Slashdot Mirror
History and Perspective on BeOS
prepp writes "Avid BeOS user Robert Renling posts his first article
about the Be Operating System." An interesting little article, with the amusing conclusion that BeOS isn't dead after all! Ah Zealots. Aren't we fun?
...it's not dead. Obsolete computer systems don't die--they just get severely marginalized...
"It take 9 months to bear a child, no matter how many women you assign to the job."
...and look how wildly they succeeded there. You can look in a thousand video editing suites now and not find one BeBox. They're all using Mac OS X now. Why? Well, aside from the fact that it works better and has a zillion times more software....
As far as I'm concerned, from a purely technical standpoint, BeOS is the BEst Operating System ever. It has absolutely everythign I've ever wanted. The only reason I don't use it is the lack of software. Can I get photoshop for it? How about Winamp? Icq? Aim? Eudora? Most importantly Half-Life: Counterstrike? Some yes some no. Despite all of its outstanding technical greatness BeOS doesn't have all of the software I need.
Windows has absolutely everything, and games.
Linux has everythign I need, or a good equivalent of what I need, and it has tools for developing software.
So I run windows and Mandrake. I would LOVE to run BeOS, it's got everything I've ever wanted. But no software. Sorrow!
The GeekNights podcast is going strong. Listen!
My major clue is that the install process seems to still require the making of a 1.44" boot floppy. That is, if you want to run it by itself, outside of another OS.
To me this speaks volumes about just how old it really is, and probably indicates it is never going to be updated to modern hardware. Also, what makes it relevant in this day and age? Can it do anything another system cannot do better? If the answer is no, or even an extravagantly technical yes (which would never matter to most users), then the world has passed it by.
The impact of BeOS was probably like Ross Perot in the 1992 presidential election. He lost, but got a large enough percentage of the vote to scare the mainstream politicians into sharpening up their act. I think this is arguably one of the factors for the prosperity of the 1990s. If I am correct, we can thank BeOS for encouraging other software makers to improve their quality/performance. Therefore BeOS benefits us even now, but we do not get the benefit from actually using it.
While I do have an affection for obscure operating systems (and the BeOS is certainly now that), the fact that BeOS is obscure is not what makes me admire the damn thing so much.
...now if only I could find a BeBox on the cheap!
As the article says, it was well designed from the beginning, and well thought out through the end. The same can not be said for any other recently modern OS, really, save for maybe OSX (and this requires one to look at OSX as a "new" OS).
Windows certainly doesn't qualify, and even Linux (which I use and love a great deal) was never initially designed or thought out to be the OS it is today. It's been hacked together over the years to add features like the ones that were in the BeOS from the start (not that the hacks haven't been good...they have...but they're still hacks)...In a way, I'm quite disappointed that Be lost out. There's still always the hope that Palm might do something fun with them, but they'll probably just screw it up...
The operating system is TOTALLY irrelevent when it comes to most users. There are only three things that matter: 1) Applications, 2) Hardware support, and 3) Applications. You can have the worst operating system in the world (Windows 3.1) and utterly destroy a clearly superior operating system (OS/2) simply because you win the hardware and application battle.
Be was dead before it started, because the ONLY hope for a new operating system is compatibility with the current application base. What I don't understand is how Be deluded themselves into thinking that application developers are going to spend valuable resources porting to a completely new operating system without any users just because it's "new and cool".
No one cares about operating systems. Say it three times.
Sometimes it's best to just let stupid people be stupid.
If you want to take it a step further, you can say that most, if not all modern operating systems are somewhat "Unix-like" in that they all implement features from Unix libc.
NetPositive seemed rather lame when I used it.
BeOS was/is a slick OS that deserves most of the praise it receives, but what it didn't need was a Linux binary compatability layer or a working implementation of Wine. People who want to run Windows or Lnux apps are already running Windows or Linux. What BeOS needed was some BeOS-only applications that gave the platform a competitive advantage.
BeOS was like a shiny new car, all polished but with nowhere to go.
-- Slashdot: When Public Access TV Says "No"
First of all, I'm not in the naysayers camp about OpenBeOS. If developers want to rewrite BeOS to scratch their own personal itch, far be it from me to tell others how to spend their personal time. However, OpenBeOS is not ready now. Yet for a vendor targeting a vertical market there are so many other available platforms to choose from now, that waiting (or even developing) for OpenBeOS simply doesn't make sense. These guys are in business to make money, not to scratch a personal development itch. If that's the market I wanted to target I would likely choose either Windows or Linux, depending on how I wanted to price my product, and how I wanted to arrange support. I might even go QT and target both platforms. The last thing I would do is hope and pray for OpenBeOS to come along and save my day when the market was there for the plucking and alternatives to OpenBeOS readily available. Just my .02...
Cheers,
--Maynard
However, it had some pretty bad nasties:
Copyright Violation:"theft, piracy"::Anti-Trust Violation:"thermonuclear price terrorism"<-Overly dramatic language.
> Dead or not, BeOS was one of the best operating systems
;-) ...), but that's because its
> I have ever used.
I won't go that far, but certainly Be had some innovations that other
OSes would do well to consider. Even today. No, I'm not talking
about the filesystem.
> If only it had the software/hardware support.
I don't think either was really a problem. It had the stuff that
actually mattered. (Emacs, Mozilla, what else do you need?
It ran fine on my hardware. Now, it has problems with some newer
hardware (USB, 3D acceleration,
development waned and stopped; it was up to approximately current
at the time of the release of R5. At the time, it had better
hardware support in some areas than Linux. (For example, BeOS had
drivers for some software modems before Linux did.) It has rotted
since things fell apart, but that's a symptom, not the problem.
BeOS needed two things. Advertising and OEMs. Oh, and there were
a handful of important missing features, such as the ability to set
colour prefs globally, but the Mac is _still_ missing that one, so
it must not be fatal. Java support was lousy, but there have been
issues with that on the Mac also, as recently as a year ago, so
again, it must not be fatal.
BeOS, like I said, needed two thing: advertising and OEMs. But
instead of trying to sell the system, Be kept trying to sell the
technology (to Apple, to Palm, to embedded markets, to game
developers, and who knows where else that they didn't make public).
I don't know whether they could have successfully sold the system
as a desktop system, but I wish they would have tried a little
harder to do that. AFAIK there was never _one_ TV commercial for
BeOS systems. I know commercials cost money, but look where not
advertising ended them. You have to try something, and the things
they tried didn't work.
> It booted faster than DOS(and I'm not kidding)
Maybe not kidding, but you're exaggerating fiercely. The time DOS
required to boot was dwarfed several orders of magnitude by the
time the BIOS needed to do the POST; to say the same of BeOS would
be a significant hyperbole. It did boot much faster than Windows
or Linux, but as the other poster pointed out, boot time is really
not a big deal to most users.
> It had one of the best browsers I've ever seen
Err, I don't know what you saw in NetPositive. It didn't seem like
a very good browser to me. This really didn't matter though. First,
most users don't care beans about the quality of the browser (hence
the popularity of IE4 in its day, which was nothing to write home
about either), and second, you could download and install Netscape 4
(which at the time was not seeming so ancient; today of course you
can get Mozilla for BeOS).
> and it was very very slim
That really only mattered for dual-boot scenarios. I will say, BeOS
is a multibooter's dream come true. "Plays well with others" could
just about be its official motto. It also had an excellent driver
model, which basically didn't require any changes when hardware was
swapped out -- very user friendly, that. HardDrake is only just now
beginning to approach this. It also had a couple of nice features,
such as having a different res and colour depth for each workspace.
> What they needed is a linux binary emulator
Way more trouble than it would be worth. An X11/GTK+/Qt library
done as a wrapper around the native GUI would have been orders of
magnitude easier to do and gained source compatibility, which would
be plenty good enough. And yeah, I know FreeBSD does it, but OSS
does a lot of things in different ways from how companies do them.
> and a well designed wine-like windows binary emulator
Even harder to do than the Linux binary emulator, because Windows
is more poorly documented (in terms of its internals and ABI).
It would also be more worth doing, but the amount of work involved
could be prohibitive, and performance would probably not be great.
Besides, OS/2 went down this path, and the only reason they didn't
go bankrupt is because IBM has lots of other irons in the fire
besides the OS.
> I stopped using it because it didn't support my NIC, and when i
> sat down to port the driver from BSD i found myself lost in the
> lack of debugging documentation and gave up.
I think Be made a mistake getting out of hardware. They got out
because Apple wasn't cooperating any longer, and they ported to
x86, and as far as it went that was fine, but while offering up
a version that will run on various x86 hardware with an HCL is no
bad thing, I think they still should have sold prebuilt beboxen,
in an x86 variety. And I think they should have marketed them.
Now, I think Palm should come to terms with the realisation that
they aren't going to develop BeOS (unless they _are_ doing so, in
which case great), and get what PR they can out of the deal by
open-sourcing whatever parts of the BeOS source code they have the
rights to. (Obviously there would be some pieces of BeOS that were
sublicensed and could not be released, like there were some bits
of Communicator and StarOffice that couldn't be released with the
rest, but that's a minor complication.)
Cut that out, or I will ship you to Norilsk in a box.
What they needed is a...
What they needed was a market. This seems to be a hard lesson for many technical people to understand.
If they had run so much as a single advertisment, anywhere, any time, perhaps they might have started finding out if there was a market.
This article contains a number of inaccuracies and omissions, which leads one to wonder if the author is not writing with rose-colored glasses firmly in place:
1. "BeOS is fully POSIX compliant" -- not correct; it would be more accurate to say "mostly" rather than fully. I could quote from the Be FAQs on this point, because I wrote the original FAQ (I worked at Be for three years).
2. USB & FireWire support -- the article states that the USB support is not very complete, and shortly thereafter implies that FireWire is supported more fully. It's really more the reverse, though I doubt if the USB code would work with much of the built-in USB hardware being released these days (you never know, though; we got the original stack from Intel). At any rate, if you happen to have a BeOS retail box, you'll see USB listed (along with the Intel credit), and no FireWire (though my most current box is for R4.5, not R5).
3. Design of the kernel -- I can't comment on a technical level, but my recollection of conversations with kernel engineers was more that the kernel was monolithic (and that we thought that was a good thing). The design inspiration was from the XINU operating system ("XINU" is "UNIX" backwards), I'll leave it to operating systems connoisseurs to determine whether that compares with the Hurd or L4, as the author asserts. Perhaps the author is thinking of a new kernel being written for the "not dead yet" OpenBeOS project(s).
In all, the article reminds me altogether too much of the many articles about the Amiga OS that I read while I worked at Be. Sad, but true. I wish those projects luck -- I miss Be and BeOS -- but I consider them wishful thinking. I've moved on to Mac OS X, and don't plan to go back.
Maybe the team now at Palm will change my mind -- I hope so!
Ok...
I tried BeOS as my primary operating system for 3 months (this was a prior release over the final). I returned once again during the final release for a very short period. My roommate was a huge fan of BeOS and he stuck with it til the end. (At which point we thoroughly converted him into a Linux lifer).
While I truely enjoyed all the very technically pleasing asepcts of the operating system... there were problems. I really did admire the file system, threading, memory management and unix style xterm/console... there were problems.
Driver support was absolutely appauling. I know, I know, it did not have an amazingly huge installed user base and these things naturally follow. However, if we are going to stand idly by and watch BeOS be touted as the world's greatest creation, lets just be honest.
Application support was non-existant. There were no decent browsers and a hand full of applications that implemented only the very basic's. I have horrible memories about non-intuitive UI implementation in the provided applications and utitilities.
They really did a great job of implementing the fundamental foundation. However, a great home has more then just a nice starting base.
Its great it recovered from application crashes and would know when to auto-magically kill problems. In fact, with the frequency in which applications died it was quite necessary. At the time, it was admitted that tracker was horrible, but was just until something better was going to be released. I had more the one instance where networking would bomb out and never ever return to a functional state. This was truely fixed in 4.5? I really could go on for quite a while. I wanted to fall in love with this operating system, but its less important features that actually effected the user in some noticleable way just never got addressed.
I'm not sure what the author is smoking, but I could take a guess its the same thing Balmer hits before he goes out to praise MS.
Yes, I'm posting anonymously too. I'm not lavishing any portion of this, these were all my experiences under the very technically elite skeleton BeOS was.