Submitted by This summer, too, the Haikuproject is part of the Google Summer of Code event. One of the more interesting projects is the Services Kit (draft document!) by Christophe “Shusui” Huriaux, which is an API to facilitate the creation of native web-enabled programs using standard web protocols and data exchange mechanisms.
At Haikuzone, Jorge G. Mare gives a succinct but thorough explanation of what the Serives Kit really is. “The Services Kit is comprised of three layers: a network layer for sending and receiving data using standard web protocols such as HTTP, HTTPS, FTP or LDAP; a data interchange layer with support for standard specifications such as XML, JSON or REST; and the API layer that provides the interface for developers to use,” Mare writes.
Huriaux explains what the Services Kit does in its most basic form. “In its simplest shape, the kit takes an URL, decide what protocol he should instantiate, make the request to retrieve a resource, ensure that the request is well completed and return an object containing all the resulting data,” he details.
The goal of the Services Kit, therefore, is to make it easier to develop web-enabled applications in Haiku using a standardised API. As such, it made sense to test the Services Kit with WebPositive, the WebKit browser for Haiku. After a lot of work, a WebPositive build using the Services Kit instead of cURL was built, and it offered compelling performance improvements.
“A little test of a huge file download over my local network led me to this conclusion: Web+ with cURL was downloading at ~250kB/s whereas Web+ with Services Kit was around 1.5MB/s,” Huriaux writes, “The core of the Kit is not yet optimized with deferred loading and other features like that, but we could expect much more good things!”
This seems like a great new addition to the Haiku landscape, and I’m sure many developers of web-enabled Haiku applications are eager to start implementing Services Kit. For now, the Kit is not yet done, but Huriax has already made it clear he’s going to continue development even after GSoC.
oh this is gonna be soo cool!
Thats nice, Haiku probably have a lot better chances of any adoption this way. Is its browser based on Webkit, I thought so? If so, can it be integrated smoothly with the Google Appstore? (or what about porting the Chromium browser even?).
Getting enough traction with native apps could be a struggle in the desktop space (even OpenSolaris could not really pull this off IMHO, standing on the shoulders of Linux efforts for some of its apps) so this is nice.
Edited 2010-08-20 22:26 UTC
I agree with you, check out this project ( http://revlin.org ) they are using Haiku as the core of the OS
Since Thom Holwerda issued the challenge to Haiku developers some years ago, they have been moving forward at a good pace. The project seems to be gaining momentum as it progresses. Of all the “alternative” OSes, I believe the Haiku developers demonstrate more “outside of the box” thinking. They also use GPL code wherever necessary, thereby eliminating the need to re-invent the wheel.
I did what now?
You did it with pink unicorn power!
In all seriousness, I think what he is referring to is your frequent vocal support in the early years of the project and your well-stated love for the BeOS over the years. Like it or not you are a somewhat influential person on these here interwebs, at least when it comes to operating systems.
Then again, he could be referring to a specific sentence you spoke three or four years ago and we all forgot about it, you included.
I’m going with unicorns.
That’s not entirely accurate, that license is pretty much a last resort for us. Generally most code in Haiku tries to work from sources with friendlier licenses such as BSD/MIT or LGPL. The only major GPLed piece is gcc, most everything else is disabled in the default build (you have to explicitly configure your build to include GPL-based code).
Indeed. And several GPL pieces have even been replaced in favor of BSD-licensed code (such as several network drivers, etc.)
It will be nice to release Haiku from glibc eventually as well
Not that I might work on it, because my aptitude in this subject is still limited, but what about calling it hlibc?
I was under the impression that Haiku was still using the FreeBSD libc except that there were a couple of functions that were changed so that they operate like the glibc implementation rather than the FreeBSD implementation.
I believe they started off with the BSD libc, but switched to glibc for compatibility reasons (since BeOS had used glibc for its “libroot” support).
http://dev.haiku-os.org/browser/haiku/trunk/src/system/libroot/posi…
Recent discussion suggests that glibc may still be retained only for the BeOS compatibility (i.e. a gcc2-compiled set of libs will still remain optionally-installed for BeOS compatibility) while a non-GPL libc replacement will be grafted in moving forward.
Sounds awesome; I have to admit though, it is rather disappointing that they’ve removed the progress chart which used to track the ‘completeness’ of each API call or feature; is there some way to track how far along they are in terms of implementation? do they have a goal for the final R1 release or is it more a situation of plodding along and focusing on the quality rather than an arbitrary time table?
I’d say it simply became too hard to track.
At this point, it’s not so much “what’s done, and what isn’t”, but rather: “What apps run, and for those that don’t, do we really care?”
I’m not sure anyone really has a clear idea of what is incomplete these days. I know the media kit is still missing encoding support, but beyond that, the parts that are lacking BeOS R5 support are either few, or unimportant I guess.
These days, people are more interested in when Haiku will support their specific hardware, or support WiFi, etc.
As for when R1 might emerge… I couldn’t answer that
There seems to be a desire for rock-solid quality before the core devs will sign off on that.
Edited 2010-08-22 05:52 UTC
The Haiku media kit already has encoding support. It was implemented last year by Stephan “Stippi” Assmus.
The development progress charts were cool and but pretty useless. The idea was nice, but the team leaders did not have the discipline to keep the charts up to date, so in the end they became nothing but make believe business, as in most cases they did not reflect reality at all. Good riddance.
I suppose that is why Clockwerk can work
Clockwerk is one; the other is MediaConverter, which also uses the Media Kit encoders to convert media files.
glibc isn’t under GPL.
You’re right, it’s only LGPL…
Ultimately, there is very little GPL code used in Haiku, unless you count all the CLI tools, which are pretty much used only for development and *nix compatibility.
Right its LGPL and closely tied to Linux/*nix systems…
other possibly more flexible alternatives should also be considered … for the c++ libc++ might be usable and its designed to be flexible and compiler independant
If the Haiku project aims for GPL-free development environment, Clang is the only potential alternative and that comes with its own C++ library.
As for a LibC….. I’d guess they’ll adopt the one from FreeBSD or NetBSD at some point to get a BSDLed implementation.
I don’t see why they would aim for a ‘GPL-free’ development environment, the licence makes no difference in this case. For system bound code, naturally they need to avoid GPL. Whichever compiler they choose to ship Haiku with should be decided on a purely technical level in my opinion, which I’m sure the Haiku devs are very qualified to do. However, given that llvm has adopted the compiler flags from gcc it should be no problem for the user to work with whichever compiler he prefers (or different compilers for different projects) once llvm has matured sufficiently. From what I’ve gathered when skimming the mailing lists, llvm compiles Haiku (after some fixes) so the time when you can build haiku from top to bottom with both gcc and llvm might be very near.
Yes, gcc and bash afaik, and given that both are standalone they have no impact on the overall licence strategy.
Does anyone know of their plans for LLVM integration?
LLVM and Clang support next then
? Soon enough C++ won’t be a problem with Clang at all 
.
The Haiku dev teams didn’t have an efficient PR years ago. It had brought many incidents of misunderstanding against Haiku. But now it’s not the case anymore.
As long as the spirit of BeOS lives on through Haiku, I wouldn’t complain.
Take a look at the draft proposal:
https://docs.google.com/View?id=dfc49mzx_7gq8fbgfw
Here’s how a pro would explain the same thing:
“We are going to implement wget/cUrl/urllib/QNetworkAccessManager-like thingie for Haiku. We might throw in some data extractors (using JSON/XML parsers) if we have the time, though we are not sure whether it really belongs to this layer.”
I believe it’s the same thing than Google Native client right ?
If so, why they didn’t use it ? (Especially since BGA work at Google for it). Anyway it is really interesting
http://www.freelists.org/post/haiku-development/Googles-NativeClien…
Thanks
Edited 2010-08-21 06:33 UTC
See my previous comment for what it is. Completely different things.
Recently used Haiku a couple of hours a day for a week, and BOY, does it suck.
I have never used BeOS, so I don’t have the nostagia thing that probably makes it interesting for some people.
Compared to modern OS’s, it seems awefully dated. It actually caused severe pain in my “mouse arm” because of all the insane menu navigation.
It generally feels old, awkward and I got a lot of freezes when putting it under pressure (disabling SMP helped somewhat.)
So I have to ask… is there a point to this project beyond the fun of doing it / nostalgia?
I am pretty sure Haiku’s UI will go through an evolution. The idea of its current UI is pretty much based on the look of the old BeOS 5.2 but they’ve done some overall refinements. However I do prefer if the tracker (taskbar) behaved a bit like the Windows one when stretched to the bottom and this means left clicking on a button activates the window and maximizing the window does not cover the tracker/taskbar.
Actually it is an alpha stage software and it reached its Alpha2 milestone, meaning that the OS is still on *unstable development stage* and putting it under heavy pressure is unfair giving the fact that it is not production ready and everyone knows that it is in heavy need of bug reports. Besides all this, I think that even its in alpha stage it still far better than other FOSS or proprietary software.
The point of this project it to provide a cleanly implemented bloat-free responsive modern OS that permits everyone to concentrate on *working* not on tweaking or cleaning the OS 8h/day. In addition on goal of the project was to implement a complete coherent OS stack from kernel to font preference dialog with focus on threading to allow maximum responsiveness. All of this was thought from the beginning not after seeing all the success and hype of threaded tabs on Chrome browser.
Being someone who has seen BeOS in it’s hayday and used it in relation to offerings from BSD, Linux, OS-X and Windows, BeOS as a an OS that focuses on User responsiveness and media handling – nothing has touched it.
I would take BeOS on a AMD K6 with 256Mb ram for running multiple video and audio streams without dropping the ball over anything that exists today.
Haiku – now it’s in Alpha 2 Stage and they have to iron out the code after rebuilding from the ground up and also adapt it to modern hardware. Couple that with tracker the OS will wipe the floor of the current desktop offerings. Kudos to the dedication of all involved in the Haiku project.
Yes the UI needs to evolve and it will but they are focusing on getting the thing out the door with 5.3 binary compatibility and then move it forward with subsequent releases.
What exists on the desktop today is fn depressing and holding back the state of desktop computing by a decade or two.
I find mac osx and win7 pretty darn pleasant to work with. The Haiku gui feels more like some sort of intellectual masturbation thing than anything else.
Try loading both modern OS’s up with 6 avi streams and 10 MP3 streams and then get back to me about how wonderful the User Environment is.
Even on modern accelerated hardware bot OS’s will crawl. BeOS was doing this on 10 year old hardware with no audio, frame drops or non responsive UI without the power of computing we have to day.
Chalk and Cheese.
I did just that in addition to having a few browser windows open, and I’m not noticing any issues at all. CPU use is hovering around 10% and i can’t even find a single stutter to complain about.
Now I’ll admit that Haiku could probably do this on slower hardware than what I’m running, but I don’t think a Q6600 CPU, 4GB RAM, and a Radeon 3870 GPU are that much more powerful than what the typical user runs today.
I did just that in addition to having a few browser windows open, and I’m not noticing any issues at all. CPU use is hovering around 10% and i can’t even find a single stutter to complain about.
I just tried it myself but I didn’t bother with mp3s. I just opened 10 AVI files instead, resulting in about 17% CPU usage. That’s not bad IMHO; 10 simultaneous videos playing and plenty of room for more. Haiku ain’t the only OS that can do that and the people promoting Haiku should rather find something else to use as metrics instead of “I can play so and so many videos at the same time!” Playing several video streams on even a moderately recent hardware just is no feat at all.
BeOS did that with a few MHz, <256MB of RAM and software rendering. I guess all that glossy stuff doesn’t help if you need a powerful computer to do what was possible 10 years ago.
While its cool, video and audio editing on BeOS is joke. There are no applications that are good enough for professional work. So even it it could run 40 streams without hickups its just a nice demo to impress your roommates.
There really aren’t many outside of Windows and Mac. Of all of the apps on Linux for video editing, only Cinelerra, and Kdenlive even come close, and it’s still a huge step down from even Premiere elements.
VLMC looks very promising, but development is moving at a pretty slow pace.
I think once there are stable, pro-quality video editing apps for Linux, it will be trivial to have them on Haiku as well.
On the other hand, audio is looking better, since some of the major audio hardware companies still use BeOS in their hardware, and will probably use Haiku when there is a stable release. Maybe there will be a version of Ardour, Audacity or Traverso when there is a 1.0 release… but until then, I doubt you’ll see many popular apps for Haiku.
http://www.tascam.com/products/sx-1.html
http://www.roland.com/
http://www.izcorp.com/
That was a cool BeOS ability 10 years ago, not so much today. Any modern OS running on moderate hardware is up to that task these days.
BeOS was cool and Haiku is cool but running 10 mp3’s and 6 videos won’t be the selling point today.
True, but it depends on wether that is because the hardware is so much better nowadays that every OS can play lots of videos simultaneously or if the OSes themselves have catched up with Beos high level of responsiveness. When I do computionally heavy things like rendering/compression/encoding on Windows and Linux the system can become slow and slightly unresponsive (particularly on windows), this never happened back on Beos so it will be interesting to try maxing out Haiku once it matures a bit more to see if this still holds true. Hell, even copying a large file slows down user interaction in windows/linux from time to time which is probably fine for a server os, but for a desktop os I rather have the file copying 5-10 seconds slower and have a fully interactive system while it copies. In my opinion responsiveness should always trump throughput in a desktop environment.
For the average user it does not matter. Heck, playing 10 mp3’s and 6 avi’s is entirely pointless anyway. It’s the kind of thing you use to show off to your mates before one of them go “sure, that’s cool but what can it DO?”.
BFS had problems with files too, you know. I remember having a couple of thousand emails in a folder and let me tell you, doing any operation on those files was the slowest filesystem performance I have ever seen.
True in it’s own – multi videos and audio tracks are useless except for wow factor BUT, considering the hardware and true plug and play in it’s day which was the envy of MS, BeOS was a potentially devastating content creation platform.
Unfortunately that was lost by Be Inc loosing focus and changing direction with their hardware ideas.
Tried that. Works fine on my aging KDE/Linux-powered laptop. The only bottleneck is my slow hard drive and that bottleneck has nothing to do with the used OS but with the cache setting.
BeOS was a great OS back then but other operating systems catched up and exceeded meanwhile.
THE HARDWARE caught up. Even today, I see Windows 7 locking up and doing its own stuff instead of responding to me. That’s not evolution, that’s stagnation.
Maybe it’s because in Win7 you can actually do stuff. In BeOS/Haiku, not so much, apart from watching a teapot spinning. Kinda reminds me of the whole Compiz spinning cube hype. The more things change…
What do you define as ‘doing stuff’ that Haiku can’t?
I can:
* Browse the web with a modern webkit browser
* Connect to wifi
* Play console games via emulators.
* Edit images with the Pixel Image editor
* Edit videos with Clockwerk
* Download torrents with transmission
* Program in c/c++/python/perl
* Chat on Jabber
* Burn CD’s
* Install numerous apps from BeBits,HaikuFire,or Haikuware
…Just to name a few tasks… Not bad for a Alpha os
I haven’t used their new web browser, but its lacking plug-ins and pixel32 wasn’t really usable last time I tried it on BeOS.
The one thing I really liked with BeOS/Haiki was/is the TranslateKit. Other OSes should really look into that feature and integrate it by default.
To stay on Topic I think this new Kit will be a good thing. I haven’t looked at the API but its good to access networking layer etc in C++ so you don’t have to mix C with C++ for simpler tasks, and reinvent the wheel over and over again.
Does it still only support WEP? I also presume 3G modems are out of the question.
While that’s certainly part of it, I think you’re overlooking the fact that a great deal of work over the last 10 years has gone into making these OS’s scale better. Look at the BKL work in Linux, for example, and the changes in Windows have probably been even larger.
Anyway, in the end what does it matter? If i can run 20 videos simultaneously no matter which OS i choose, why wouldn’t i go for the one that i like best?
Because BeOS never locked up or slowed down. Ever. Right.
So you’re “disprooving” my comment about KDE/Linux with flaws of Windows??
Yeah.. makes sense…
That’s like saying that Haiku sucks because NetBSD has a flaw of some sort.
And yes, hardware also caught up. Now I have GPU acceleration for many videos. HD playback with only 2% CPU utilization, baby!
Works fine, actually, on my win 7 peecee.
But playing a buttload of videostreams must be the worst metric ever conceived, since 99,999% of all users play only one stream at a time.
As I already mentioned on another comment here, the purpose of this sort of demo is not to show the practicality of playing 10 videos at the same time, but to demonstrate the efficiency of the system under heavy load; the point being, no matter what you throw at Haiku, you will always enjoy a smooth user experience, even if you are using low end hardware.
Given the right (high-end, hardware accelerated) PC, Win 7 (or Linux) may also play multiple streams w/o a hitch. The point is, though, that using Haiku you don’t need a supercomputer with hardware acceleration to enjoy a system that does not annoy you with the occasional UI lagging and/or mouse jerkiness.
For example, I have a middle of the road dual core PC with 1GB RAM, and Haiku runs very smoothly not matter what I throw at it (ie., compiling Haiku, playing a video plus web browser, email client, IRC client and plus a few more apps running). On the same machine, Win 7 is a joke that it feels like I had downgraded my PC to a Pentium with 64MB of RAM, and Ubuntu runs acceptably, but the UI shows lagging and it becomes overall jerky (ergo, unusable) when, for example, I compile Haiku in it.
People have got used to the sort of subpar user experience that Win and Linux has delivered over the years, and unfortunately the notion that you need super-machines to enjoy responsiveness has settled in a lot of people’s minds. In a way, the multiple video demo tries to demonstrate that it does not have to be that way.
Some people put emphasis on that special feature – even the Haiku Users Guide:
http://www.haiku-os.org/docs/userguide/en/tracker.html#navigating
I myself don’t use it much. Haiku has a rich set of keyboard shortcuts and also an optional filemanager mode which is more browser-like, which some people prefer. Haiku has all the UI features that BeOS borrowed from MacOS (classic) and Windows. Things like Alt-Tab, spring-loaded folders and cut/copy/paste in the file manager.
Haiku can’t do so much more with the graphics feel until it gains support for accelerated 3D graphics (perhaps through Gallium) which would allow more complex composition, partial translucency, live window thumbnails, etc.
About the freezes.. I don’t see any system hangs on my quad core, a few crashes a month maybe (with everyday use), but I’m using non-release builds. The system is at the mercy of its device drivers, and ultimately, one’s hardware, so it’s not necessarily “Haiku itself” that is to blame, but that is perhaps a bit of an excuse.
I know the Haiku devs work hard on code quality. It’s just difficult to build a bug-free system that supports most hardware out there. It’s actually quite amazing what Haiku has acheived with so few people.
I would have liked for Haiku to be a bit more revolutionary, but the project had to focus and a straight clone of BeOS (+select improvements) was what the community could find consensus around. Haiku will continue to evolve as long as people keep working on it.
Whether or not it’s worth our time.. Whether or not it’s worth -your- time.. it all depends on what you want and expect.
I have never used BeOS, so I don’t have the nostagia thing that probably makes it interesting for some people.
Compared to modern OS’s, it seems awefully dated. It actually caused severe pain in my “mouse arm” because of all the insane menu navigation.
I have to agree here: I only tried some trial version of BeOS years and years back and never used it for more than a few minutes so I lack the feelings of nostalgia. With the rose-colored shades of nostalgia missing I can only say that Haiku looks plain hideous to my eyes.
Yes, technology-wise it’s got enormous amounts of potential and I can bet my ass that it’s going to be a strong contender for certain things in the future, but if they wish to make it attractive for people without the rose-colored shades they really need to work on the looks. And I fear that’s not going happen; given how many of the core developers are die-hard BeOS-fans they will not want to ‘tarnish’ it with some fancy effects or more modern looks.
Well… Haiku’s app_server is themeable to a certain degree (Decorators).
But as there is no 3D acceleration, most of so called modern effects are simply not practical and it is actually better to keep things simple.
Considering that GUI is not accelerated I’d say it looks quite ok and professional.
I don’t think that nostalgia is the sole (or even primary) driving force behind retaining the BeOS visual appearance in Haiku. Minimalism was always one of the main design (and philosophical) goals behind the OS – a minimalist, understated interface is consistent with that.
Of course, aesthetic preferences are just about the most subjective & widely-varying opinions that people can have. Even leaving aside nostalgia, I find that the mix of subtle/understated UI & clean but slightly-cartoony graphics is easier on my eyes than just about any other UI I’ve used.
To me, that says that the visual designers paid close attention to detail – but they also had the restraint to avoid going completely overboard with gratuitous visual effects (as opposed to effects that serve a useful purposes, E.g. to enhance usability by giving better visual feedback).
Actually, there’s been significant discussion among Haiku developers about visual changes to the OS – just not for R1. E.g. there’s a “Glass Elevator” sub-project that’s been around almost since Haiku’s inception, with the goal of more long term, “forward looking” changes after the immediate goals of R1 are met:
http://www.haiku-os.org/glass_elevator
And there already have been a number of visual changes to Haiku that, while subtle, are immediately-obvious to any long-time BeOS users. And unlike much of what I saw of ZETA, the changes have all been for the better (at least to my eyes).
There is a general perception that, because they’re recreating an “old” OS, they must be die-hards who are clinging to the past, etc. But I don’t think that the age of the OS is very useful indicator on its own – for one, in many ways BeOS was quite advanced for its time and had features that have only recently become common in mainstream OSes (IIRC, “future-proofing” the OS was one of Be’s goals). And every indication I’ve seen is that that the Haiku devs are entirely realistic about changes that the OS will need if it’s going to have any viable future.
I don’t think that nostalgia is the sole (or even primary) driving force behind retaining the BeOS visual appearance in Haiku. Minimalism was always one of the main design (and philosophical) goals behind the OS – a minimalist, understated interface is consistent with that.
Could be, or could be not. I don’t frequent Haiku-related boards or follow their discussions elsewhere, I’ve just stumbled across a few blogs and a few discussion threads every now and then and in those I’ve seen lots of people with those rose-colored shades screaming murder at every idea regarding enhancing the looks of the UI. So yeah, I admit that I might have just been looking in the wrong place and gotten the wrong impression but that’s how it is.
Of course, aesthetic preferences are just about the most subjective & widely-varying opinions that people can have. Even leaving aside nostalgia, I find that the mix of subtle/understated UI & clean but slightly-cartoony graphics is easier on my eyes than just about any other UI I’ve used.
Of course, there is nothing wrong with liking such style. People have different tastes and all that. I was just saying that I doubt the common populace will find Haiku’s looks pleasing and will call it ugly or out-of-date.
I, too, find Haiku hideously ugly, but as I said, I really like the underlying technology. I do wish the devs luck with the project and I hope to see Haiku going strong for years to come.
but they also had the restraint to avoid going completely overboard with gratuitous visual effects (as opposed to effects that serve a useful purposes, E.g. to enhance usability by giving better visual feedback).
There’s plenty of ways to enhance visual feedback without going overboard, and I agree to an extent: all those glass-effects in Win7 are rather annoying. Such gimmicks are mostly useless from usability standpoint.
But well, let’s see…I use Compiz under Linux and I have configured this one plugin so that the windows lose opacity and color saturation the longer they are idle. Ie. a window that I haven’t used for a while only has 20% color saturation and 80% opacity. On the other hand, the window I am using has 100% opacity and 100% saturation, and any windows I have just interacted with has the same. Hardly a gimmicky eye-candy effect, but gives plenty of useful visual feedback. Though of course, not everyone would like it.
you never had a document open on the second screen while working with something on the first one, do you?
you never had a document open on the second screen while working with something on the first one, do you?
Umm. What does that have to do with the example I gave?
I think he means that sometimes you need to have a full opacity view of windows that aren’t active (like if you are using another window’s content as reference for something, not exactly a biggie though since these effects are/should be easy to turn off).
Well, I can’t say I find it hideously ugly, but I can’t say I find it particularly visually pleasing either. That said, I doubt that this has been any kind of priority with the development team apart from Stippi adding some gradients and a nice set of svg icons. I’d wager that once Haiku get’s 3d acceleration (through Gallium most likely) we will see more entusiasm/effort going in to modernizing the look of the gui. Also there are the aforementioned decorators which allows third parties to customize the GUI and thus allowing for ‘themes’ to be made. As for the menu system, sure it may take nesting to the extreme, but seriously who launches their apps from the menu other than in extreme cases these days? I use shortcuts/launch icons for practically all the non cli apps I use and I doubt I am the exception. Currently the focus is on providing a stable, capable and performing system, once that is in place then making a better looking gui will be a piece of cake in my opinion
As for the menu system, sure it may take nesting to the extreme, but seriously who launches their apps from the menu other than in extreme cases these days? I use shortcuts/launch icons for practically all the non cli apps I use and I doubt I am the exception.
Truth be told, having the ability of using icons on the desktop is a really poor excuse for having a messed-up menu. First of all, if you happen to use lots of different applications and occasional gaming you’ll end up with horribly cluttered desktop. Secondly, it still makes those cases where you have to dig in to menu any better.
But is it so bad? I mean for things like applications/preferences (which is what you will most likely launch) it’s not a hell of alot of menu nesting is it? Certainly no more than launching applications from the windows start menu unless I am mistaken?
I always wonder about comments like this about the right click navigation: it’s an additional feature, it doesn’t replace the other and usual means to reach your data or applications.
It’s completely okay to dislike it, as no one forces you to use it. Just stay with what you know, other people happen to find this little feature exceptionally helpful at times.
Those are not SVG icons! Stippi didn’t just design some of the icons, but also invented a brand new file format. And this was actually a quite important improvement! The HVIF, or Haiku Vector Icon Format, is so compact that the icon usually fits in the file’s inode in the file system. This means that the icon data is right there when a directory is to be listed, without first reading the content of separately stored icon files, bringing very welcome speed improvements. Rendering HVIF is also much, much faster than rendering a general purpose vector format, especially SVG which is XML based.
Some details on the format;
http://www.haiku-os.org/articles/2009-09-14_why_haiku_vector_icons_…
Very well said. The Haiku UI is not about nostalgia, but is simply a manifestation of living up to the BeOS philosophy of minimalism that is both pleasant and — more importantly — does not get functionally or visually in the way of the user.
Edited 2010-08-22 13:53 UTC
“Is there a point to this project beyond the fun of doing it / nostalgia?”
Yes. I also have never used BeOS before using Haiku. Compared to other operating systems and user interfaces, I prefer to use Haiku. I like the consistency between applications. I like the responsiveness of the user interface. I like the clean look and nice default settings. I like the simplicity of installing and uninstalling applications. I find that the user interface better fits my work flow.
So the point of making Haiku is that, similarly to how not everyone wants to use iOS or Windows or Mac OS X or Linux, there are people who want to use an operating system like Haiku. Well, especially when it considered stable enough for an official release.
As another user of Haiku (Yes, i use it as my day-to-day OS where possible, i’m posting this from Haiku
) I have to admit that the main reason i use it is because of it’s UI and the general feel of using it.
The media-friendlyness and sub-10-second boot times help too.
Ever since the emergence of netbooks I’ve been saying that Haiku would become the perfect OS* for such a device. The boot time, the rich multimedia capabilities and the ultra-low overhead all come together to make a great low-end netbook OS.
As much as I love OS X, I’d take an eeePC or similar with Haiku over a 12″ PowerBook any day simply for the portability and speed, not to mention the price. A good condition used 12″ PowerBook (the smallest Mac laptop ever made) can cost upwards of $400, and it is limited to running Leopard or outdated Linux distros. A netbook can multi-boot Haiku, Windows, Linux and/or FreeBSD resulting in a well rounded and very capable device for under $200 new.
Damn, now I’m itching to restart a project I had begun and abandoned several years ago: Putting BeOS r5 on a Toshiba Libretto 110ct laptop and getting all the hardware to work right. Looks like I’ll be dusting off the r5 CD and trolling eBay tonight…
*I know it’s not there yet; there are still a lot of issues to work out before it becomes a good mobile computing OS, not to mention a good desktop OS. Of most immediate concern to me is lack of wireless and accelerated video drivers. I haven’t lost faith though!
As a note to my previous comment, I run Haiku on both a Desktop system and an eeePC 901 successfully. The EeePC has reasonable graphics speed with the Intel Extreme graphics driver (I can comfortably watch video on it), and Wireless (WEP only ATM) is supported using the wifi card out of an old 701. It’s enough to get by at home and on unencrypted wifi networks in town.
In general, Haiku supports all my systems very well out of the box. I’ve yet to have any real problems with Haiku when using it as my main system, apart from the lack of software (Outdated Bittorent clients and alpha quality messenger software abound)
Overall, the current state of Haiku is a very positive one. I’ve had terrible experiences with Linux, and have never really seen eye-to-eye with X, which puts me on a prejuciding foot when coming to UNIX-like systems. Haiku is a very capable breath of fresh air when it comes to open-source operating systems, and things like the Services Kit mentioned in the article are starting to push it in a more desktop-orientated direction, which i think UNIX-like OSes provide very poorly in
About bittorrent, if you have Qt installed you can use the newer Transmission client with qt gui.
http://qt-haiku.ru/index.php?option=com_rokdownloads&view=file&Item…
Transmission 2.04
Released: 13.08.10
Seems to work very nicely.
But if you need a messenger client (as in msn messenger) i think you are out of luck.. allthough if you use ICQ, Jabber GTalk/Ya.Online/LiveJournal.com or Mail.Ru QutIM might do the trick for you.
The Haiku devs is doing the best they can. While we are thinking in “look-and-feel”, and app level, they are thinking about multiple hardware to run the stuff, and make all more stable.
If you wants to go in a particular direction, please join, and contribute. Will be very appreciated.
The topic of web services degraded to “I don’t like how the window manager looks” and “the menus are too difficult to navigate” crap arguments.
I used BeOS for years and found it far superior to Windows and Mac OS at the time. If BeOS can manage to come to Beta, support all the modern networking requirements, I’d hop on board and develop apps for it – I assume others, old timers and newbies alike will do the same.
It has a wonderful, minimalistic interface that is functional, pervasive multi-threading, a great file system (maybe not the fastest, but great to work with), and the API was easy to work with as well.
I use Windows 7 and Mac OS X mostly these days, but I would trade all the glitz and glam in a heartbeat for a version of BeOS that had the look and feel of 4.5 and up that I could work on for web development.
AFAIK, this is the first time that an Haiku/BeOS news stays on the top of screen for a whole day without any other news entry.
Speaking about UI, is very strange some Haiku users/developers talk against eye-candy stuff in the system, but praises the Win7 and OSX ones.
About Service Kit, the student don’t mention any commit deadline..
Personally, I don’t see a whole lot of eye candy in OS X, at least that isn’t functional. Some examples:
Dock Magnification: Helps someone like me who has 30 or so dock icons to better see which one I’m trying to open.
Expose: I have it set to the upper right corner, and with a flick of the mouse I can see all the windows I have open. This is a useful and logical effect.
Spaces: I don’t use it myself (I’ve never gotten used to multiple desktops) but OS X’s implementation is less flashy than say Compiz on a Linux distro.
Stacks: Very handy for getting to all the other applications I don’t have room for on the Dock; I changed it from Auto to List though, as it kept trying to do a grid and that was just too much for me.
There are many more examples but those are representative. Of course, there are also a few gratuitous and non-functional effects, such as translucent menus, minimize animations, window shadows, and my personal pet peeve, the translucent menubar.
Then there are some eyecandy-esque things that are missing, such as native window shading, that would make the UI even more intuitive.
I’m with you on the Win7 eye candy though. As much as I do like that OS (compared to previous Windows versions, anyway) I don’t see anything remotely useful about fat, blurry translucent window borders and grossly oversized buttons. Don’t even get me started on Aero Peek (why the hell would I need to see my desktop icons if I still have to minimize my open app to select them?), and Shake (it simultaneously vents and induces frustration, what a concept!).
I think the most useful new UI tweak in Win7 is the Snap feature; When I’m in 7 I use that all the time, and it’s actually not that flashy.
I don’t know what to think about Haiku. Its APIs sound great, but when it comes to actually operating the system as a user, I always find it plain horrible (ugly, complicated and technical, small UI controls and heavy menu navigation just like in the 90s…). They achieve some great performance, but their goal of resurrecting BeOS sounds like an inability to accept its past failures. And so on…
I don’t like to use this OS right now, and I don’t want to code on an OS which I don’t like. But still, I see some potential. So I’ll just wait and see what it becomes as time passes…
Edited 2010-08-23 09:01 UTC
The failure of BeOS wasn’t one based on technical capabilities, or a lack of user desire for a solid computing platform with decent performance.
This thread is both very exciting and very frustrating to read. It’s exciting because it reflects the the fact that Haiku is nearing a tipping point. It’s stable, the speed is good, Bugs are quickly fixed, lots of classic BeOS apps run fine, the development environment is mature, it’s gaining visibility. That’s all good.
But there is a lot of misinformation about Haiku. There is a lot of “arm chair quarterbacking” by folks who are interested but haven’t taken the time to install Haiku and take it for a test drive. Like for instance: “and also adapt it to modern hardware.”
Haiku runs fine a lot of modern hardware for basic purposes. Lots of drivers are supported. Here are instruction for installing Haiku as a dual boot. It also runs embarrassingly well on old hardware:
http://haikuware.com/wikis/doku.php?id=tutorials:install-haiku-besi…
The example of BeOS running many video streams without dropping frames is driving me crazy so let me address that. This was a great demo for 10 years ago. But just like the spinning teapot was cool, then wasn’t cool, I think its time to find a better demo for Haiku.
What is the major strength of Haiku? It is the application APIs and the “mini” kernel. In operating systems theory class you learn about the trade-off between through-put and response time. An example of through-put is database transactions per second. Linux is very good at this, Solaris is/was best at this. Because the kernel was designed and tuned for throughput. BeOS was designed to be the best at quick response to user-centric applications. “It feels fast” means it responds quickly to GUI events. Why would one OS feel faster and not block on GUI operations? Because the application APIs and kernel were designed for minimal latency in the kernel.
These days I think you could probably play 6 or 8 video streams on Windows or Mac with no dropped frames. th4e hardware is so much faster and the video subsystems are much improved. It also not that exciting because how many videos do you need to play at once? Maybe 1 or 2. So playing multiple video streams on Haiku is no longer a great demo.
So what is a great demo on Haiku? I think we need some new ideas here. We need to write some applications that reflect what people do on PCs today AND use the application APIs to show it off in Haiku.
So I give this challenge: Write a simple app for Haiku and use the application API. It’s fairly simple and very straight forward. There is an excellent tutorial for programming apps on Haiku. It starts with no assumptions in chapter 1, teaches you basic C++, and moves up to basic API/GUI programming by chapter 20. If you have never coded before but are interested, start here. If you are a seasoned coder but have never written a Haiku app, start here but skip ahead:
http://www.haiku-os.org/development/learning_to_program_with_haiku
I have more to say but maybe this is enough for one post.
Happy Haiku hacking
OK, more responses to more comments.
“Do they have a goal for the final R1 release or is it more a situation of plodding along and focusing on the quality rather than an arbitrary time table?”
I used to ask the same question, when is the final release, so I can use it. After using Haiku A1 I was surprised at how much I could do. And with A2 now in the rear view mirror, Haiku is “good enough”. There are a growing number of Haiku users who are switching to it for their daily use. Sure, these guys are early adopters but they are able to get their stuff done. The number of regular Haiku users is growing. You don’t need to wait for R1. You can use it now.
“Does anyone know of their plans for LLVM integration?”
Grzegorz has done a proof of concept. But this does not have high priority. It doesn’t add any functionality:
http://haiku-os.pl/node/1244
“With the rose-colored shades of nostalgia missing I can only say that Haiku looks plain hideous to my eyes. ”
I would agree that the Haiku look is a bit dated. It needs to catch up to OSX and Windows 7 in several aesthetic areas. Until now the project has been focused on functionality, and rightly so. Better to get it working first, make it pretty second. One of the major strong points of Haiku is that you have a single C++ API. This means that Haiku GUI can be easily extended by refactoring the look and feel code. If someone with serious window manager chops wants to step up and become a hero, this task is waiting for you. One person could make a huge difference here and could make a real name for them self. Just do it.
“While its cool, video and audio editing on BeOS is joke. There are no applications that are good enough for professional work.” I know of 3 audio editing apps that work on Haiku. 2 of them might be good enough for semi-pro work.
http://haikuware.com/directory/multimedia/audio/audio-editing/
There was commercial video editing app on BeOS but it is no longer supported. Writing a good video editor is a real challenge. I have seen numerous projects started and abandoned on Linux. It would be nice to see a basic video editor on Haiku. I wouldn’t expect to see professional apps of any kind for a while yet. First we have to show clear advantages to using Haiku, then add lots of users, then come commercial apps. This is our next challenge.
“I haven’t used their new web browser, but its lacking plug-ins and pixel32 wasn’t really usable last time I tried it on BeOS.”
Yes, not too many plug-ins. But WebPositive is surprisingly good. It lacks many bells and whistles but it is quite good at the basics, and quick. I ran Google apps with no problems. As for image editing, try a native app like ArtPaint:
http://haikuware.com/directory/view-details/multimedia/graphics/ima…
Edited 2010-08-23 18:15 UTC
Its still alot of C code there. I hope they change that in future versions.
Most of the apps are vapour ware. Cold cut was good though.
I runned them all. I used BeOS for maybe 6-7 years.
The BeOS style multiple video demo is still VERY useful. The problem is that the purpose of this demo is misunderstood. This sort of demo is not meant to show the usefulness of Haiku, but to demonstrate the responsiveness of the system under heavy load, something which is at the heart of the Haiku user experience and that we seem to agree is a major strength of Haiku.
Look at it this way: running multiple videos w/o skipping a frame, while spinning the Teappot in a system that only has software opengl and runs in VESA mode video is a very effective way of showing how efficient Haiku is at using hardware resources.
OK, but the videos had better be interesting 😉
Of course! My favorite is the one that shows BGA giving his OpenBFS presentation at the Kansai Open Source Forum conference in Japan. Another one ranking high in the list of demo videos is JLG giving his speech of support to Haiku at our Google Tech Talk a few years ago at the Google headquarters in MTV.
The problem is that all the competition can do the same thing, so it’s not really showing much that’s unique. Unless maybe they can’t on netbook hardware that’s slow enough? I don’t have a netbook to test, but certainly they all run with ease on a modern desktop.
I think a better strategy might be showing quick 30 second youtube clips, where you:
open a netbook and boot
start an email app
make a quick reply to someone
shutdown
all while a split screen shows another computer that is still trying to boot up.
I honestly don’t see myself running Haiku on a desktop system anytime soon, but if it’s really that quick then i think that could be a major selling point for people lugging around laptops, iPad clones, etc.
The single most common reaction from first time Haiku users is a “wow, this thing is faster and more responsive than anything else I have ever used” type of comment. There is a reason for this. Windows, Linux and Mac users are used to sluggishness and the occasional choppiness in the interface, so they consider it normal.
Try Haiku on the very same hardware that you use your other OSes, and you will finally experience what having a responsive user interface really means. Only then will you come to the realization of how sluggish the other OSes actually are. This cannot be explained with words; you have to experience it, and once you do, I am pretty sure you will agree that it sucks to go back (to Windows, Linux, Mac, etc.).
The thing is: there’s no easy way to measure responsiveness, otherwise it would easy to benchmark OS and to improve them..
*So* there’s no easy way to demo responsiveness.
Still one possibility would be a ‘split screen video’: show some activity such as starting an application in parallel on Haiku,Windows,Apple.
It might not be easy on Windows or OSX but applications can be scripted. On Haiku you can easily do application scripting with hey. Hey uses BMessages, another nice feature of Haiku:
http://www.birdhouse.org/beos/bible/bos/ch_scripting6.html