Slashdot Mirror
After 8 Years of Work, Be-Alike Haiku Releases Official Alpha
NiteMair writes "The Haiku project has finally released an official R1 alpha, after 8 years of development. This marks a significant milestone for the project, and it also debuts the first official/publicly available LiveCD ISO image that can be easily booted and used to install Haiku on x86 hardware. Haiku is a desktop operating system inspired by BeOS after Be, Inc. closed its doors in 2001. The project has remained true to the BeOS philosophy while integrating modern hardware support and features along the way." Eugenia adds this link to an article describing the history of the OS, along with a review of the alpha version."
Operating Systems are not trivial and hardware support is a real pain. It takes years even for large communities to do this and even a community as big as Linux's doesn't always get it right, neither do some companies for that matter. They look as if they're a small team trying to do a great deal.
I remember using BeOS on an old Pentium 166MHz with little RAM and being able to play many songs, browse and play videos and the same time when Linux and Windows struggled to do any one of these on the machine.
Sure, most people won't be interested, but variety is the spice of life and if some of the good aspects of BeOS get adopting, it will be a good thing for everybody.
"If anyone needs me, I'm in the angry dome."
It'd be better if they all came to a consensus on where libraries go and follow the Filesystem Hierarchy Standard and a package system.
That which does not kill me only postpones the inevitable.
Congratulations to the Haiku team. Back when Be closed its doors, I remember there were several projects to recreate the OS, but most people didn't expect any of them to succeed. This announcement proves that wrong. BeOS was a fantastic OS and with Haiku making strides toward a stable release, the legacy can live on. Although it's taken a while to get this far, writing a full operating system from scratch takes a long time. Even large companies with dedicated teams generally take 5+ years to build a new OS, so 8 years for a group of volunteers to release a working system is quite reasonable. Once again, congratulations and thanks for all the hard work you've put in over the years. Although only an alpha, this release is quite stable and usable. Your efforts have certainly not gone unnoticed.
For every post, there is an equal and opposite re-post.
Haiku != Linux. Amarok is very deeply connected to KDE/Qt APIs which are, of course, not implemented in Haiku. Although ports can be considered (Firefox is a must), maybe a player designed for Haiku's APIs would be best for Haiku at this stage, even as a showcase.
No one has really answered you so far, surprisingly. I don't really know BeOS internals, but having toyed around with it as an ex-Amiga user looking for a modern equivalent (like many others), I can give you the general idea.
Basically, it's this: unix sucks.
Lol, it's flippant, but for all the greatness of Unix and Linux, especially compared to Windows, there's a definite truth to this. The problem is that unix is a few simple (and strong) principles from the early 70s, upon which nearly decades of evolution have occured. The fact that this was even possible is a huge testament to the flexibility of those core principles. Nonetheless, most of the evolution since is essentially a big hackish attempt to keep Unix up to date. For instance, go to phoronix and search for graphics stack. You'll find a lot of discussion about Xorg, the Linux kernel, graphics drivers, GPUs, libraries, the linux console, and how none of them are really consistent or integrated, and the problems that result. Moreover, Unix was originally designed for many users sharing a huge, expensive computer. It's not really designed for personal computers at all. Arguably, this distinction isn't so relevant these days.
BeOS, on the other hand though, is an attempt to make a modern, coherent, friendly, desktop operating system for personal computers. It's designed to be quick, to have a logical stack of libraries that cooperate (such as for audio and graphics, again, unlike Linux's audio/graphics stack).
Essentially, the point is just to build a modern system, and dump all the old, legacy cruft that just gets in the way. It's an attempt to draw a line under the past, and say, "OK, that's the old way. From now on, programs should use this stuff instead, so everything looks good and runs well, and integrates nicely."
So, a unix-like kernel with a pretty window manager is modern?
Damn. That's some strong kool-aid.
That's still Free Software by the FSF's and most other people's definitions. What it is *not* is copy left. So yes, you can make non-free derivatives. But the rest of the world will still have the previous, open source releases available. You even have the freedom to create a GNU-focused Haiku release if you really wanted to - it might be worth it, just for the looks of horror at the idea of a GNU/BeOS (I'd use it!).
Apple would have gotten a better operating system for their purposes out of BeOS, but they got Steve Jobs with NeXT. Or was it the other way around?
I think what you're looking for is "NeXT purchased Apple for negative 429 million dollars."
Village idiot in some extremely smart villages.
Apple did learn; they have their own retail stores. They don't rely on companies that make most of their money selling MS products and MS-related products for their business. Microsoft can't offer the Apple Store a discount on Windows if they don't sell OS X.
Be failed because it messed its customers around. Their first releases were for PowerPC and ran on Macs and their own hardware. Then they added support for x86, and didn't provide cross-compiler toolchains, so most third-party apps became x86-only and the people on PowerPC were left in the cold. Then they announced that they were going to switch focus to BeIA, and frightened third-party commercial developers away from BeOS. Then they turned down Apple's offer, demanding ten times what Apple was willing to pay, and eventually had to sell to Palm for around 20% of Apple's offer. Plam did very well out of the deal, paying $11m for the company and then getting $23m from Microsoft in settlement of the suit over anticompetitive practices.
I am TheRaven on Soylent News
Ubiquitous, indexed, filesystem metadata. I didn't need an address book app, or a music jukebox app with BeOS. MP3 tags were extracted and stored as filesystem metadata and so I could browse my music by artist, album, genre, and so on, from the Tracker. Linux, Windows and OS X all, now, include extended attribute support in their filesystems that make this possible, but they are not used.
I am TheRaven on Soylent News
So you are basically saying that if something already exists, we shouldn't create new things?
-- dnl
It's worth noting that the CLR is heavily based on the Smalltalk-80 VM
In what sense? Pretty damn sure it's not the code; if it is design, then can you please explain in more detail what CLR design decisions are "heavily based" on Smalltalk VM?
C#, semantically, is almost identical to Objective-C, it just has more C++-inspire syntax.
Uhh, seriously, WTF? Since when is a statically typed OO language "semantically almost identical to Objective-C", which is based on Smalltalk's message-passing, dynamic object model, with its hallmarks such as the ability to handle arbitrary messages sent to your object, and redispatch them elsewhere?
Remember that, originally, there were 2 main families of OO languages - one static, started by Simula-67, another dynamic, started by Smalltalk. Objective-C has Smalltalk all over it; on the other hand, C++ is definitely a Simula grandkid, but so is C# - in fact C# is perhaps even more so, since virtually every Simula concept has direct representation in C#, including such bits as single-inheritance, value/reference type separation or virtual concept and keyword.
Something that's semantically almost identical to Smalltalk (and thus much closer to Objective-C) is Ruby.
Regarding F# - it isn't really all that modern as such (I mean, it's explicitly just another CAML dialect!), though it does have some nifty ideas in it like active patterns or units of measurement (which have been seen elsewhere before, though). The nice thing about it is that it's an attempt to take a mostly functional language with roots in academia, and put it in the mainstream by teaching C# and VB developers to appreciate the power it gives, and sticking support for it into an IDE they already use daily.
That about matches what I've read of the whole affair. Didn't know that Palm bought Be for so little though; that's been a harsh lesson for someone I'll bet.
Does anyone happen to know why Apple only wanted to pay about $115M for BeOS, when they eventually paid something like $400M for NeXT? Did they just think NeXT was worth more (that they'd need to spend a lot more developing BeOS maybe), or did they just run out of options and get desperate by the NeXT stage, I wonder?
If any one wonders how does Haiku look in an OS, here is one screenshoot
http://commons.wikimedia.org/wiki/File:Haiku_Screenshot.png
So, it looks like the Be theme in KDE?
Kind of a joke there, but also a bit of a serious comment, too: can't tell much about an OS from a screenshot like this.
Bow-ties are cool.
What advantages does that offer the OS and what are the downsides (lack of binary future compatibility when the compiler changes?)
I used and loved BeOS, and AmigaOS, and I still don't care about Haiku.
BeOS was amazing because it was written by a group of dedicated developers with a razor-sharp vision of how to design a great OS.
Haiku is an attempt to copy what those guys did a decade and a half ago.
One is really a lot less exciting than the other.
For all these years, I have held onto the Spellswell source code, and kept it safe, knowing that someday the Phoenix of Haiku would rise from the ashed of Be, Inc. (Or rather, I just don't like to ever throw anything away.)
I also still have all the protocol specification documents. I just gotta organize them and throw them up on the web again.
Word Services actually still works on Mac OS X, but not yet with Spellswell. We never did Carbonize it. Eventually Working Software was dissolved, and we all went our separate ways. But I expect I'll release an OS X-Native Spellswell at some point as well.
Some things never die... Spellswell was originally published by Green, Johnson Inc. before Mike Green and Dave Johnson split up into Cassady and Green and Working Software. My understanding is that it could check Microsoft Word 1.0 documents on the 128k Mac. It was a huge hit, before Microsoft added a built-in speller to Word.
A lot of that code from 1984 is still in there, for example an incredibly elaborate dictionary file format that provides compression while at the same time being editable.
Request your free CD of my piano music.