1. I'm an OS junkie. I love installing, fiddling and learning new OSes. I installed BeOS and used it for about 2 years (no joke) back in the day. LOVED IT. I even had it replace my Slackware instance for about 6 months as my primary system.
2. I love seeing OSes still actively developed even if they aren't popular. I think it keeps driving ingenuity and pushes other people. But, I don't get why we have to have rewritten older systems...lets try to advance OSes.
3. The interface has to go. It was clunky back then and now it is downright horrendous compared to nearly any interface we have. Gnome might feel like 2004, but this feels like 1997.
4. The biggest problem with BeOS back in the day was lack of applications. I don't see this being fixed with Haiku unless they have translation layer.
EDIT -
One thing I forgot to mention about BeOS back in the day. It booted in about 15 seconds. It was amazing. Right now the fastest I've seen is an Ubuntu box getting from start to internet in about 7 seconds w/ a SSD etc but for 1999/2000, it was amazing. Astonishing, really.
I would also add in this edit that I think the state of UI on current OSes is a bit disappointing. I think Ubuntu is trying w/ Unity and even that has a ways to go before I see any real innovation in interfaces. It feels like we've got a modality that worked 15 years ago and we've sort of massaged it, but haven't really innovated or pivoted on the concept too well.
It's fun that people are so different. I would give a lot if I could have BeOS Tracker, Deskbar and GUI fitted into, for example, a Linux system instead of Gnome.
BeOS GUI was minimal, slick, and one of the rare GUIs that you can actually use for system maintenance. That's a tipping point for me: Linux desktops GUIs are there because a desktop computer apparently has to have a desktop GUI. In BeOS, it was there so that you could operate your system. Using a shell on Linux is the equivalent of the "Start" button; using a shell on BeOS always felt like a hack: you didn't need it for system control.
I have first gen ( dual 66 ppc ) and second gen ( dual 133 ppc ) beboxes. i ran far more than 4 movies with no hiccups. so much more. the media performance on the BeOS back in the 90s w/ a tiny amount of RAM still blows away any OS I have used.
Ever GUI OS I have used before that or since then has seemed like a bloated slow old pig in comparson.
4. The biggest problem with BeOS back in the day was lack of applications. I don't see this being fixed with Haiku unless they have translation layer.
I don't know... it seems like most of the benefit of Haiku is in the unique APIs. What's the point of using it if you're just using apps from other platforms?
Would you have said the same thing about iOS? That iPhone should've had a translation layer so it can run Windows Mobile apps, because the app store was empty?
At some point, if you're trying to build an innovative platform, you just have to grit your teeth and do what you can to attract native apps.
Sure, I agree with that statement in principle. Though I don't see Haiku capturing a market the size of even WebOS or something like that, it just doesn't seem that there is enough space for it.
The only reason I mentioned translation layer was to get people using it...but it they plan to write some awesome apps for it, all the better.
It really is a catch-22. Look at Ubuntu and Wayland. Getting to Wayland means having to write quite a few apps that are there on X11 stack, but if you just use a pass through or translation layer, what incentive is there to write more apps? So, yeah...you are right, sometimes you just have take it for a bit with few apps if you think you have the audience. Clearly Ubuntu has the audience. Does Haiku?
Widespread adoption doesn't seem to be a goal, but really as long as the POSIX APIs are supported in a reasonable manner (...or you are Windows), it doesn't matter much. It's nice to use, and it gets to take advantage of portable open-source software which gives it a larger community of developers than (say) Plan 9 and Amiga enjoy, so market percentage doesn't matter much.
I've always like the BeOS interface... then again I also really liked CDE. There are plenty of things I would want to change about the Be interface that Haiku implements but, I think that would be iterative not replacement.
Having loved BeOS when the free demo version 5 came out (I was very late to the game), and also being an OS junkie, I was eager to try out Haiku.
The interface certainly looks 1997, but it feels reasonably smooth. It is lightweight and fast, just that one has to click a lot. So, it feels 1997, when GUIs were a ittle more responsive. (I do really wish the WM would do some more things for me, like SloppyFocus) ...and don't ask about the file manager.
As far as boot speed goes, on my laptop it's a couple of seconds. I'd use it more if not for the below-mentioned issues. (Booting from a disk; the live CD is as slow as you might expect.)
The software support is not bad at all. There's a nice browser (webkit-based), bash, and I didn't have trouble getting a couple of mainstream interpreters running. I believe they're using a slightly older gcc, but I had no trouble getting a couple of my personal projects to build (rsync'd from somewhere or copied from my /home, since Haiku can mount ext3), or a few other programs that I wanted. If something is POSIXy, it'll probably run. Ruby and Python worked for me, as well as a small language I'd put together.
The only thing that keeps me from running it as the main OS for my laptop is drivers. The wireless support is still in development, and the version I have doesn't know when I've plugged in my headphones; my laptop is mostly used to play music, ssh to a larger computer, and browse the web. So, without headphones, I would definitely disturb coworkers, and need an ethernet cable to use besides.
Still I recommend trying it out. It's still under development, but it's solid and fast already. (Just need the danged sound/wifi drivers to work...)
3. The interface has to go. It was clunky
back then and now it is downright horrendous
compared to nearly any interface we have.
Gnome might feel like 2004, but this feels
like 1997.
I use windows 7 at work and os x 10.6 at home (as well as ubuntu 10.10 in VirtualBox) and I can't see what feels like 2004 about gnome, or what feels like 1997 about this one.
In fact I can't off the top of my head really think of any major UI improvement between windows xp (2001) & 7 beside the clever resizing/positioning of windows & the new task bar.
> 4. The biggest problem with BeOS back in the day was lack of applications. I don't see this being fixed with Haiku unless they have translation layer.
Keep in mind that beos is basically posix-friendly. You can already compile a lot of applications which work on linux. Firefox needed some changes for example, but works now. Applications like vim compile without any changes at all - just get gcc, make, make install.
The reason that I'm excited about Haiku: a fledgling open-source OS that isn't based on the Linux kernel. It has been awesome to see Linux find its place on the desktop and mobile, yet, the homogeny is rather stifling and it's disappointing to see there's no alternative to maintaining UNIX's legacy (X11, filesystem layout, my personal dislikes). Haiku seems to have the potential to fill that void.
Haiku icons are actually stored in a vector format instead of bitmap. That means all icons are resolution-independent and antialiased. Also, since the format was designed for icons, the filesize is so small that the data will usually fit in the same block as inode data, so the file takes effectively 0 space on the HD!
OS X's Spotlight is a remnant of BeOS. Apple hired Dominic Giampaolo who did most of the design and implementation of BeFS early in the 2000s. Both systems use metadata stored in the file system for indexing, but BeFS was more through and had better query support. There was hope that a filesystem close to BeFS would be the default OS X FS.
The fact that BeOS is single-user is news to me. I have to wonder what the security model looks like without a 'root' concept - this article's dismissal of the problem seems pretty shallow.
Obviously the "Change the World" title is just linkbait. Thanks for renaming.
Plan 9 has no super-user. Each server is responsible for maintaining its own security, usually permitting access only from the console, which is protected by a password. For example, file servers have a unique administrative user called adm, with special privileges that apply only to commands typed at the server’s physical console. These privileges concern the day-to-day maintenance of the server, such as adding new users and configuring disks and networks. The privileges do not include the ability to modify, examine, or change the permissions of any files. If a file is read-protected by a user, only that user may grant access to others.
IIRC BeOS had no security model; basically everything ran as root all the time. You might say that it had no concept of non-root. Remember that Be viewed Mac OS 8.0 and Windows 95 as their competition, not Unix.
I can't see what Haiku offers that modern OS X doesn't. Drag and Drop between applications? Got it. Combining soundcards to use in aggregate? Got it[1]. Object Oriented APIs? Had it since NeXTSTEP (and they're called 'Kits' as well).
The only thing that seemed mildly interesting to me was the routing of audio between apps, but this is mostly achieved these days using ReWire or Soundflower (admittedly not an ideal solution).
It seems like a fun hobby (I'm into OS nostalgia myself) but suggesting that it's going to have any effect on 'the world' is pretty silly.
I use Mac OS X every single day and I would freely admit it is a giant bloated memory sucking pig. My finder was using 100 megs of memory yesterday. Why? Who knows but for me I'd like to have a lean OS with good features rather than one that I need to reinstall every year because it suddenly becomes a sluggish, memory sucking pig like my Mac OS X machines have over the last 5 years.
Haiku isn't a memory pig- I'd like to see that ideal come back in OS design.
And yes, I am transfixed with 'memory sucking pig' as a phrase right now.
While I wouldn't disagree that an OSX install can get slower over several years, I can't help but think what you want is a bit of a case of 'the grass is always greener'. There are plenty of lean linux distributions out there, but the reality is that they cut things out.
Modern operating systems do a lot to improve your experience as a user and provide various services. A lot of tech people seem to want their system to have as much CPU and RAM free as possible, but then what are you going to do with it? Are your apps actually running out of memory and swapping?
Yes I am having a problem w/ 4 gigs of memory, running minimal numbers of applications with swapping and slowness. Every year I reinstall Mac OS X and it gets better for a bit, then it starts to come back and I do some maintenance and it gets somewhat better and repeat til around the 12-15 month mark when I have to reinstall again.
I totally agree. The main reason I play with Haiku is that the UI is really responsive. OSX seems to be optimized for screenshots, looking pretty in a manner that often gets in the way. I get beach-balled a lot on the Macbook I have at work, and most things are sluggish. When it's not loaded down, you still have to wait for animations to change desktops and minimize windows. This is even true of my iPhone (again, issued by work) which, even if it's not overburdened, it does make you wait for animations.
Not to single out Apple; most UIs are like that nowadays, including my personal phone (N900, flashy/slow animations and everything, but tolerable). Responsiveness and speed in GUIs seem to be relegated to semi-obscure X11 window managers. It would be nice if "usable" got a higher priority than "flashy", without neglecting flashy, but it seems to be one or the other for now.
Some people clearly want that, though, so it's hard to blame anyone for selling it to them.
The spotlight interface is still inferior to the searching mechanisms in tracker. The technology underneath is the same technology but generations ahead, but the delivery is poor adn the result is far less useful. Unless Apple rebuilds their GUI this will be permanent because search is bolted onto finder, whereas it's close to the design ideas of tracker.
Tracker is much more rapid to navigate around than any other GUI I've used. I find haiku workspaces far more fluid to use.
Some disadvantages opver OSX is that OSX is really unix, whereas Haiku just presents as it. Most people won't care about this, but if you were porting unix utils yourself you'd probably be aware of it.
BeOS/Haiku use C++ as the primary development language. This will be offputting to some. I'd expect the APIs are easy enough to wrap and that the haiku community will produce first-class python wrappers or similar. They've had them in the past.
> suggesting that it's going to have any effect on 'the
> world' is pretty silly.
There are some cool apps around that were originally developed for haiku, and then moved to other platforms. We might see some more of this.
I can't see what Haiku offers that modern OS X doesn't.
That's not really the point. Haiku's intent is to keep BeOS alive. At the time (over 10 years ago), BeOS was the epitome of a modern OS. The pickings were slimmer back then.
Of course, it took almost that long to make Haiku happen, and the OS world has changed significantly since then.
For a "modern OS" BeOS lacked a lot. Shoddy hardware support at best, a terrible network stack, did it have a printing subsystem? I know they wanted to rebuild the networking stack I don't remember print being good. Internationalization was non-existant.
It was an RTOS with a UI on it, sexy and clean, I'm not a hater but a long long long way from NextStep or OSX.
Replace shoddy with limited and I could agree with that. BeOS died very young.
Pretty much all those points could be made at the inception of any OS. OS X had a nice jump on that because it was NextStep and had been developed for well over a decade.
To be fair, much of the needed work was accomplished in the Dan0 development release that was leaked near BeOS's demise. This included the BONE network stack and some 3D support. Everything about Dan0 was, of course, very much development code, but it was definitely a marked improvement over R5. Haiku was able to use some of the Dan0 high level kits in place until rewrites could be made.
Clearly the Apple/Be Inc. issues were sorted out to Apple's benefit, but it's still neat to see what could have been. OS X had its own issues in its widely released beta, after all. ;)
Having tried Haiku, I like it a lot. I stopped using it, though, because it's really hard to program in it: most languages aren't supported, and others have large math errors.
BeOS was amazing. It simply smoked the competition. It was sleek, functional, good-looking, fun to program for, and had blistering speed (remember it playing 6 different videos each on a side of a rotating cube?).
It was not free software, but with features like that, who cared, right?
Well, its community cared when BeOS was mutated into an "appliance OS" and then later sold off and closed down.
I'm not optimistic about Haiku. As soon as it becomes popular (and I don't see why it wouldn't -- it's aimed squarely at knocking the socks off of end-users), I bet all the core devs get scooped up by some company which will then add proprietary lock-in features and sell it to users.
I used to run BeOS on a Umax c600x280 Mac clone (PowerMax, Umax, Motorola days). I loved it because the UI was responsive. There was no lag. It also introduced me to the command line.
The previous year, I tried an mklinux install. I couldn't tell upside from down. I didn't know enough to "man" a command, let alone which commands to man.
I'm sad Be never became OS X. I wonder what Haiku/Be would be today. Remember the first OS X version? It was slow and buggy. 10.1 was usable.
Maybe Apple would've made a slow, buggy OS B. Back then, it was on.
Tiles answered the main part of your question (Steve Jobs), but to respond to the embedded part, Palm never did anything useful that I know of with their purchase of BeOS, which is unfortunate.
By the time Palm bought Be Inc., they had all but stopped doing anything useful with anything. Other than the Treo smartphone line they bought from Handspring, they only had a few good models from 2002-2005, and since then they've only achieved notability by not going bankrupt and instead getting bought.
i'm sorry, but there's no way that it can become popular with that limitation. security is too much at risk here, didn't we just manage to convince microsoft that computers shouldn't start as root all the time?
also, seeing that it's geared towards multimedia, there's no way that any real multimedia company will adopt the software with no file permissions and user access to all root powers.
It's okay to make things that aren't intended to become popular or to take over their niche commercially. Haiku's makers loved BeOS and are just remaking it. It's very charming and human.
That's not true, Applications could still run as root but be signed and subject to approval as they are now. The only difference I can think of is they just wouldn't be sandboxed any longer.
1. I'm an OS junkie. I love installing, fiddling and learning new OSes. I installed BeOS and used it for about 2 years (no joke) back in the day. LOVED IT. I even had it replace my Slackware instance for about 6 months as my primary system.
2. I love seeing OSes still actively developed even if they aren't popular. I think it keeps driving ingenuity and pushes other people. But, I don't get why we have to have rewritten older systems...lets try to advance OSes.
3. The interface has to go. It was clunky back then and now it is downright horrendous compared to nearly any interface we have. Gnome might feel like 2004, but this feels like 1997.
4. The biggest problem with BeOS back in the day was lack of applications. I don't see this being fixed with Haiku unless they have translation layer.
EDIT - One thing I forgot to mention about BeOS back in the day. It booted in about 15 seconds. It was amazing. Right now the fastest I've seen is an Ubuntu box getting from start to internet in about 7 seconds w/ a SSD etc but for 1999/2000, it was amazing. Astonishing, really.
I would also add in this edit that I think the state of UI on current OSes is a bit disappointing. I think Ubuntu is trying w/ Unity and even that has a ways to go before I see any real innovation in interfaces. It feels like we've got a modality that worked 15 years ago and we've sort of massaged it, but haven't really innovated or pivoted on the concept too well.