional books with coding wisdom inside and switch to dedicated Haiku work instead hacking some BeOS apps here and there. And one of my doubts preventing from that step was just BSD license.
Maybe it was caused by YT influence, though.
The appropiate place to make any suggestions to Haiku and its developers is the forum on the Haiku website http://www.haiku-os.org/forums/ or the OpenBeOS discussion channels on irc.freenode.org #openbeos now #haiku or by email.
The appropiate time to have done that re licensing and legal status would have been a long time ago before any legal negotiations and licensing agreements began.Preferably before any developers agree to join the project and commit to working in it.
Telling the people who are working for some years under a different license to change it is not constructive,and this is certainly not the place to do it.You are lucky any Haiku developer or BeOS user has spent the time it takes to read such posts,let alone answer them at all.
Considering its unlikely YT will last the winter, that shouldn’t worry you as much. They’ve got no chance of making real money, their investment is going to be burning fast considering they have full time coders…. They’re going to die.
>They’re going to die.
There is a BeOS based company, paying full-time money, as you put it – to their employees, allowing people to live by coding for BeOS again, and you sound like you wish them to fail.
Shame on you.
Their CTO publicly slandered me as “unsuitable to be a member of the community”. They sell a shoddy product. Of course I “want” them to fail.
I’d prefer not to see some of the developers they have out of a job – theres some very good people there – Rene, Voidref, Francois, etc. But the company as a whole has serious problems – bad management, a bad CTO, bad PR. This will just damage Haiku in the future as people may associate them with YellowTAB and the problems of Zeta.
did i say that?
I rather meant that this company uses my hard work getting all my free time WITHOUT ANY “thank you” but rather presenting results of that work as its own achievement.
See story with Decaf and Firefox credits, at zetanews archives or at bezilla blog. And this is quite common practice fro them, it seems.
I have nothing against their bussiness, lets do what they wanna. I just don’t wish that such companies use my potential work without any obligations in giving something back.
One is Haiku, the other is ReactOS.
I’d like a version of BeOS that runs on modern hardware, has modern drivers, and has a slightly updated UI.
And I’d like a Windows Compatible OS I can run on my older hardware so I can give it away to friends without Pirating a Commercial OS, and have an OS where the programmers believe in making it small and bug-free rather than bloated and buggy.
Haiku IS taking a long time. I’m hoping I’m right in reading between the lines in thinking that a bootable version of Haiku (a major milestone) will be possible in under a year.
ReactOS is currently bootable, but not feature complete and it’s fun to watch it grow.
I think a LOT of the PR problems with Haiku could be solved by more communication. A weekly or monthly newsletter that says what’s happening.
WINE is taking forever, but the communications on the WINE HQ website let’s people know that it’s coming along, and new things are being added every month.
I think this the ONLY think Haiku and other projects do wrong. They put communicating with the user base (or potential user base) very low on the totem pole.
And this is a place for a Non-Programmer, but Web-Savvy conributor to help…
Projects need not only Programmers, the need web developers, documentation writers, help screen writers, Artists, etc…
Uhm, I didn’t mean you.
“Considering its unlikely YT will last the winter”
See, Kian, this is nothing special about YT in this sense, bussines is harsh by definition, and more ore less every company dealing with money and targeted to profit will behave
in less or more in similar way, until it is reach enough with cash and experience to take care about public relations and future. YT was rather example close and familiar to our small community.
Whe Sinclair reaches the top of home computing with ZX Spectrum machines, some bad business decisions make the empire fall. Remember that Sinclair invented the windows display system with the Sinclair QL computer in 1983
I was a Sinclair User, and I was a BeOS user, but my little Speccy gets me more fun hours than the Internet and my BeOS runs everyday to remember me that somtimes it was fun…
… but Be realistic, I need AutoCAD to make my work of everyday, need some other apps that BeOS can’t offer by now. I can’t connect my Palm !.
Well, I have a frien that is working by now with the Spectrum Stylesheet to control its home finances and saves it to the tape, yes, a cassette…
Well, I think this is going so far…
Bye.
^..^
sabreman
An OS is only Dead when the last user cannot use it for their everyday use. As you’ve just said, you know someone who’s use a Sinclair for everyday use at home. Then that platform is not “dead”.
However, anyone who calls BeOS, RiscOS or AmigaOS “Dead” should look at the large number of people still using them as their everyday OS, both for work and home. Although much of the RiscOS users are under emulation (VirtualAcorn) these days.
Kevin, did the _CTO_ really tell you that? You’re not confusing the CTO with the CVO – Bernd?
> … but Be realistic, I need AutoCAD to make my work of everyday, need some
> other apps that BeOS can’t offer by now. I can’t connect my Palm !
I need to use openssh everyday, there’s no (good) openssh for Windows, ergo Windows is a dead OS. I only use it for amusement now and then when playing old games.
Sorry, the *CVO* slandered me. Not the CTO. Thanks to tic for reminding me what weird title Bernd was using today.
Chief Visionary Officer… hrmm….
Actually, I just read an article about them different titles; expect more of that coming… The CTO is actually very friendly. And well, I think the CVO is, too, although I haven’t talked to him as much.
It’s overpriced.
I think it ought to be priced at no more than $39.95US.
It’s just not worth over $100.00
YT might be surprised at a simple business practice, Pricing your product to sell….
At $39.95 (even $29.95 for a downloadable edition), people who would otherwise not touch it for over $100.00US, might try it.
I certainly would.
If YT is hurting for cash… Lowering the price to increase sales would be smart.
Especially since Zeta is Beta product… And its competitors in the same space are free (BeOS PE, and PhOS).
I don’t mind paying for an OS. But I DO mind OVER paying for it.
Whe Sinclair reaches the top of home computing with ZX Spectrum machines, some bad business decisions make the
empire fall. Remember that Sinclair invented the windows display system with the Sinclair QL computer in 198
wait, what? windows display system? i thought windowing had been around for quite a awhile before the ql.
Dear Michael and others,
It might way past time to remind you (and others that would rather prefer GPL licensing to MIT in Haiku) that *absolutely no standard or reference implementation was, is or is to be coded under the GPL license*. It’s illogical. A standard is there to be used as a guide for others to implement, improve upon, create a product, package it, sell, whatever. GPLing a standard defeats the very purpose of calling it a standard.
Haiku is BeUnited’s standard reference implementation for a Open Standards BeOS Compatible Operating System. Not by design, but things happened to be this way. Were Haiku GPLd, it just couldn’t be considered a standard.
Now, for those wanting a GPLd Haiku-ish OS, praising how many more developers it would attract, we could view this in two different ways, and this addresses another big complaint I’ve read in this board: binary compatibility. There are no GPLd projects aiming the recreation of BeOS in a binary-compatible way. Period.
So, let’s break binary compat and “keep the BeOS spirit”. How about AtheOS? It geneted a flurry of optimistic views in the past, it was touted as the Next Big Thing, but now the excitement went down to zero. And anyway the project was pretty much dead the last time I checked it. Most newcoming, GPL-attracted devs moved on to Syllable, and I can concede this: it did attract some devs. However, considering AtheOS had been in development for several years by then, and considering that Syllable has been developed by GPL-adhering, innovation-hungry people since the fork, *it’s absolutely unimpressive so far*. That’s the general consensus from what I’ve seen in the comments on the LiveCD4. So what if it boots, has a GUI etc? AtheOS did that. I somethimes fail to see the point of this project as it fails to bring the atmosphere, the “oohs”, the nostalgia of the sweet memories and the hope for a new, true BeOS? I’d rather run Workbench on UAE.
However, I just responded this question: “I’d rather run Workbench on UAE”. So it’s a matter of passion. They do it because they enjoy doing it and they don’t give a rat’s feck to what I think. And I’m happy they’re enjoying it; they’re at least doing something *useful and fun* instead of sitting in the peanut gallery and whinning about licenses or binary compatibility.
So, now this leaves us with BlueEyedOS. It’s got X11 as the GUI engine, and Linux as the kernel.
Last posting on their site: May 2nd, 2004. Saying that OT did compile and link. Wait a minute, so it was linked against *stub*, empty functions? Even I could do that! Where’s the progress in there? From the best of my knowledge, they actually did some original work, then used some Haiku code and now they’re trying to tie everything together. And the previous message states that they’re having trouble to open a SF/Savannah account to upload the project and have it open for the general public. For Heaven’s sake, how many projects have been opened on these sites since last February? How about the message before that one? It dates from last December. One news posting every 3/4 months.
I’m glad that *even* when they’re dealing with the site transition and the non-profit paperwork, the Haiku team is seldom silent for that long.
Now, Cosmoe. It ditched X11 but it’s got Linux. Guess what? They leverage on the work of Haiku, Cosmoe and Syllable.
Unless we’re running in circles here, there’s absolutely no progress bein made.
Wait, we are running in circles. I explicitely conducted this message to get to this point: Haiku is where the real progress is being made, even if it’s behind the curtain. The man behind OpenTracker, Axel Dörfler, is the leading Haiku developer. Actually, if you read the OT changelogs, you’d be surprised by how many of them are also part of Haiku. The people behind Mail Daemon Replacement are also in the team. François Revol (mmu_man) did some substantial contributions. Marcus Overhagen, who developed the ICH AC’97 and two Gigabit Ethernet drivers, is there as well. Caz, who ported an obscene amount of emulators and SDL games to BeOS, is there as well. Looncraz was but a script kiddie 2 years ago (Hi Loon!
), and now he’s quickly maturing as a developer. And he’s coding with Haiku in mind.
Bottom line is:
a) The other opensource projects aiming to recreate the BeOS experience are either experimenting much-less-than-blazing development speed, irregardless of adopting the GPL, X11 or the Linux kernel, and despite the fact that most of them already come from a WORKING, SOLID, GUI-BOOTING base (AtheOS);
b) The most prominent BeOS developers bar the former Be engineers are already at the heart of Haiku.
Conclusion: Please, do yourselves a favour and step out of the peanut gallery. The whinning is happening because you’re not seeing progress being touted left and right, because the developers aer *busy coding*. However, discussion does happen, all the time, and the devs do get involved. So I suggest you people to join the mailing lists, watch the progress happening in realtime as I do, and, if possible, contribute something (even volunteering to update the site!) instead of whinning out of misinformation.
**END OF RANT**
The CVO is the infamous Bernd. He can be nice, but he has a very short temper and often serious issues are caused by him misinterpreting something you say in English. He has a great habit of putting his foot in his mouth at times. And doing things like slandering BeOS users for being involved with BeBits or the BeOS Journal.
is openbeos at year two or three? you might reserve judgement regarding the lack of developers that haiku has until you compare it to the total number of developers that linux had in year two or three.
I have a feeling linux was not doing so much better at year 2 or 3.
Kian Duffy wrote:
>I just want and hope Haiku to stay under a real free licence – BSD or MIT.
Well, MIT is a very free license,yes, but here we were discussing about *number* of developers…that prefer one license instead of another. Yeah, if I can have Linux developed under MIT it could be even better, but actually it isn’ t so (it don’t exist).
Let’s call it fashion, GPL or what you like, but the fact is that Linux
has grown quite well and with many supporters.
To do an example on the other side, OpenBSD uses BSD, but you can’ t just copy their CD images,
their team works hard, it has some goals (i.e. security),
and with this CD-IMAGES® escamotage, they survive.
So there are case and case, where to choose license, and how to boost a project.
With GPL, it seems there are more developers…that’s all.
Ah…and in my first post I wrote about many Haiku developers that hate GPL…that’s why I wrote “Les jeux sont faits”.
Yes, I know that if openBeOS project started as GPL, many of them they didn’ t help it, it’s their choice, right 🙂
GPL imho is unconfortable, but with LPGL…
Ok, “les jeux sont faits” 😉
Its just entered the third year. Just.
Be may be dead, but we’re alive. Haiku is making progress. You’re all spoiled by the current mode of “release early / relase often”.
Yes, it’s tempting to cut corners and use Linux 2.6. In many ways it’s an awesome kernel. But, it’s probably close to impossible making a application platform (or “OS”) based on Linux without it ending up a mere Linux distribution with the usual Linux apps. And there are aspects of Linux which we, the BeOS community, do not enjoy. Technical ones. Licensing ones.
Anyway, here’s my vote of confidence in the Haiku project, and here’s hoping I can find more time to contribute in the coming months.
Thank you.
meianoite, thanks for the detailed post and explanations. That’s exactly what I was hoping to see — proper discussion on licensing, and the pros-and-cons of adopting the GPL. (Some people got the impression I was dogmatically saying “GPL good, all else bad, Haiku should use it or it sucks” etc. Not at all; I just think it’s something worth considering.)
“*absolutely no standard or reference implementation was, is or is to be coded under the GPL license*. It’s illogical. A standard is there to be used as a guide for others to implement, improve upon, create a product, package it, sell, whatever.”
OK, but then you can end up with multiple standards, which isn’t good for anybody. If Haiku takes off, you might have five companies all making their own variants, adding closed features and modifications — and they won’t want to release them, in order to maintain their advantage. So software developers would have to target multiple Haiku-flavours and it gets messy. I understand what you’re saying, but it may not work out so well in the long-run…
If Haiku was GPLed, modifications to the core couldn’t be kept under wraps; in this way, there’s less chance of serious fragmentation and incompatibilities emerging.
“However, considering AtheOS had been in development for several years by then, and considering that Syllable has been developed by GPL-adhering, innovation-hungry people since the fork, *it’s absolutely unimpressive so far*. That’s the general consensus from what I’ve seen in the comments on the LiveCD4. So what if it boots, has a GUI etc? AtheOS did that.”
Yes, AtheOS is dead now, and Syllable is carrying on where it left off. Bear in mind that AtheOS was already feature-complete in many respects, in that it had a full GUI, network stack, development toolchain etc. — so there’s hardly anything revolutionary for the Syllable developers to add.
Besides, a huge amount has happened in Syllable, but most of it has been under-the-hood (drivers etc) rather than snazzy end-user features that really stand out. Due in the next release (0.5.4) is a brand-new desktop and file manager, an incredible project by Arno Klenke, which gives Syllable a massive visual overhaul. So development is progressing well (not as rapid as everyone would like, but decent nonetheless), with over 500 mailing lists posts this month and all sorts of things in the pipeline:
http://msa.section.me.uk/sdn/
(by meianoite)
http://www.osnews.com/comment.php?news_id=8114&limit=no#274430
What happened March 2004?
http://haiku-os.org/statcvs/loc.html
OK.
Perhaps we must say Dead, as Kian Duffy said, when last user deads…
So we have to search for another definition of dead. How about : hibernate ?
My Spectrum is hibernating until I play some tapes (that is some times a week, so isn’t dead and isn’t hibernated)
My BeOS is really hibernating, so I used it twice a month.
My Windows isn’t hibernating because I used everyday.
My QNX is hibernating since first week of use cause my USB modem isn’t working
So hibernate could be a better word than dead, but it’s like Woody Allen said: it’s like Anything else.
> OK, but then you can end up with multiple standards, which isn’t good for anybody.
That’s a bit of a leap. I cant think of any current standard (ie ANSI, IETF IEEE etc) that is under the GPL. Of course, we could be talking adhoc and informal standards but in the BeOS world there’s something kind of like IETF, BeUnited. Come to think of it, BU is more like OpenGroup but anyway….
BU is the organization that sets the “standards” in this case so if a company goes incompatible, well, hey, you’re officially no longer standards compliant.
Also note that it is quite possible to make an incompatible Linux distro without violating any license. Since Linux “allows” non-GPL’d kernel modules (the nivida gfx drivers come to mind) to be loaded it is possible to replace most/all modules with your own closed ones. Voila, a very incompatible Linux without violating any license.
> I understand what you’re saying, but it may not work out so well in the long-run…
Worked out pretty well for the *BSD’s and most other non-GPL projects.
> If Haiku was GPLed, modifications to the core couldn’t be kept under wraps;
> in this way, there’s less chance of serious fragmentation and
> incompatibilities emerging.
To continue the above discusson, this would still be possible since the BeOS kernel is highly modularized so anyone wanting to be incompatible could just write their own app_server or whatever module and be incompatible without breaking any license.
> Besides, a huge amount has happened in Syllable, but most of it has been
> under-the-hood (drivers etc) rather than snazzy end-user features that really
> stand out
That’s pretty much like Haiku then, lots of changes under-the-hood.
SD wrote:
>See story with Decaf and Firefox credits, at zetanews archives or at bezilla blog. And this is quite common practice fro them, it seems.
It’s always nice to see your own hard work (with lots of tweaking and bugfixing) demoed on german TV as if it was YT’s own. I wish they would have contacted me before. Afterwards I posted an article asking what all that was about, and the CEO called me a ‘Zeta hater’ for posting that (which I am not). Although I must admit I’m not a big fan of CEO’s that resort to namecalling.
The GNU coreutiles/sharutils/etc were checked it. Most of the missing CLI apps
meianoite, thanks for the detailed post and explanations. That’s exactly what I was hoping to see — proper discussion on licensing, and the pros-and-cons of adopting the GPL. (Some people got the impression I was dogmatically saying “GPL good, all else bad, Haiku should use it or it sucks” etc. Not at all; I just think it’s something worth considering.)
Well, I was not addressing YOUR specific concerns about the GPL; anyway, I believe you’re the only one who still cares about this discussion. I guess I’d just have to repost that message when this discussion resurfaces again.
OK, but then you can end up with multiple standards, which isn’t good for anybody.
That’s a risk that might happen. However, it hasn’t happened with X11, and when it happened to BSD, it was “for the good”, as the big innovations always flow from one project to the other. The BSDs are more like siblings with different personalities, but in the end they just happen to share mostly the same DNA. So in a way the project can ger fragmented, but in the other hand History has proven that this could be a good thing.
Which is in flagrant contrast against the Linux distro fragmentation, in which they rarely truly contribute anything to the other distros. You don’t run YAST on Gentoo. You don’t get Portage on Fedora. When such things happen (like linuxconf being used in some non-RedHat distro), it’s a community effort which would be better spent improving something else. IMHO.
If Haiku takes off, you might have five companies all making their own variants, adding closed features and modifications — and they won’t want to release them, in order to maintain their advantage. So software developers would have to target multiple Haiku-flavours and it gets messy. I understand what you’re saying, but it may not work out so well in the long-run…
You still fail to see the point that Haiku is NOT a distribution; it’s more like a body of standards. It’s no surprise that they are not booting a GUI: it’s not the point to do this just to show some VESA graphics, a mouse pointer, a blue background, and some ugly-looking windows. A GUI will be ready when it’s ready to be pronounced The Gold Standard. This is the same in regards to the other kits: some are production quality (MDR was recently donated to the CVS tree and is now considered an integral part of Haiku; then there are the Matrox/Nvidia/Neomagic/Radeon/AC97/… drivers, all of them under the Haiku umbrella as well) and usable right now, some are beta quality, some are alpha. None stuck in the “planning” phase.
As I said, those kits are not being released (as in announced, touted, made easily and readily available) to the public for the sole reason of not being production quality by the (extemely high) standards of the BeOS community. Those curious enough can download and test the kits; there are tips on how to get them on the Haiku site itself. What they won’t do is endorse using those kits, apps and drivers that are not production quality, because that would backfire in an extremely painful manner. I hope you agree with that point.
You see, there’s no distro ready, that’s obvious. However progress is made in several parallel camps. They’re not completing each task linearly, one after another, but everything is being tackled in parallel. So, as NathanW put it, I guess people will just be surprised when Haiku “sprouts from the ground”. Actually, it won’t come out of the blue, because great progress has been done though all these years. It will just seem that it came from nowhere to become something big becaust it won’t be released in a distro form (and notice: BY SOMEONE ELSE. Haiku is not a distro company, they only provide the source and that’s all. Sort of how CrescentAnchor is distributing DragonFly BSD) until it’s thoroughtly feature complete.
Also, you’d better read some mailing list archives instead of relying on the progress meters on the site. I’d be *very* sad with their “progress” if I relied on it. It gives false negative impressions. Trust me.
If Haiku was GPLed, modifications to the core couldn’t be kept under wraps; in this way, there’s less chance of serious fragmentation and incompatibilities emerging.
I seriously believe that this won’t happen. First, because BeOS as-is still runs on a variety of hardware, old and new, and some won’t bother to “upgrade” if they can just run BeOS on the legacy hardware and Haiku on modern hardware. Apps are still being developed, ports are being made, and people just want to run their software on the platform they like most. Take MorphOS as an example: it runs on modern hardware (Pegasos boards) and it retains compatibility with most of the older 68k Amiga apps, and Amiga is still run on legacy 68k machines to this day. Breaking compatibility in any form will backfire at those who break it. People have expressed their concern about Longhorn breaking compatibility and how that would backfire at Microsoft; how come in the case of Haiku people just become doomsayers and say it’s an inevitability? People are not stupid and the BeOS community is at least as united as the Amiga community; they don’t want and won’t let incompatibilities happen.
The second reason is that not everybody wants to trash their old hardware for one reason or another. I still have my NES. I still have a Performa 6360. Same goes to my Pentium 133. The latter 2 are used *daily*, perhaps more than the newer Athlon XP on NForce2 mobo. When one gets comfortable with their setups, there’s really no reason to change unless absolutely necessary (like a hardware failure). Becoming incompatible is the worst idea for those kind of people (which are the majority of the users of any OS that got past the status of “curiosity”, “toy”).
Yes, AtheOS is dead now, and Syllable is carrying on where it left off. Bear in mind that AtheOS was already feature-complete in many respects, in that it had a full GUI, network stack, development toolchain etc. — so there’s hardly anything revolutionary for the Syllable developers to add.
My point exactly. So I hardly understand why people tout Syllable’s qualities and fast pace of development when the hard work was already done, and ditch the efforts of the Haiku team when they came from ground zero (except for the kernel; that one was perhaps 15-20% ready for its intended use in Haiku)
[continues…]
Besides, a huge amount has happened in Syllable, but most of it has been under-the-hood (drivers etc) rather than snazzy end-user features that really stand out. Due in the next release (0.5.4) is a brand-new desktop and file manager, an incredible project by Arno Klenke, which gives Syllable a massive visual overhaul. So development is progressing well (not as rapid as everyone would like, but decent nonetheless), with over 500 mailing lists posts this month and all sorts of things in the pipeline:
http://msa.section.me.uk/sdn/
That’s absolutely great, and I congratulate the whole Syllable team. I have nothing against them and I wish them the best of success; I’m just a little ticked off when people start comparing stuff that can’t really be compared, like saying “look where Syllable stands now, and it’s being only 2 years in development!”, or how “AtheOS and SkyOS are mostly one-man projects, and they are MILES ahead of Haiku!”, when both are false, misleading statements. They are both hobbysts OSs (and I don’t mean that in a prejudiced sense) which are nowhere near production quality, and that’s the whole reason why people join their teams: to take them to the next level. I mean, seriously, how many features are really complete on those OSs? It’s not difficult to show some progress when you expect nothing more than the “works for me” and “fulfills my needs” statuses; hence the pace of development when only a single pair of hands is driving the project. When people start demanding more and more hands join the project, the perceived speed of development usually decreases, because more is done under the hood (as you just said). Then again, as foundation is better laid, the speed picks up again (hence the new flurry of excitement over the 2.6 series of the Linux kernel). I believe that, by now, you understand that
1) Haiku must already fulfill a whole community’s need, and aim for more. Their standards are much higher than those of hobby OSs. In recent times this became valid to Syllable and SkyOS as well. I see no miracles in the development of those projects, neither in their former fast pace of development, nor in their current mature, well-thought-out-albeit-slower-than-before progress;
2) Because of that, it will be *way more impressive* than people expect, because it’ll be feature complete *and* as mature as serious users would expect. It’s only natural that it takes longer to achieve these goals.
And please, keep in mind that Haiku R1 is only the beginning: it will work like R5, bar some really bad bugs. It will be faster (OBFS already is, and the app_server and kernel are going to that direction too: just check the archives of the newly-created haiku-optimization list. They already got speed improvements of 20% on areas that BeOS was considered to reign supreme).
Those who follow their progress closely just can’t lose their faith, because progress is definitely visible and almost touchable. Join the lists, you won’t regret it.
I’m completely, thoroughly, entirely flattered.
Thanks, Jonas ^_^
I guess it was when some of the GNU CoreUtils code was updated to newer versions, and when the MDR were submitted to the CVS.
No voodoo
uh. sorry about that comment going all bold from some point on.
(the preview button doesn’t show the markup effects =P)
The MDR isn’t in CVS yet. It was entirely the GNU stuff. Another massive jump will happen when MDR goes in, and then eventually when OpenTracker goes in there’ll be a third huge jump.
Sorry for some spelling mistakes, redundancies and bad grammar as well ^_^
Not a native speaker, bear with me.
The MDR isn’t in CVS yet. It was entirely the GNU stuff. Another massive jump will happen when MDR goes in, and then eventually when OpenTracker goes in there’ll be a third huge jump.
Oops. My mistake. I guessed it had already happened as Nathan has warned the subscribers to the CVS list that it would happen soon, and that was maybe 2-3 weeks ago.
So it was GNU stuff. “Nothing to see here, boys; move along”. ;D
Be Inc.: It looked into the eye of the ‘redmond’ T-Rex and ‘focus shi#tted”.
BeOS: Left for dead but still shows eye movement and nerve twitching.
haiku is not beos. Haiku is a new open source project which will emulate SOME of the features of BeOS. Its sounds as though the team is already modernizing elements of the OS and trying to correct for mistakes made in beos.
the arguments here suggest that many of you perceive haiku as beos. It is not. It is really just a new OSS project which is luck enough to be able to draw fans of beos.
that gives haiku a leg up on a lot of other open source projects. Haiku will also evolve and it may still be unique in its product offerings when it is available. After all bloat is the route for the other operating systems.
Haiku has received very very little publicity which is ultimately why it remains a relatively small group. That might change. Linux was once small as well. You might remember that there were once very few computer users and no internet users. What is today may not be tomorrow. nothing is static other than the chronic negativity of a few who continue to criticize this worthwhile effort.
So give these guys a break and quit the negativity. They are doing something cool and i applaud that. What do you do with your spare time? They code. We criticize. Which is better?
Market success may or may not appear but they do have the luxury of time since they are open source.
Notably, the audio/visual world still has no optimized OS for intel hardware and that is a market which i don’t expect to see filled anytime soon.
Ok, folks!
Haiku STILL HAVE A BRIGHT FUTURE.
Haiku will be finished in two years.
Haiku isn’t a market product to think in a timeline.
Go Haiku!!!
Michael Vinícius de Oliveira
~ BlueEyedOS.com Webmaster ~
Wow, a freakin’ black square. Am I supposed to be impressed?
I’ll be a lot more impressed when I see a Haiku live CD, or even a screenshot for that matter. Until then, Haiku is just the biggest piece of vaporware since Duke Nukem Forever.
Er, while the attention of so many BeOS people seems to fixed on this thread, can anybody tell me what “Bone” is? The only thing I know about it is that it’s apparently illegal.
Replacement kernel-level BSD compliant networking stack. Improves stabiliy and network speed for some; improves POSIX compliance for all. But its leaked and illegal.
300 posts, must be an OSNews record
Vapourware, as you put it needs for there to be no code to show of any development progress, just verbal statements that the project is alive. My understanding is that Haiku has more than hot air supporting its progress. The fact you seem to purely ignore. There are many parts of BeOS that can be replaced by working code from the Haiku project. If that canstitures “Vapourware” then I seem to exhibit ignorance as to what “Vapourware” means.
Get a life!
p.s. The waiting is worth it and already there are parts of Haiku making its way into PhOS.