Hey! Welcome to my site. You may want to subscribe to my RSS feed. Thanks for visiting!
BeOS and I
A little more than a decade ago, I used to love BeOS. It booted faster, was more stable and reacted much better than Windows 98, which I was using as my main OS at that time. I hoped to see it expand and take over from Windows as the dominant operating system in the world. This was before I knew about free software. I was very unhappy when Be Inc went down and sold BeOS to Palm. They managed to release some code under open source licenses before that happened, but since then BeOS was pretty much dead. It was also right about then when I started playing around with Red Hat Linux.
Enter Haiku
In 10 years things have changed quite a lot, and a group of people have been working on a complete BeOS replacement called Haiku (also previously known as OpenBeOS) that’s released under the MIT license. They announced a new alpha release today so I thought I’d give it a spin.
Haiku Installer
The Good
It boots very fast with no annoying flashes or disturbances during the boot process. I guess it would do a very good job of replacing any legacy BeOS 5 systems that are still out there already. The interface is also very responsive, and the installer isn’t too bad either, although it crashed when I only gave the virtual machine 64MB of RAM, which it runs fine on after installation, it installed fine having 128MB of RAM. BeOS is nice for very small devices. It should give you decent performance on even a Pentium II machine. My old BeOS 5 machine used to be a 75mhz Pentium with 32MB of RAM and it used to run just fine.
Haiku Desktop
The Bad and the Ugly
Pretty much everything that happened in the last 10 years is in the rest of the world is in a horrible state in Haiku. USB support is dodgy, power management is basically non-existent… oh, and if you want Wi-Fi, it’s fine as long as you have an Atheros card and only want to connect to unsecured networks. It’s essentially useless for laptops.
Even though it’s a unixy system, it’s still a single-user system. Everything is owned and ran by user “user” and group “root”, and there are no plans yet for fixing this. The /etc/passwd file literally just contains one entry. If I run the program called “teapot”, my configuration gets stored in /etc/teapot, not in my home directory! In my opinion the time for single user desktop systems have long gone, and it might have been ok in the mid 90’s where you just had to compete with Windows 98, but these days it’s really not sufficient. People have come to expect their computers to be much more safe and secure than that.
I think BeOS was a real nice system, and in many ways ahead of its time, but it’s dead and I think it’s best if they rather just leave it that way. I can’t see the point in spending another 2-3 years to make a system that is already obsolete in design, when those efforts could rather be spent on more worthy projects. Perhaps I’m just biased since I’m a big fan of GNU/Linux, we’ll just have to see.
Browse Feeds
YouTube Channel
Technorati Profile
LinkedIn Profile
Facebook Profile
LastFM Audioscrobbler
Twitter feed
FSF Member
Mugshot Page
Identi.ca Feed
Whoisi Page
Digg Profile
Plaxo Pulse
September 18th, 2009 at 6:33 pm
This is exactly what I’ve been saying. Good post, I share your opinion.
September 19th, 2009 at 9:00 pm
Hi Jonathan.
I’m in 2 minds about this. Firstly, I remember Beos as an O/S that could really thread. My test was always to format a 3.5″ floppy and then copy files and use the system. Only my Amiga system could do this reliably. On PC, OS/2 Warp was better than Windows & well, back then, I couldn’t get X running on my Trident vga card to compare linux honestly. However, Beos seemed to be the real deal. Fast and capable, but needed apps support.
In terms of the Haiku release, I hope they can build from the current alpha release to a beta with more drivers, possibly using intelligence already gathered in the unix/linux world to increase compatibility with wireless, etc.
However, my big concern is that every O/S seems lightning quick when its not finished. If you took a linux system of a few years ago and stripped out a lot of the daemons that run today on Gnome/KDE, it would be faster too.
In the same way that linux used to be much faster than xp for normal use, as functionality on linux has grown and matured, that difference has gotten to be negligible for most desktop tasks. In fact, XP seems just as responsive in many cases to Linux if you remove realtime antivirus.
So, as I slowly head back to the point, if Haiku continues to mature for another 4 years, will it really be faster than Linux is then?? The optimist in me says yes and I wish the developers well.
Having recently spending a lot of time using osx on my hack, I see a lot of areas where Linux can improve from a usability point of view so its good to have competition. If their community is passionate enough about Haiku to want to go up against linux, then good luck to them.
Your point about effort being reused in other projects reminds me of the one about gnome vs kde or amarok vs banshee vs rhythmbox. Its an effect of our community. It bugs me too, but I realise it fosters competition and a desire to improve competing projects. However, that’s an argument for another day.
September 21st, 2009 at 12:41 am
I cannot share your oppinnion istoff, because I am watching the haiku development quite about 6 years now and I have to say the developers are really really god damn serious about the performance. So they absolutely want to keep the BeOS Speed feeling even with more drivers and better sub systems (kits).
Its a micro-kernel based system that has some technological down points but in Haikus implementation you won’t feel them thx to the good work of the devs.
I think Haiku has a chance, not becoming a main stream desktop, but a interesting alternative for netbooks or low end computers.
Hardwaredriver problems you might encounter right now are very near to be solved. The OSS Support allows porting of FreeBSD OSS Audio drivers. The WLAN Stack will support all FreeBSD WLAN Drivers very soon, so I think they are much quicker than you might expect them to be.
@Jonathan:
“[...] obsolete in design. [...]”
I don’t think so. From a technical point of view its state of the art, even if some code is still missing or experimental. Hey its an alpha thats normal.
Haiku as a alternative to various linux distributions is a worthy project I think. The more choice the better.
There are lots of unique features or lets say usability ideas that might push the desktop in a big way.
I think its quite elegant to manage files only in my file-manager and not to have any external apps, like mail client or contact app or music management software that come with their own databases. The filesystem and its metadata is something big linux distros have but are not using. (well there is beagled that uses it for faster indexing but thats only 1/4% of what could be accomplished)
I think its quite a good idea to have a graphical user interface thats not build for servers but for desktops.
Single user desktops might be old, but maybe better and easier to understand for computer beginners.
The security argument here is something I don’t understand. Why is a multiuser system more secure than a single user one ?
Ah and that reminds me of a study I did for myself asking 100 Ubuntu users if they have more than 1 user account. Guess what the result was. 98 people had only 1 user account. 2 people had more but were mostly using only 1.
So I see not really a must for multiuser here.
Even a family most often uses one account because the main thing that they are doing is surfing the net.
I am using haiku quite a long time now (from pre-alphas until the alpha) and never missed multiuser I have to admit. (I missed more some apps)
Apps is a good keyword, thats the only problematic thing I see right now. There are some project that where started to port apps over or create apps and there are some BeOS r5 apps that are compatible but all in all there are not to much apps.
So Haiku needs apps !
But wait there is another problem. Haiku mainly supports only apps written in c++ and the beapi. So there is no global platform independent development kit.
There are some easy to learn languages though as yabasic and python but gtk+ and qt4 are not wanted.
Its a little bit strict by the haiku development team. But they are mainly fearing the performance issues and the “not feeling native” problem. That will cause some big problems I think, because a normal developer wants to develop software that is widely used on many different platforms.
But at least there is a chance. Thanks to the unique structure of Haiku and the Translator principle its possible to create high quality software that might fit some multimedia points Linux Distribution cannot fit right now.
September 21st, 2009 at 8:12 am
@Leszek, multi-user systems (such as unix-type systems) are more secure because a user doesn’t have direct access to the entire system. Printing, file sharing and other services would run under its own user, which means that if a user account is exploited, it will often only affect that user. Programs that are run as a user also only inherits their rights, which makes it hard for virusses to spread.
September 21st, 2009 at 10:19 am
@jonathan
This argument is best held over a beer, but frankly, I don’t care if the attacker gets my user only and not root. i’d rather rebuild the system than lose my photos / source code, etc. My linux o/s is the most frequently replaced part of my systems. I have /home backups that are years old, which are far more valuable to me than a compromised O/S. I of course realise that this is reversed in a corporate environment where security of the o/s is paramount, of course. I don’t dispute the validity of your overall point, just saying we should not underestimate the potential damage done by losing access to your user credentials due to an attack.
@Leszek
I hope that the level of control / caution exercised by the Haiku team will help us avoid a similar fragmented set of competing communities like we deal with today on Linux.
September 24th, 2009 at 10:18 pm
Haiku is neat, yes. But so is Amiga, which, like BeOS, died for lack of software. Unless Haiku promoters have some magic marketing trick up their sleeves, it’ll end up in the same place, six feet under, due to the same cause.
And it seems to be a silly effort anyway: is it worth reinventing the wheel to begin with? Does it promise to be *far* better than Amiga? I think not but i’d love to see Haiku developers reply to this one.
No, we’re stuck with the damned if you do and damned if you don’t range of non-choices [Windows, Mac, Linux].