“BeOS is a marvel of software engineering and creativity. We all know it, use it, and love it. We are also re-creating it for the pure fact that we want to see it continue and thrive. But before it can be successful three things need to happen” Editorial from BeUnited‘s Simon Gauvin about the future of OpenBeOS. Also, yellowTab and BeUnited are to join development and promote the open source BeOS standards.
OpenBeOS and the Big Bright Future
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
“The clones currently have a 0.1% share of the desktop market. Developers are trying desperately to make these clones easy to use based on X (KDE, Gnome) but have not succeeded in providing the level of simplicity required by today’s average user. The main reason is that Linux does not solve a problem on the desktop that is more compelling than the cost of Windows”
The very last sentence is exactly what I tried to told “Camel” and a lot of linux zealots.
Now, all I have to say is: “Go OpenBeOS, GO!”
🙂
I think it’s a MAJOR mistake to never sell OpenBeOS. I say, unless it’s in a pretty box and available at Best Buy or CompUSA or Circuit City or WalMart or whatever, it’ll never be able to compete with Windows or even fill the gaps that Windows doesn’t.
Let people walk by the pretty box in the aisles 5 times. The fifth time, maybe they’ll pick it up. The 20th time, maybe they’ll buy it. Lather, Rinse, Repeat.
I think selling it for a realistic price, say between 19.99 and 39.99 is a smart choice. Of course, like commericial Linux distros, you should be able to download ISOs, but it should be for sale too.
And hey, if no one else does it, thanks to the MIT license, maybe *I* will. Anyone have any capital to invest?
I think this is great! I hope the vision outlined does come true! I have heard nothing but great things about OPEN/BeOS. I would like to try it out, where can I learn alot and find installation file?
This is an incorrect statement by Simon regarding the desktop market share owned by Unix.
Linux has about 0.35% of the desktop market, not 0.1%. And if you count *BSD and Solaris when used as “desktops”, it is about 0.5% overall. That is five times more than what Simon wrote.
BeOS had many great goodies in it and was a quite fun system. However, its creators made the ultimate mistake of base its APIs on C++. Frankly, I think they should have avoided this error. Even at the time BeOS was designed, it was starting to become obvious for many of us that C++ was real crap. The guys at NeXT did the right thing with the Objective-C choice, and now Mac programmer are happy using cool, dynamic object-oriented frameworks inside Mac OS X. You will not compete with C# & the CLR from Microsoft, java based systems or Mac OS X (Objective-C) with something based on C++, IMO.
Unless there is some amazing breakthrough on any of these hobby platforms, including OpenBeOS, these operating systems will always be good for hobbying around and not much more. They are fighting against time as the world evolves around their efforts.
I think Simon Gauvin sees the edge of this reality, but doesn’t really want to embrace it.
Windows did not succeed because of clean elegant design. Microsoft marketed it, bundled it, and wrote amazingly good apps for it.
Linux on the server side is being marketed, supported, and there are some really great apps for it.
As OpenBeOS will not have the infrastructure in place to support really great apps for 2-3 years, it is destined to be nothing more than a quixotic dream.
Maybe they should get a bizdev person on board and figure out how to buy/license BeOS from Palm? Start out with a solid base and then build something really fabulous on top of it?
Time is of the essence. And timing is of the essence. If there were an OS project targeting the state of the world 2-3 years out, it would not be recreating an old OS. It would be figuring out what platform will the apps of that time need to really shine. And how do we build that platform and some compelling sample apps.
Amiga/AmigaOS was more successful than BeOS will ever be because it offering something compelling. It allowed the user to experience computing like they never had before with amazing speed, color, graphics, video, sound, etc. With the Amiga platform, NewTek was able to offer video solutions to professionals that were a fraction of the cost of what they had been paying. What happened with BeOS? Basically nothing as there never were uses of it or applications that would have made switching to it a compelling choice for the user.
Looking at the OpenBeOS site, I could not find the answer to what they are aiming for. Simply, what is the goal for what OpenBeOS is going to offer to the user 2-3 years from now?
It had better be more interesting than the amazing features that OS X will offer. And it had better be more compelling than the revolutionary system Longhorn will be, with a full database file system and a hardware accelerated 3D GUI and a killer suite of apps all tied together with web services and XML. And it has to be more compelling than Linux which will offer a good desktop, a good kernel, a great file system, and many many server apps.
#m
I’m nearly speechless after reading that article. The man nearly comes off as a lunatic. His marketing opinions are all fine, but he has a few outrageous comments in there.
our dream OS is around the corner!
Riiiiiiiggggghhhhhht. As I’ve said before, I would love to see a functioning OpenBeOS. But “right around the corner”? They don’t even have a basic system yet. Face it, folks; there’s still a very long way to go. It took Be, with more than 50 full-time engineers and some (limited) help from hardware companies, years to get the BeOS out the door.
A vast majority of Windows users would not remain with Windows if they had another reasonable choice.
To say this, “reasonable choice” must mean, “runs all current and future Windows software perfectly, as well as running all current and future hardware at least as well as Windows can.” OpenBeOS, as far as I can tell, currently has plans to tackle neither of these two issues.
It does not take a genius to estimate that some percentage of these users would like to use an operating system that is simpler, more stable, faster, and cheaper than Windows to run on their x86 based applications.
It “does not take a genius” to know that current versions of Windows are rather stable and are apparently fast enough for most people, since a lot of fast, dirt-cheap hardware just isn’t selling. And these are current versions of Windows, let alone what will be out by the time OpenBeOS finally ships. So the selling points become “cheaper” and “simpler”, which are valid in themselves, but not particularly compelling enough to say, “Ditch all of your software, hardware, and training, and switch!”
OpenBeOS has the opportunity to fill this HUGE untapped market!!!
This one’s my favorite. The “huge” market of people that run few or no applications, have no need to exchange data with other people (Office formats, etc.), and run on an extremely limited set of hardware. Not to mention people that can understand why a program won’t run, why a particular website won’t load, and why they need to verify any new hardware against a short compatability list before buying it enough to not be too upset.
So here we are, at the enviable position of being at the begining of a new era in desktop computing, creating a version of an operating system that has the potential to change the world.
Why is it that nowadays people are so enthralled with “changing the world” that sound logic and philosophy goes out the window? And I don’t just mean with computers. Why not simply focus on making a good operating system? Until that’s done, you have nothing to market; no solution even if you find a problem.
OpenBeOS may well turn out to be a great project, but Simon does no one favors by promising way more than can reasonably be delivered within the near-term. No doubt marketing will be incredibly important, but having a product that can boot beyond a very basic kernel level wouldn’t hurt, either.
<sarcasm>
yeah paul…C++ is such crap…that’s why most programmers use it
</sarcasm>
in fact it’s number one in the world and programmers aren’t locked into using C++ in anyway…ever hear of language bindings? i guarantee you can make Obj-C bindings for the BeOS API along with Python and Perl and plain ole C…some things will be a little more cumbersome…but come on…objects are so much cleaner than “handles”
C++ isn’t crap…gcc’s implementation until 3.1 was crap…but that’s another story
OpenOffice for OpenBeos?
Having said that … I love the hell out of BeOS. Even if it is a “hobby” OS as some have alluded too … it is one sweet hobby.
Be did not fail due to it operating system … it failed as a company due to a relentless fight against a Huge Mega Micro Sloth ass sitting its gigantic butt on the competition.
I believe to this day that BeOS should have been “given” to the geeks at the college and university level. Position your product correctly and market the living heck out of it.
After two years away from the BeOS community – I have returned and seen a living community of coders and supported coding away and keeping an excellent dream and OS alive.
Keep on coding!
Troy
OpenOffice is a pile of junk compared with Microsoft Office. It is a forgery that is much worse than the original painting.
If any platform is going to do another office suite, note that people want something MUCH BETTER than Microsoft Office.
And whatever you do, no matter what, do not let Miguel de Icaza and his bunch of Microsoft cloners on the platform.
REPLICATION is not INNOVATION.
#m
Doomed by C++ virtue : who says C++ has to be the only API for programming under BeOS?
Another quixotic OS dream project? : who says obos/beunited is trying to take of the world? It would be a nice side effect, but, not a requirement
Currently, obos is shooting for BeOS 5.0.3 compatability, no more, no less (well, maybe a little more).
I’m nearly speechless : didn’t you know all BeOS users are dreamers? these are the ideas of many in the BeOS community. it is not a bad thing. realistic, maybe not, but, dreaming never hurt anyone
Simon, please fix the numbers
: Eugenia, did you just break rule #7 
(btw, just kidding)
I may be wrong, but I seem to remember reading somewhere that the OBOS team isn’t planning to do distribution. I believe that the plan is to leave that to third parties, like perhaps YellowTab, to package it up and maybe polish it up for end users (or maybe even shrink-wrap it). Sort of like the OpenOffice > Sun relationship.
(And yes, I know that technically it’s “OpenOffice.org”, but if anyone is really pedantic enough to care, they can forgo pointing it out and go piss right off).
Where is Be today? Trapped in some sort of Palm business unit?
Where is BeOS today? Gathering moss inside Palm’s code vaults?
BeOS was some fine coding, don’t get me wrong. I still am a big fan of efficient code and elegant design.
But the mentality of the company was flawed from the beginning. There was no goal, no purpose other than to craft an amazingly good low-level operating system. All the media functionality never came to any sort of fruition as there was no focus on providing good ISV support. There was no focus on building higher layers of support for application programming.
In general, that was Be’s largest failing. The company did not understand the needs of ISV’s.
After their collapse, BeOS should have been released to the world as open source so that its evolution could continue. It could have continued to have an architecture board or other oversight group to make sure quality was maintained.
Instead it lies fallow.
It sounds like no one has picked up the phone and talked to Gassee and asked him what it would take to pry BeOS out of Palm’s death grip.
#m
Not because *you* hate C++ mean every programmers hate it. In fact, personally I became a BeOS addict especially because of the full C++ API. I just plain loved it. And from all the BeOS dev forums and usenets I knew, developers felt like me : attracted by the C++ nature of BeOS.
I’m fluent in a dozen of language, and everytime I have the choice to pick one for a project (hobby or real work), I never hesitate and jump in C++ when fit best.
I have no idea why you think the whole OS is doomed because it use the #1 software dev language.
Personnaly, I’ll kill to get Borland C++ Builder for OBOS 🙂
I am so disappointed when reading that main people coordinating BeOS rescue are so blind to FREE software principles and still count on corporate sector.
I am not a coder and I really wanna help OpenBeOS whenever and however I can, but heaving read this I feal lik ethis is going to be another waste of my time (after BeOS).
Is this great technollogy really doomed to bad managment and marketing ? Why protect corporate interestest … what did thay ever done for BeOS?
I don’t see why coorporations couldn’t build their products (applications or services) on GPLed BeOS replacment.
The Best 3D software is being ported to Linux without any problems!
Is there anyone Who Can really say something smart and make me not leave the (Open)BeOS community?
EUGENIA PLEASE PUT A POLL within BeOS community What do thay prefer GNU/GPL or MIT or Mozilla licence (which is maybe best one)!
> A vast majority of Windows users
> would not remain with Windows if
> they had another reasonable choice
Hahaha, get real! Other than ideological geeks, who cares about what OS people use to send email? For the vast majority of users, OS choice is not even an issue. Windows works real good, and I don’t see why or how a new OS can become so compelling.
> Windows is currently the ONLY choice of desktop
> operating system for 99.9% of the x86 users.
Well, you know the saying: there are three things: lies, damned lies, and statistics!
> Linux has about 0.35% of the
> desktop market, not 0.1%.
Hehe, that’s a difference of 0.34%!! Not so huge, is it? And how do we know that 0.1% is less correct than 0.35% anyways? I have heared figures, from ‘reputable’ places, of up 4.3%, and even 5.7%. It all depends on who is quoting stats and for what purpose.
It sounds like no one has picked up the phone and talked to Gassee and asked him what it would take to pry BeOS out of Palm’s death grip.
actually, this has already been attempted by BeUnited I believe. Someone with more knowledge on the financials requested, and the letter from Palm please help me out on this.
Michael, As soon as it was announced that Palm was getting Be, several groups, most notably BeUnited got together to try and get it in the open somehow. They were told to buzz off unless that could come up with several million dollars (10mill sounds about right, IIRC)
So, if you have a few million laying around feel free to contact BeUnited, im sure they would be more than happy to hear from you.
HaveFun!
WhiteRabbit
should’ve read the fine print! no
allowed! sorry about that.
While it is true that OpenOffice.org is not yet ready to compete with Microsoft Office directly, what about all of those normal folks who cannot afford to spend over $500 for an office suite when they might utilize 10% of its functionality.
I think, like Appleworks for MacOS, OpenOffice.org could provide a nice, INEXPENSIVE (free or Sun’s StarOffice), basic office suite for OpenBeOS if it were ported. Your basic user only does a few things on a computer: Games, email, web browsing, basic word processing, media playing, and maybe finances. If the OS can be released with a good browser, decent word processor (or office suite), and a good media player (music and video), it could be a main OS for many. BUT, as has been said, those apps alone don’t give Windows users a real reason to switch. On the other hand, without those apps, users will have reasons NOT to switch or even try it.
A popular home use of computers is gaming. Without appropriate drivers, and OpenGL or DirectX (through wrappers, etc), it will be tough to get companies to develop games for the OS. If you have to reboot into Windows to do too many things, use of the OS will fall by the wayside.
One last bit of rambling… For acceptance and userbase growth, any new OS must solve the old chicken and egg argument: The OS must have good apps available to be adopted by a significant user base. There must be a user base for many companies to spend time making or porting good apps.
>Hehe, that’s a difference of 0.34%!! Not so huge, is it?
It actually is. The fact that is 0.35% does not mean that it is close to 0.1%. That is 3,5 times bigger. So, if let’s say, the 0.1% is 2 million users, the 3.5% is 7 million users.
> I have heared figures, from ‘reputable’ places, of up 4.3%, and even 5.7%
No, these are definately not right. Linux as a desktop is less than 0.4%. My stats are pretty much the same from 3 different *big* sources.
“I’m fluent in a dozen of language”
is english one of them?
Eugenia, 0.35% may be 3.5 times bigger than 0.1% but remember there is such a thing as a margin of error. 0.35% is still WAY too small to be considered a reliable number. Especially when you realize people may have multiple copies of linux or maybe they installed it and didn’t like it after a few days.
I loved BeOS, and I’m excited about the prospects for BeOS clones. However, as the author pointed out, without a compelling reason to switch, no one will. This project is years away from reaching any critical mass, if it is possible to do so at all. Remember that in a few years OS’s are going to be in a different state than they are today. Can OBOS take into account changes required to have a modern OS by a few year’s standards? I wish them well, but I believe the project will always be in the realm of a hobbiest OS.
Apps
That was the whole idea of OBOS project – there are enough apps for BeOS, re-create API and most of them should work.
You don’t need OpenOffice – GobeProductive kicks ass out it.
All the media functionality never came to any sort of fruition as there was no focus on providing good ISV support
Apparently, you discovered BeOS after “focus shift”.
There was another “focus shift” even before that, although it wasn’t called this way – it was called “IPO”.
After IPO Be Inc. drastically reduced amount of information available to general public.
And the “focus shift” finished what was left of ISVs.
OBOS project is trying to get new BeOS under old terms –
use it, enjoy it , improve it.
Palm’s death grip
I don’t know why people so angry about Be for not releasing BeOS sources to public. It’s their baby, they can kill it if they want. They choose it to keep it for themselves.
Wait for new Palm OS. And moreover for the roadmap that should be announced after.
I only wish that Palm would continue to develop Dano in the future as well.
Long Live BeOS.
ciao
yc
I am not a coder and I really wanna help OpenBeOS whenever and however I can, but heaving read this I feal lik ethis is going to be another waste of my time (after BeOS).
If you’re not writing the code, you can’t decide what to do with it, it’s as simple as that.
However, with the MIT Licence being compatible to the GPL, no one’s holding you back from rereleasing OBOS under the GPL if that makes you feel better.
but I’m quite sure that OpenBeOS will fail, eventually.
There are two possibilities. The first is that progress will simply be too slow, and developers who were once gung-ho about the project will begin to withdraw, until it stands as nothing more than a reminder of wasted efforts past.
The other possibility is that OpenBeOS will be a success for a while–they will reach the point where they have copied everything that was in R5. At that point, the project will spiral out of control. Writing a technically advanced OS is much easier than writing one with a slick, easy to use interface. Once they no longer can copy directly the example set by BeOS, OpenBeOS will degenerate into another, slighly higher performing linux-type OS.
Let’s face it: you need a company with real designers and real testers to produce a good interface. OpenBeOS will not have these, so once the source material runs dry, it will falter.
Quote:
“Its creators made the ultimate mistake of base its APIs on C++”
You don’t know much about programming, do you?
Quote:
“i don’t see why coorporations couldn’t build their products (applications or services) on GPLed BeOS replacment.
The Best 3D software is being ported to Linux without any problems!”
1 – Because they wouldn’t make money (remember loki?)
2 – What software is that? I dont recall autodesk porting 3dstudio or autocad to linux.
The author’s three main points are good, but he and everyone else in the OpenBeOS camp are missing one point: hardware support. This project is dead without it. And no, “hardware support” doesn’t mean taking a driver that Be Inc. wrote and patching it to support a newer revision. Any “BeOS replacement” movement based on Linux will at least have NVidia and ATI drivers, full P4 support, etc. By the time OpenBeOS is actually functional, the hardware it does support will be five years old!
To Z.B.:
“Why protect corporate interests … what did thay ever done for BeOS?”
Be was a company. It turned its backs on the BeOS world, so all it’s corporate partners left. This left no new apps for the OS, and no user-base (an Internet Appliance user base??? whew! where were they that day?). Now to get the user-base back (assuming it left), OBOS must provide ‘platform protection’ (my own wording
) by giving the OS free, and letting people (oh, I’m sorry) “corporate interests” sling the OS on their own. It may be crummy that those who wrote a bulk of the code may see no $$, but it will create a larger user-base by flooding the desktop market with a free, fast, working OS that can sit in Win or Lin’s lap.
Corporate interests are the only thing that can keep an OS going for very long. Hobbies are hobbies, desktops are different. GoBe, AbiSuite, Wildcard, and many other co’s took a lot of time and effort (which = $$) to make BeOS software, they deserve protection against their platform failing for reasons beyond their control.
And he said “Let there Be.”
2 – What software is that? I dont recall autodesk porting 3dstudio or autocad to linux.
He was probably refering to Maya or Softimage XSI. And BTW, 3D Studio is IMSO far from being the best 3D software out there.
This is always the nasty issue. Assuming we get to OBOS R2 (or whatever the new name’s gonna be), I would suggest a driver-compatability layer for Linux/BSD drivers. Keep in mind, I’m no C++ coder and my knowledge of the Be API is rather limited, but since they are technically similar (does Posix mean anything here?) I would think a Lin2Be driver adaptation recompiler?/emulator? would be worth diving into. Keep in mind, I didn’t mention Win, since even God doesn’t seem to have the API.
About the GPL licence, this is old news but may be of interest:
http://www.thelinuxreview.com/entry.lxp?lxpe=105
I’m not sure if it’s a media bias or not, but it seems that most news stories and postings I read about Stallman peg him as a radical communist. Is this true? For the most part I agree that the GPL is far too constricting for most businesses to even take a look at it.
First things first: talking about Linux is very complicated because:
a) there’s those GNU things and the impractical (but fair) GNU/Linux discussion;
b) KDE (and Gnome, AFAIK) are not Linux, and it matters a lot here (GUIs against Windows); these DEs are supposed to be Unix-oriented and not Linux-oriented;
c) everybody mocks Linuxers; “Linux zealots”, OTOH, usually like (and learn with) other Unixes, including but not limited to, *BSD, Solaris, *IX, *UX etc. — the name zealot should be reserved to fanatics who want to exclude the competition.
That said, albeit I’ll continue to be a Linux user, I sincerely wish OBOS success. I never ran BeOS, but 99% of those who used talks wonders about it — so who am I to disagree?
Linux itself evolves steadily, and I believe it will be a good desktop alternative sooner or later.
Nevertheless, if OBOS can become as good as BeOS at multimedia, I’d find it excellent.
But don’t make it closed, if I’m allowed to comment on this… and I’m not being idealist: closed desktops can’t beat Microsoft Windows (in acceptance, that is). Maybe OSNews is a site about technical aspects, but marketing or public acceptance do play a role in an OS survival, as evidenced in the BeOS case.
What would you do that Mr. Gassé (?sp) and co. didn’t try? Don’t suppose they were naïve: that would be your first mistake.
HTH.
It seems like quite a while ago that OBOS set out to choose a new name
for the project. I would think that they’d have chosen by now.
Regarding the recent newsletter article about the choice of going with
MIT rather than GPL… I think one of the main reasons why GNU/Linux
enjoys so much success is that the really hardcore contributing devs
know that their work is garaunteed to remain Free. To these people,
working on OBOS is not an option. That’s the appeal of the GPL I see
anyway.
Yes, going GPL may alienate some company out there looking to use
OBOS as a base for something like a synthesizer OS… then again, said
company could always hire more programmers and do it themselves
if they’re really that worried about having to provide source code
(though you’d think they’d be making their money on the hardware).
Prediction: go GPL/LGPL, watch the GPL/GNU/Linux-on-the-desktop
programmers (who are looking for a change) rush in.
stew wrote:
> However, with the MIT Licence being compatible to the GPL, no one’s
> holding you back from rereleasing OBOS under the GPL if that makes
> you feel better.
Huh? Is this correct? I was under the impression that only the copyright
holder could re-release the code under a new license.
Regardless, this would constitute a fork, correct? I don’t think a fork
is the answer.
my 2 cents.
ps. careful what you read about what folks have to say about RMS —
he seems to be a popular target these days.
There’s one line in the editorial that confuses me:
>> OpenBeOS = Mac + UNIX + Performance + x86 + open source + 0$ <<
Unix???? The goal of Be for BeOS was not to make BeOS a “desktop unix”. Some unin compatibility (POSIX) is OK, but I don’t want it to bekome just another unix. Blue Eyed OS is going this way by making Linux BeOS-compatible.
I stand informed and corrected about getting BeOS out from the death grip of Palm. However, it does sound like no attempt has been made since the economy crashed. Things likely have changed.
It’s great to see all the thinking and effort that goes into the Hobby OS realm. It at least is doing what our sad educational system is not doing — teaching people how to write real code.
There is room for so much innovation in the desktop OS market it pains me to see people recreating the past.
And it pains me to see the idiots at Palm sit on a depreciating asset so long and do nothing with it.
If I were passionate about BeOS, I’d find a way to buy it from Palm and then start a Linux-like effort to evolve it. I would not spend 2-3 years cloning it.
#m
This guy’s like “Build a faster, cheaper, and more stable OS, and the Windows users will come running!!”
Yeah, whatever .. take another hit off that crack you’re smoking, buddy.
Truth is, unless the OpenBeOS team realizes that nobody besides hardcore techies gives two shits about the OS, this project is going to enjoy about the same desktop marketshare as Linux currently does.
When you start having to defend your apps by saying “Yeah, but most users don’t need those features”, you know you’re in trouble.
It’s all about apps, games, and hardware support. At the end of the day, nothing else matters, and certainly not the OS. In fact, I’d go as far as to say that the OS itself is rather insignificant. Why else do you think people put up with Win9x and rebooting 3 times a day or more, instead of switching to Mac or Linux? I did it, and so did everybody else I knew. Why? Because that’s where all the ‘good shit’ was
John is a fictional character in this post
– John can work on any project he likes to in his spare time.
– if John likes jazz, fine, let him listen to it. Its his choice. You cannot host polls asking people to vote on which music John should listen to.
– if John wants to use a pencil instead of typewriter for a personal project, its his choice. So what if you cannot read his handwriting.
– if John likes Casper the Ghost more than the Simpsons, fine,thats his choice. So what if he belongs to 0.01% of the population.
– if John wants to write MIT licenced OBOS code during his free time, thats fine too.
Thats what the OBOS group are – a group of people who enjoy similar things, and are working together to recreate what used to be a masterpiece. They enjoy working on it. Let them be.
Potentially (a thought which is at the back of everyones mind), OBOS might become a viable desktop OS, good enough to be a users primary OS. OBOS might become attractive to big business then, which might bring newer drivers and new features. An old masterpiece might then be used as a base of something new and something great. But this hypothetical situation is far in the future, and today, John just wants to have fun and work on a hobby project. You know, hobby, just for fun, no $$$ or world domination plans. Some people collect stamps. Some people assembly plastic model aeroplanes. And John likes recreating an OS.
I’ve been running Lycoris Desktop/LX as a desktop for a few months, and it’s nice, for what it is.
I still run Windows 2000 on two machines at home, and run Linux on two servers. It all works really well.
But, but but. I still miss BeOS, and it’s really a shame my newer hardware won’t run it (I’m still looking for an older donor box).
So what’s the attraction? Explaining it to someone who has never used it is difficult, but here’s my best shot:
My father owns a japanese import. He likes it; it runs really well, it’s very reliable (it’s a 1993 and still going strong with minimal maintenance) and it does the job.
The last time I visited him, he waxed poetic about his old 1984 VW Jetta. He LOVED that car; so did I (I bought it from him and kept it until 1997.) It did the job, too, but it did so with an efficient flair that his japanese import lacks.
You can call it engineering, or production value, or whatever else makes sense to you — the switches felt expensive; the steering communicated everything about the road; it did exactly what you asked of it, eagerly. It was just plain fun to drive, even when you weren’t driving fast, hard or far. It was a great car.
I still miss that car, too. I don’t miss the car I replaced it with, and my current car (japanese import) makes me want my Jetta back.
THAT’S what BeOS is. On paper, you can’t explain why you like it so much, or why you miss it when it’s gone…but once you use it, you love it, and you think everyone else will enjoy it just as much as you.
Provided it supports your hardware.
Many of you seem to think that BeOS was some great operating system, however, it was FAR from it. Sure, it had/has a great, well designed kernel/architecture, but does not these things a good OS make.
BeOS lacked any bit of polish, such as a good printing system, a good game kit, etc…. All these things are what make an OS great.
In fact, BeOS had a lot of mediocre to just plain crap componants (like the networking stack).
BeOS was still very much in development, and should have been on version 1.0, not 5.0 (working on 6).
Isn’t it time for a new dream?
BeOS was great at the time.
Today BeOS is locked up and pretty much given up for dead because Palm is poorly managed.
All the BeOS clones are a waste of time and energy beyond people diddling around.
Potentially anything can happen.
But it rarely does.
Instead of fantasizing about some rebirth of an old OS, why not create something cool for the future?
http://www.protocopy.com/osgui.html
There are people who believe in innovation and defining a new state of the art, not just replicating Windows.
Find them and team up with them.
#m
Daniel:
I’m very much in support of OpenBeOS, however, if they had/do opt to go the Linux route, it will lose any support from me.
Why? Because the Linux kernel is an archaic piece of crap, mantained by a group of people that do not seem to realize any sort of modern day computing concept.
Using Linux just for driver support is flat out a waste of time and effort.
I’ll second what CPUGuy pointed out.
Beyond a high performance threaded kernel, a good file system, and a high performance windowing system, Be was still very much a stripped down hacker’s OS. It didn’t even recognize the second processor on my standard chipset 440BX motherboard. There were probably a few other good bits, but not many. BeOS was an in progress OS to be sure.
BeOS could have been so much more had Gassee opened it up for public development.
But he didn’t and history shows us what happened.
#m
<soapbox>
LOKI? Loki?
Loki did not die because there was no market. Loki died for the same reason Amiga and OS/2 died. The same reason that kills off almost every IT project or company.
Mismanagment
This is the greatest threat to BeOS. Not Zealots not competition, but some executive dork that decides that the project should fold.
This is what Be fans should focus on. Stopping executive dorks from ruining the system. The dorks who say its not yet ready for release. Who slow down development by not sharing critical information. And escpecially those who say I want it all for themselves.
Eliminate the dorks and BeOS will survive.
</soapbox>
Did ANY of you ever Pay for the software OpenBeos is developing? dont get me wrong dont mean to offend but WTF is your Problem? Somebody with a good heart tries to do something better to the community (probably the only one that bothered….all you others just critisize lay , lay back and then wait for something else that happenes….I mean Damn this guy probably spent all his lifesavings in making of OpenBeos and he tries to make a dream come true (don give me that oh but it IS just a dream thats so stupid) THIS Guy actually makes something for YOU…
This is so low……Instead you should be cheering and perhaps i dont know if you had some spare money…….donate em……
I dont care if im going to get flamed after this because everybody knows its true and thats all i need
One thing I noticed is the people who seem to mouth off the loudest against how the OBOS is being run also say that they are not programmers! In other words all talk, no action. Another is that they seem to not note that OBOS has been running less than a year and probably will have a working GUI before it is a year old. Look at Linux, it took far longer to get to the same point of development that OBOS has already reached. That is because the OBOS team has a clear and fixed goal to aim for instead of fighting each other on what to do next. I am sure Linus misses the early years when only was delevoping Linux or he knew all the developers involved. Also for some strange reason some people insist that once R1 is finished all the developers are going to quit? Why? I am sure that there are many ideas on improving OBOS to follow, look at OpenTracker and the changes already made there compared to the original desktop. G.E. while not well organized already is looking into improvements to be made, just going thru it’s mailing list is sure to give the developers additional ideas. As for the other OSes out there, the two main ones Windows XXXX and Linux don’t really seem to be delevoping/advancing themselves that much. Windows changes are more to insure the flow of money to Microsoft, what real new good ideas have they presented for desktop users that makes thier OS better to work with? Linux is still cleaning out bugs and tweaking the code for performance, but again what real improvement to the desktop user are they offering? Linux is also fragmentted, no one improvement is shipped with all versions so developers must leave out emhancements that will limit which users can use thier programs. Atleast all the BeOS clones are trying to reach agreements thru BeUnited, this means many improvements will be shared by all versions and/or a min. common standard is defined to make it easyier of a developer to write a program that will compile/run on all BeOS clones. If Microsoft does get too greedy in the future (I said if!) a lot of people will look at other OSes and at the rate it is going BeOS clones will still be easyier to use. As for drivers, yes they are needed, but anyone checking BeBits will realize most drivers are not patch code but original programs, people like me do with drivers for BeOS right now, and if we can get the specs of the hardware we will be writting them in the future too.
Earl, one thing that I think is GREAT from a UI standpoint in the upcoming version of Windows (codenamded Longhorn) is the database filesystem… yes it was in BeOS (sort of), but they didn’t make any real use of it.
With Longhorn you will no longer need to know things like directory structures, rather, it will be groups, you have a movie on the database, it shows it up in your movies group, a piece of music, in the music group, etc….
Longhorn will be using Yukon (the next version of MS SQL Server) as it’s filesystem, which stores info as XML metadata.
CPUGuy, what are you talking about not used? Email uses the database, People uses the database, MIME uses the database. Basicly you could not use half of the non-game software on BeOS without the attribues and thier indexing.
I for one am ecstatic about the OpenBeOS project. I was one of the lucky few that experienced BeOS. Yes, it was rough around the edges, but it had so many things going for it and it was FUN to use. It sounds to me like the OpenBeOS project is addressing the shortcommings of BeOS, while duplicating all of the things that made BeOS such a pleasure to use. I can’t wait for it. From what I’ve read, printing is better, networking is better, the kernel is better, OpenTracker is better, and the FS is faster. I was sad when Be lost its focus on the OS, because they never completed it and made it fully useful. I am hoping that OpenBeOS will be what BeOS almost was.
I don’t feel like MS and Apple have widened the gap with their “technological innovation”. Sure the interfaces are lickable now, but I never could understand why anyone would want to lick a monitor, or want to get hungry staring at pieces of candy all day
. Not only that, but the dang things are slow and bloated. I long for the days of BeOS’s speed and tiny footprint.
If I had an ounce of programming talent, I would love to contribute to OpenBeOS. But alas, I don’t. So, I’ll just have to settle for being a grateful end user when it arrives.
If any OpenBeOS programmers are listening – I know it wasn’t in BeOS 5.x, but could you please implement multi-user functionality? If that had been present, BeOS would have been more useful. Be never got around to implementing all of the pieces.
And to the naysayers: Yeah, BeOS is a little older technology now, but when comparing ages, look at *BSD, Linux, Windows, and even OS X. They all have a great deal of legacy design or code to them that heavily predates BeOS. If age/legacy were the deciding factor in an OS’s success, most of the above would have been dead long ago. Linux took a while to ramp up itself, so give OpenBeOS a chance.
Whenever a comment is made that is not soothing, benign, and politically correct the thought police come out.
I believe the header says “Post a new comment”. It does not say “Share your undying devotion and love” nor does it say “You must state your support of this topic”.
While there is any privilege left that resmembles freedom of speech, I intend to use it.
I don’t care if you don’t like Borland products, Intuit products, the Microsoft Office collaboration model, all things I have worked on.
It is the mark of a great developer to listen and not take things personally.
For the success of a development team, it is important to weed out all the insecure developers because they make life a living hell for technical support, marketing, quality assurance, other developers, etc. An insecure developer takes things personally and attacks the person trying to communicate with them. They are not someone you want to work with.
#m
Even if OBOS is forever limited to “hobby-OS” status, I’m confident it will be the most interesting and fun Hobbie-OS available. Hoorah to the OBOS team and it’s supporters! The potential of a functional OS that is neither Linux or Windows remains an unknown, since there are none (except Mac, of course – but that’s not an option with the hardware that most people own).
Regarding the editorial, I believe that maketing truely is the key – to any new product. Word-of-mouth can work too, but it can also backfire since it can’t be controlled like advertising or promotions. Word-of-mouth might also take years and years to spread to average home computer users. Not to mention that it can easily be undone by negative “buzz.”
“A vast majority of Windows users would not remain with Windows if they had another reasonable choice.”
That’s one point of the column that I don’t agree with. It should read, “The Vast Majority of Users Will Run Whatever Came with Their Computer, No Matter What.”
And no matter what, I will continue to use BeOS alongside Windows, Linux, etc.. Special thanks to all of the programmers who ocassionally release software to BeBits. Continued growth and use of BeOS PE is a forth point that could have been added to the article. Without it, the release of OBOS would be met with, “Be-What?”
Sincerely,
Bob
I appreciate everyone’s point of view. Thanks for participating in commenting about this article. Eugenia is correct about the 0.1% market share should be 0.34%. Wow, it’s really BIG isn’t it 🙂 Some of you have great points about the OS not being significant, which I agree, and make a point of in the article. Technology is not a solution, it’s just a means to an end.
Other comments I find refreshing are about needing apps, games, and hardware support. Clearly this is the case. Making OpenBeOS able to run X apps is on the way, making it run Windows apps may come some day, making it be the best game platform may be a great new market, BeOS on PS2 🙂
Hardware support it not hard if you think of the level of hardware support Linux has and that much of the driver code can be ported very easily to OpenBeOS.
The article was a call to arms for the OpenBeOS team, to rally around a ‘dream’ as some of you call it (I call it a vision). We want to work with every OS, and especially Linux, and I hope that being open source provides a few converts in the Linux camp to give OpenBeOS a try, and maybe participate in its development… 🙂
Thanks again.
Simon
With Longhorn you will no longer need to know things like directory structures, rather, it will be groups, you have a movie on the database, it shows it up in your movies group, a piece of music, in the music group, etc….
Longhorn will be using Yukon (the next version of MS SQL Server) as it’s filesystem, which stores info as XML metadata.
In theory, the idea of not seeing the file system works well…in theory. When you isolate things down to a small subset of total things on your disk, it sounds nice to have a specific grouping for Music, Documents, Pictures et cetera. This is actually how things are arranged now on Windows and MacOS. Under the database system it would be more rigorously enforced. However the idea that I have to fit every conceivable file or project into a database hierarchy sounds like fitting a square peg into a round hole. There are many examples where this abstraction will not work, and will in fact make things harder to retrieve. If Longhorn doesn’t allow direct access to the file system, I think this would be a step backward, not forward.
By using a SQL server as a file system (something WebFlow pioneered in 1997 and I am sure others did as well), it enables dynamic categorization of “files” without having to know about the underlying positions of the files.
Instead of building a folder based on location and having a location-based tree represent the categorization metadata, you can use SQL to create dynamic views of your data items/objects/files/whatever.
For instance, I can tell the file system:
“Show me all the Word files I have worked on in the last two weeks” and save this as a “directory”. The user does not need to know what drives these files are on, which directories, etc. The database handles it. I can then attach arbitrary metadata and do queries based on this metadata.
Beyond these simple examples, having transactions, commits, rollbacks, views, fancy queries, metadata updates, etc., all built-in to the file system will make application programming much more interesting.
Microsoft is going to radically reinvent applications programming with Longhorn.
Personally, I will be glad to have the “file system” be based on a modern database engine.
#m
After reading the comments, I’m not quite sure what to say. I’m still excited about OBOS, though.
Many people seem to be overlooking a few basic things. Most importantly, OBOS has an entirely different heritage from Linux to build upon, and a wide open future that Apple and Microsoft can’t even begin to imagine.
BeOS, of course, was a proprietary OS like Windows, not like Linux. Except they had some good design ideas and a fairly well-publicized API, both of which have been a great help to the OBOS people. If anything, it looks like R1 of OBOS will already be better than BeOS.
Much of Linux is GPL already, so why should OBOS also be GPL? OBOS is on a different path, a different target than Linux. And using the MIT license creates so many different possibilities for OBOS, perhaps too many possibilities.
But I think OBOS is in a good position to benefit from the best of both open source and proprietary software; Gobe Productive and Mozilla are good examples of each type.
Of course there will be difficulties for the future of OBOS, some problems are perhaps inevitable, but in this case, dealing with those problems actually seems worthwhile. It’s not a matter of trying to improve a serviceable OS, but rather a case of taking a great OS and finding the best uses for it.
All IMHO, of course…
Let the fun begin!!
Remember, with a database as your file system, there will be no “files” or “folders” as we know them today. I think “Yukon” will change computing forever…and I’m not sure it’s for the better.
As for OpenBeOS, I couldn’t be more excited. I wish I could code for them, instead, I’ll wait patiently for R1.
Let’s all make a pact to forget about JLG and forget about Palm. Let R1 come out and then let’s talk about OBOS as its own entity, not a continuation of the BeOS. If we think of OpenBeOS as replacement parts for the BeOS, which it currently is (right?), we are limiting it and our own imaginations. This thing, in my view, is a brand new OS that happens to be VERY based on an older existing product. Cheers and Good Luck to the OBOS team.
The guys at NeXT did the right thing with the Objective-C choice, and now Mac programmer are happy using cool, dynamic object-oriented frameworks inside Mac OS X.
Contrary to popular belief, most OS X applications, even iTunes, iMovie, AppleWorks, etc. AREN’T written in Obj. C. And also, C++ is used much much more than Obj. C, so back then, it was a pretty smart choice to pick it over Obj. C.
Riiiiiiiggggghhhhhht. As I’ve said before, I would love to see a functioning OpenBeOS. But “right around the corner”? They don’t even have a basic system yet. Face it, folks; there’s still a very long way to go. It took Be, with more than 50 full-time engineers and some (limited) help from hardware companies, years to get the BeOS out the door.
Believe it or not, he is right. Be took much longer making the OS simply because OBOS is cloning an OS, while Be was making the OS. They don’t have to go into much design debates, nor try out this design and that design before deciding etc. Look at where they are now, the kernel is inching closer and closer to being a fully featured clone, BFS is already in alpha level (and have finish cloning all the features in the original BFS, though it hasn’t gone through stress and bug testing) and so on. In other words, Simon is correct.
Be did not fail due to it operating system … it failed as a company due to a relentless fight against a Huge Mega Micro Sloth ass sitting its gigantic butt on the competition.
But it was rather, “Be tried to market via over zealous users”. There wasn’t any compelling reasons for majority of Windows users back then to switch. But Be Inc. blames it on Microsoft. Heck, they don’t even have proper OEM deals…. In other words Be is the most pathetic anti trust claim ever made against Microsoft.
OpenOffice is a pile of junk compared with Microsoft Office. It is a forgery that is much worse than the original painting.
Agreed! However, OpenOffice.org is getting more and more features, if someone does a proper port to BeOS/OBOS which doesn’t use stuff like UNO etc. while having a standard interface, it may as well be a Office killer, no?
Where is Be today? Trapped in some sort of Palm business unit?
Actually, Be had been liquidfied, while prior to this all of its IP assets where sold to Palm.
I am so disappointed when reading that main people coordinating BeOS rescue are so blind to FREE software principles and still count on corporate sector.
Because OBOS and BeUnited knows it can never ever make it in marketing their products.
I don’t see why coorporations couldn’t build their products (applications or services) on GPLed BeOS replacment.
They are talking of things using OBOS code, not running on OBOS. That means, just say NVidia wants to make a really good OpenGL accelerated driver, it just takes the current generic driver and adds to it, without releasing the source code. (it could do the same on Linux because XFree was under the same license, though the kernel parts couldn’t be closed source).
EUGENIA PLEASE PUT A POLL within BeOS community What do thay prefer GNU/GPL or MIT or Mozilla licence (which is maybe best one)!
I think the people who should choose the license would be the developers itself. If you find this disturbing, find a few other developers, fork it and license it under the GPL. If I was a developer behind OBOS, I wouldn’t want some bunch of people to pick which license they want for MY software (and of course, I would certainly pick MIT X license).
what about all of those normal folks who cannot afford to spend over $500 for an office suite when they might utilize 10% of its functionality.
They could use MS Works, or even WP Office.
. Besides, everyone uses a different 10% of the features, a lot of people already have Office and don’t need something else unless it does it better. Currently, OpenOffice.org has less features, an uglier interface (come on, who’s idea was to use Times as the default font for the menus etc.), and having to be reeducated, and also not being able to share documents with MS Office users flawlessly. Nothing.
The OS must have good apps available to be adopted by a significant user base. There must be a user base for many companies to spend time making or porting good apps.
This is my own idea, having a new API for the OS
– A transitional API, for Win32 developers. That means they can develop on this without doing a massive rewrite, and also faster ports. Make this API also available on Windows, and make it real good (by having a really good IDE behind it for example) so developers wouldn’t feel burdened to use it. So in other words, developers can have one codebase for two OS.
Just my ideas.
You don’t need OpenOffice – GobeProductive kicks ass out it.
Though most likely Productive wouldn’t be available to OBOS. You need to recompile Productive 2.x for OBOS, something Gobe probably won’t do, plus that 3.x and perhaps in the future 4.x wouldn’t be available for BeOS users, well…. :-p.
The other possibility is that OpenBeOS will be a success for a while–they will reach the point where they have copied everything that was in R5. At that point, the project will spiral out of control.
Actually no. After R5, the developers would improve on their parts of the OS. Things to come out in R2 is like being multiuser, a more modern API (e.g. more classes), BONE, OpenGL support, backbuffering, a new UI look, etc.
1 – Because they wouldn’t make money (remember loki?)
2 – What software is that? I dont recall autodesk porting 3dstudio or autocad to linux.
Try Maya :p. Autodesk’s apps is mostly 2D, except for 3d studio, but that isn’t the most use app out there.
By the time OpenBeOS is actually functional, the hardware it does support will be five years old!
Actually, they plan to have driver support just like how Linux did it.
Linux itself evolves steadily, and I believe it will be a good desktop alternative sooner or later.
I believe that Linux would make it at the desktop when every project out there advocates standards (for example, an app for SUSE would run on Red Hat, a KDE app would look like a GNOME app on GNOME).
However, it does sound like no attempt has been made since the economy crashed.
The economy back then was much more worse than now, so the price would probably be jacked up.
Earl, one thing that I think is GREAT from a UI standpoint in the upcoming version of Windows (codenamded Longhorn) is the database filesystem…
And to think it would be release before R1 of OBOS, OBOS would look like an OS made for that old Pentium you have.
Making OpenBeOS able to run X apps is on the way, making it run Windows apps may come some day
Unless Microsoft happens to allow such a thing to happen, no you wouldn’t be able to run Windows apps on OBOS. See Wine.
I use BeOS as my primary OS at home. The only reason I boot windows is for Quicken, and the only reason I boot linux is for checking out Cosmoe and for Java (my bread and butter). I am not an active developer with OBOS, but do, occasionally contribute comments/ideas to OBOS and GE. I am a member of BeUnited.org have been working on the Mozilla port (when time permits) for over a year.
I respect everyone elses opinion here. One of the reasons I still use BeOS is NOT becuase of its technology. Its more because of its community of followers, and, that I can do almost everyting that I need to get done on a computer under BeOS.
When I go to most linux-type news sites, and read the article comments, I usually leave the site, as it does not seem to be a place where I would like to converse. Most places where BeOS people get together are really rather enjoyable. OSNews was like that once, but, more and more I find myself leaving and not commenting, as discussions here are becoming more and more rediculous. So, before I get OT too … end of rant ….
“Personally, I will be glad to have the “file system” be based on a modern database engine. ”
Im the biggest fan on putting most transactional data into relational dtabases since im a dba. But to call relatioanl databases modern is ridiculous, Oracle came out with a relational database before there was even windows.
The only thing ” revolutionary” about putting a relational engine in a single user desktop is the bloat. This is really the Oracle approach on development. On their non database products everything needs a relational database(i.e Internet Directory) when it really doenst need it. OF course Oracle’s database itself is a top notch product that isnt for lan environments.
BTW a transaction on a single user desktop is not that important if you think about it, at least nothing tha a JFS cant handle faster. With 1 user mysql beats the crap out of sql server and oracle, because it really isnt a transactional db by default. its a flat file db that supports the sql language.( yeah I know about they have the new trasaction table types).
I true sign of ignorance is to not know how old most things really are. Very very little is new in computers except the ignorance of the new comers straight out of .edu. MS probably hires way too many bright eyed high IQ college types who simply don’t know that what they are inventing has usually already been done before but possibly too far advanced for it’s 1st time appearance.
So OBOS is resurrecting a 5yr old OS, so what, it’s still better (give or take some) than most of what is supposed to be new today. As for XP being current, well it looks alot like W2K (almost a great OS) + more_drivers + mindless_eyecandy + License_management + call_home_spyware + Boot_up_lock_in + hidden_updates + Palladium_on_the_horizon + incredible unnecesary complexity to me, so thats why I am so anxious to get off the Windows bus. XP & its future versions might better be called Win1984 or big bro edition. No thanks! Fortunately OBOS does not have or want most of those XP features.
As for whether OBOS takes over the world, I bloody hope not, but 1% or less would be the best scenario. I don’t want my computer illiterate relatives using the same OS as I do, otherwise the OS will get market rengineered for them instead of me “typical but lazy developer power user”. I do hope that OBOS makes development alot more friendly by reencouraging some of the ports that BeOS made so difficult. I don’t mind if Qt, GTK, Wine, wx, winBe, Java, Mono, Smalltalk, Lisp, even X get ported, the more the merrier. If it works well enough, I’ll use it.
So I hope MS continues to drive 97% of the world crazy, stealing their money, their life style info, controlling their choices, what better motivation than that to look for something else. Put enough pressure on people, & some will move.
Ironically I am using BeOS (any OS would do) to recreate a modern version of a microprocessor that died 10 yrs ago, was dreamed up 25yrs ago, by a company very similar to Be that lost all it’s IP to a boring giant, that still uses that IP in ways that would dismay the original architects many of whom still work for said company.
Guess what, Intel, AMD, IBM, Sun, the x Alpha team are all picking on the corpse of that dead chip. The best ideas don’t die, they just take a few iterations before becoming mainstream, the bad stuff like 16bit x86 takes a long time to kill off 1st. So whats in common, pervasive multi threading.
Uh, SQL Server might be based on the relational model which itself is old, but the server itself is very new technology.
With full support for XML, text indexing, metadata indexing, etc., it is a very capable database engine.
On many of the multi-user transactional benchmarks, SQL Server is the fastest database server available.
Only Oracle and IBM have anything that is comparable.
As for a journaling file system, the language semantics for dealing with transactions ain’t pretty… and there is no high level support for doing all the fun stuff you can do with a SQL server 😉
You’ve gotta consider something here. In 2005, you will be running a 6Ghz+ processor and have 1GB+ RAM and a 200GB+ hard drive. Many video cards will offer 512MB video RAM if not more. Standard ethernet will be 1G or 10G. Standard wireless will be 100mbit+.
#m
“You’ve gotta consider something here. In 2005, you will be running a 6Ghz+ processor and have 1GB+ RAM and a 200GB+ hard drive. Many video cards will offer 512MB video RAM if not more. Standard ethernet will be 1G or 10G. Standard wireless will be 100mbit+.”
Michael, can I have some of what ever you got there.
The 1G ram & HD are easy, could do that yesterday.
I don’t think CPUs will get that fast so soon, a little problem called heat gets in the way as well as the really big effort needed to get far below 0.1u that would be needed. Even if the clock does go up, the general performance will not be reflected by CK freq. Maybe too many people reading about the TSMC & Intel work on high GHz stuff in EETImes. A 10GHz ALU today doesn’t make a 6GHz cpu tommorow.
Fact is a simple but well designed cpu (or few) at a few 100MHz that used cheap SRAM as main mem plus a DRAM based HD, would knock your socks off far more than any CPU at x GHz with tiny sub 1M cache, 500cycle rand access DRAM & spinning disk HD. Most CPUs I believe don’t deliver on the myth except on multimedia tuned codecs etc & thats because the streaming data is essentially amplified from say a 56Kbps vid stream to say 200Mpixels/s & thrown away. Real apps that grind that much data die in performance at the fringes of the cpu on the DRAM & HD. This explains why the Intel alien ads show off the streaming codec performance. If they just copied a full zip disk to HD, the advert would be >30sec & would show off how slow the PC really is.
Proper computer design used to be about building a pyramid where each layer was say 10x faster & smaller than the one below, so there would be 5 or so layers, HD at the bottom, ALU speed at the top. The power was measured by the hight of the pyramid.
Today, its about sticking a tall flag pole on a short fat ugly cpu & claiming tallest (fastest) in the world & a small gust of wind knocks the pole over reducing hight (speed) back to nothing.
The Video card stuff, I am completely lost on why any body needs more that 4M. I use 16/32M cards just for very high rez 2d text & surfing, most people use fairly low rez monitors 1280 etc. So the only thing left must be gaming. Uhh, must be kids spending their parents money on $500 gaming cards & Nikes.
As for the networking, I don’t see many people even wiring up there homes for 10/100 let alone 1G.
End of rant
I think they’ll manage a 50% increase in clock rate over the next three years.
What does 2005 look like as far as we know today?
Intel 0.07um process is expected to be released in 2005. The clock speed of Intel’s NetBurst architecture processors is expected to be around 8-10Ghz on Intel’s 0.07 micron process by 2005. The core voltage of these processors is expected to be around 0.85v, and these processors are expected to have around 400M transistors (around 10X greater than the P4).
PC5400 DDR II SDRAM is expected to become available in 2005. PC5400 modules will operate at a clock speed of 667Mhz, giving a memory bandwidth of 5.4Gb/s.
All in all, when we are able to run Microsoft Office inside of Doom V in 2005… life will be good 😉
#m
I’ve been in this semi biz >20yrs as an IC designer, Intel is currently at .13u for P4s, most every body else is a little or much further behind at .15u or mostly .18u. People who are not doing cpus are much further behind usually at .25u. The .13u game is definitely a rich mans club with literally 2 or 3 members.
0.1u is the next target & that is a few yrs away. There is no fab equipment out there to make anything less than .13u. Only marketing droids can make these wild claims. Now some components in any technology such as serial LVDS IOs can run 2 or 3 times faster than cpu logic speeds so a 1GHz cpu could have 2.5 GHz serial links. Xilinx Virtex Pro FPGAs with the embedded 300MHz ppcs offer serial links upto 2.5GHz for the Router market but delivery is er real soon, ie you can get smaller protos.
Most all the 10GHz stories on Optical networking router guys are all being toned down, the market for extreme speed just went bust so did a couple of these 10GHz IP shops. Most of these guys are now working on what customers are actually buying rather than what they dream about. Since Intel also got into this biz, they will probably tone down their Networking retoric too.
“0.85v, and these processors are expected to have around 400M transistors (around 10X greater than the P4)”
0.85v is scary, the nmos thresholds used to be that & it takes 3 of those to make a safe inverter. This means the device threshold is below .3v which is a very bad place to be for lots of device reasons. Bet the IDS curves look like shite. 400M doesn’t really mean anything anymore, probably mostly sram. You already have 256Dram chips that have equiv to >500M devices each chip.
“PC5400 DDR II SDRAM is expected to become available in 2005. PC5400 modules will operate at a ”
Well they used to say, “it the latency stupid”. These DRAMs may burst blocks of data at faster & faster rates, but their random access times have decreased only about 5x from 250ns to 50ns in 25yrs. So if my single threaded app misses cache at 6GHz, I will be waiting >>300 cycles for data. Most of my mem access are random & bigger than the cache, so I am dead at 6GHz. Bandwidth means nothing unless you are a codec.
Back to OBOS, bet it will have about 1% of the OS market before you see a single x86 Intel cpu that is really 6x faster than a 1GHz P3/Athlon on most useful operation (not codecs).
Life will be sweet when I can measure the speed of my PC & it compares favourably with the marketing fantasy.
” Intel is running 4Ghz P4’s in the lab today”
With OC or not?
I saw pics on some Overclocking site maybe from Computex of a 3.6GHz P4, & a report of a Japanese OCed 4GHz P4. I will accept those as valid, but that means it will take a 2-3yrs before Intel could match that in a safe conservative product that might end up in ordinary users PC.
But it still comes down to massive heat removal. It has always been known by IC engineers that devices can work 2x faster if you can cool the substrate to -70c or so, so you basically jump 2yrs into future if you can handle that kind of cooling & not even the gov does that.
And since when did MS Office need 6GHz for word processing. That implies pretty bad SW design on both it & Windows, I remember using WriteNow on a 25MHz 040 Mac being quite sweet. A lot of the kind of apps that are clamoring for the Hammer with 64bits address could use 6Ghz, but not most 32bit stuff.
My ‘dream come true’ would be a partnership between OBOS & Linux.
I currently run Debian on my server and love it. A little more work than my W2K server, but it just ‘feels’ better. I’ve definately converted. My SOHO server will forver be Linux. Desktops? Windows without issue. I’ve tried Redhat, Mandrake, and several flavors of my beloved Debian. I just still don’t dig Linux on my desktop. BeOS on the other hand I loved on my desktop and ran it quite some time on my laptop & main desktop box. The Beos install was about the easiest, cleanest install I’ve ever done.
I’d love to have a 2 CD set: 1 linux distro, 2 OBOS. Just walk in into a room w/ 2 machine, drop a CD in each and have a tight little server & desktop setup in a heartbeat with no $$$ to Microsoft. Get OpenOffice ported over to OBOS and a good gaming kit and things are looking bright-n-shiny.
Linux (wonderful server) + OBOS (hopefully wonderful desktop) = my dream.
But, it’s still a dream
I think the Cosmoe & B.E.O.S projects look interesting on paper, since they both promise Linux Beos hybrids. If you have to run Linux apps on the Linux kernal then one of those might be something to look at. Long term Cosmoe also promised Carbon API as well. But I haven’t seen much tangible on the latter yet.
Carbon12,
You hit the nail on the head there my friend. Linux as a server and OBOS as the desktop is a PRIME example of what can happen to make small enterprise level wins for the open source community. beunited.org is working on standards to make such communication possible. For small companies that need email, browsing, forms processeing, or any turnkey system the Linux/OBOS combination is a real ROI story. Take for example the following small scenarios:
– in a Library
– Internet Cafe
– Retail Store
– Small K-6 schools
– Doctors office
– any small companie
These examples are were such companies would nomally turn to Windows 2000 Server/Windows XP combination to build a sytem. If an enterprising company were formed to supply solutions for such companies with a Linux/OBOS combination I think they would be successful. The reason is that the price would be unbeatable, the support would be immense, and the technology is easier and sound.
I think that many many many small enterprising young people could start such companies all over the world to promote and make use of such a system… $$$
Don’t wait until OBOS is out, use BeOS PE now. It’s a match made in heaven 🙂
“Looking at the OpenBeOS site, I could not find the answer to what they are aiming for. Simply, what is the goal for what OpenBeOS is going to offer to the user 2-3 years from now? ”
Try the FAQ. They state everything pretty clearly to me. Sorry if someone posted this already, I haven’t had time to read the whole thread.
OpenBeOS is OpenSource & might be free to use, Distro’s will be sold, dudes.
The Pro Version of upcoming ‘Zeta’ has got a DVD writer App for any kind of DVD-Writer as long it is Data.
They got the freshest gcc (3.1x), freshest Python, Perl, Personal Java 2 RE, etc…
Anyway – the most important is: it’s going on with BeOS although Be Inc. itself is recruited to polish the Major PDA OS to bring it ‘up-to-date’…
Sounds good to me – with a working ‘RDesktop-Client’ (also on development) we can talk seriously in ‘biz’ ways, guys….
For those who say that they are only recreating old stuff, check the Glass Elevator mailing list. That’s where they discuss what’s beyond r1.
And GoBe Productive probably won’t have to be recompiled, as they are going to be completely BE-compatible. Corret me if I am wrong.
Yet more people that don’t seem to realize the TRUE limitations of BeOS.
BeOS is NOWHERE even close to ready to be running library, the cafe, etc…. in fact, it’s barely even ready to be run on a single user desktop.
BeOS just can’t do things like file permissions, system policies, heck, it doesn’t even support a multiuser environment yet, all things that you HAVE TO HAVE in order to run an OS in the environments listed.
In short, BeOS is a LONG way from being anything that has been mentioned, yet you seem to think that it can be done right now. Come back to reality for a second here.
The issue with C++ in the implementation of an OS is called the “fragile base class problem”. In C++, it is very difficult to change or add methods or data to a class without breaking binary compatibility because of how C++ stores a class. It’s almost impossible to completely replace one of your base classes without breaking things.
C++ stores classes as structs; functions are regular C functions with their names mangled to support virtual and overloaded functions. Adding code to the struct changes the layout in memory, thus breaking binary compatibility. In addition, when the language changes (or the compiler’s implementation thereof), the name mangling must change, which is why there are such large issues with GCC breaking binary compatibility of C++ programs.
Objective C, by contrast, uses a separate segment in the file to store classes and methods (called messages in Objective C); this implementation, loose typing of the language, and its support for “additions” means that binary compatibility will almost never change.
I don’t think Simon was actually suggesting that OBOS would run the world or the library or the cafe or doctors office DP center. I am not sure the people in those places would want Linux either, depends how it is set up.
I think he was suggesting that since many of these locations are public places that might be tax payer or small business supported that funds used to pay for Windows licenses might be better used for other things when all that is needed is a surfing station. For the special needs of librarians & doctors, I know there must be a ton of special Windows SW for them, but that has zippo to do with surfin boxes.
The last time I looked in a public library in the most select expensive town in MA, the room was full of clunky old p200s running win95 & netscape, atleast the whole thing was on a cable modem (thanks to the cable industry library program).
I don’t think that many libraries run anything more sophisticated than that (maybe Macs instead) for the surfing room. The aspect of files is moot since the floppy disk slot is locked out. The only issue with BeOS or Linux might be the censor ware, I don’t think there is any, & some libraries might be forced to use that for political reasons.
In my local hospital, they did have some special purpose surfin PCs in a wrapped package to hide everything. Guess what the OS was, big clue here, yellow tabs. Actually the BeOS desk top was well hidden, looked like it might have been some eVilla thing repackaged with the browser locked into a local medical site. Same thing could easily be done for other restricted help stations.
If somebody asked me to set up a zero license surfin box, I would highly recommend BeOS with Mozilla. I believe some of the cafes in Japan use BeOS for just that. Linux would be a fine alternate. Both of these have very low PC HW demands too.
*sigh* another guy talking about benchmarks again(TPC or whatever those clowns are called), benchmarks that have no basis in reality like a 32 cluster configurations on rather simplistic data models.
Eweek did a also did a benchmark on just one server config showing mysql beating the hell out of SQL Server with 1000 connections(real world 1000 connections wont happen on both platforms). But Like I said mysql is faster because its
really a flat file db with the sql language which are faster then transactional databases,which you dont need on a single user consumer OS.
Dont get me wrong, I like sql server and administrate it currently.
But,I bet you believe the marketing crap that sql server can handle 30,000 simultaneous connections.
> “I’m fluent in a dozen of language”
> is english one of them?
Computer languages, that is.
For human languages, I’ll always be a moron 🙂
if c++ has a fragile base class problem, then what would be the best language in existence today that caan be used to write a new os from the ground up? what languag would u use? the language should be fast on x86, portable, aallow for complex application, be 100% oo. any ideas?
“if c++ has a fragile base class problem, then what would be the best language in existence today that caan be used to write a new os from the ground up? ”
The eternal problem : you get more speed at the expense of flexibility, or you get more flexibility at the expense of speed.
With the fragile base class of C++ you get direct access to the members, but it’s harder to modify in the futur (there’s few tricks, tho, by adding to all of your class a unused buffer, so in case of futur changes, you can still use this buffer to add members, avoiding binary incompatibility).
With a more flexible approche, like the Objective-C (AFAIK, never used it personnaly), you get all the flexibility to changes your class, but there’s an additional execution layer. Sometime you can’t humanly tell the difference, but sometime it can be a big performance cost.
Both approach are great, always depend what’s your goal. In case of BeOS, the primary goal was a very fast and lite OS to get best performance dealing with multimedia. In that situation you definitely want a “closer-to-the-machine” approach like C, C++ or even assembly.
If you have a BASIC compiler that can generate proper binary format (without external BRUN or the like), then you can make an OS with that language. Just don’t expect the best performances …
All OS:s are great in the sence that they show other OS:es what can be done.
If OpenBeOS turns out great – well…let the linux/windows guys copy interesting stuff in it…
…especiallt multimedia sucks ass in linux (havn’t tried preemtion patch though…) so any future preassure for improvement in these areas can only be good.
SVERIGE!
Why do you want your OS to be a multi-user system so badly? It seems to me to be another buzzword, pet-rock, whatever, that people just believe must be had. Out of every OS/GUI I have used to this day (DOS, Deskmate+DOS, Win+DOS, OS/2, AmigaOS, Linux, Win95/98/ME, WinNT, Win2k, WinXP)… the best user experiences I’ve had are on the single user operating systems. I have access to my own data when and where I want it. Permissions, user profiles, administrator and root accounts… this stuff is just NOT something I care to deal with. I serve no data, I run no office, I need no security internally or externally. All multi-user does for me is complicate access to my own data, and on times when I’m not remembering the architechture correctly while managing accounts, loses my data (how many of you have lost tons of files on the Desktop when you changed your accounts and deleted the “user” folder? Can’t be just me…).
Why? Because we need more than one person to access the computer, because it adds a level of security, there are many reasons, and if you think that just because you don’t need it that nobody else does, then you really need to go have your ego checked.
Theoretically, just about any operating system could do anything that any other OS can do; it’s mainly just a matter of writing the software and drivers for it. In practice, some OS’s are obviously better at certain things than other OS’s.
While I can see the value of multi-user systems in server and business situations, I don’t see any big advantages of multi-user for home users and multimedia situations. I can’t see a compelling reason why OBOS needs to be multi-user, unless OBOS is going to shoot for certain niche markets.
The server market and the business-desktop markets are probably not good targets for OBOS, but if anybody wants to write the necessary software…
😉
Its like all the other features that people want to add to OBOS, most of which I hate like pie menus, curvy rounded Window Tabs yuck.
For the most part I also want a single user, no security system but I wouldn’t mind a multi user security rich option that can be turned off at install or added as a service pack.
Just as long as it isn’t like W2K with thousands of extra things on the HD I can’t get rid of and as a consequence spies & records everything that happens no doubt in the name of bloody security requiring quite a few utilities to be added to counter all that activity.
KISS