On the topic of Wayland, video games like Unigine Oilrush have no problems drawing at the refresh rate with the right hardware and driver combinations while Xorg performs extremely well on my system. In the few cases where Xorg does not, I have found flaws in software, such as usleep() being used in konsole's draw routine. While Statements from users claim that Wayland is some wonder cure, statements from Wayland developers say that any improvements would likely be imperceptible. I have also yet to see an explanation as to how a protocol that ran smoothly on hardware from the 1980s could be a performance bottleneck on hardware today.
As for "outdated and less drivers", Linux's hardware support is not a superset of FreeBSD's. Furthermore, the drivers that Linux does have can often be buggy. r8169 is a good example of this (although not necessarily on all supported chips). Several Gentoo developers have had problems with it and multiple kernel developers working on the BQL patches stated on mailing lists earlier this year that the driver is broken. Between merging buggy hardware support and not having hardware support, I would prefer it if Linux opted to forgo hardware support. Distribution developers like to be told when they are beta testing buggy software. Having it in the tree when it is known to be broken simply wastes our time.
On the topic of init systems, systemd and its dependencies add 2 million lines of code to the early boot process, which is a significant reliability concern. Furthermore, its lack of support for other platforms raises questions regarding code quality. I would consider not having it to be a feature, rather than a liability. At the same time, Gentoo has its own next generation init system called OpenRC that is implemented in 10,000 lines of code and works on both Linux and *BSD. It has been offered to the FreeBSD developers, but if you must have it now, you are always welcome to install Gentoo FreeBSD. Interestingly, Lennart Poettering has stated that systemd is incapable of starting X and needs Wayland to do it. It is a bit odd to me that 2 million lines of code (when including dependencies) could not include code for launching X, but somehow, Gentoo has managed to do it with an init system that only has 10,000 lines of code. That is not systemd's only issue, but it is food for thought.
On the topic of Pulse Audio, can you tell me what it does for you that that OSS does not do? I mean something that you actually use. With that said, I find it hard to believe that the audio framework should keep you from using an OS. With that said, we could probably get its benefits if were to devise some kind of OSS forwarding support for SSH.
Lastly, FreeBSD has multiple distributions. Here are some of the more popular open source ones:
- FreeBSD upstream
- Gentoo FreeBSD
- ZFS Guru
Wow FreeBSD 10, that was fast...
Would I need to upgrade then in the near future?
How long will FreeBSD 8.2 be supported and when will the upgrade be close to impossible?
Those are questions I am interested in
To compare ZFS FreeBSD and Linux:
1. Linux hangs from time to time if you move large files and videos hang sometimes
2. FreeBSD does ZFS with compression and Apache etc. running in ~100MB of memory
3. Linux desktop with ZFS uses ~10GB of RAM, there is this bug to apply module parameters if it is the root fs, so you cannot influence the arc cache and it has to be a module since the licence won't allow it in the kernel to e.g. use the kernel commandline
http://www.freebsd.org/security/security.html#sup for support dates, but take them with an unhealthily large dose of salt.
Upgrading shouldn't be a big problem - the handbook and FAQ cover it quite well, and the people on the IRC channel (##freebsd) are helpful if you have any questions.
However, by saying FreeBSD is pathetic I don't think he was talking about stability. Linux and Unix like systems are the most stable. The problem with BSD is it's always behind, because of the lack of manpower. The features that are comming to FreeBSD 10 are already present in Linux.
Will this shut up people who complain about PA and say it's a layer? What's more funny Windows and OS X are also using something like this, so if everyone is using such layer now, people should shut up, right?Uber-audio daemon.
http://en.gentoo-wiki.com/wiki/Systemd#startx, it is not clear what you are talking about. systemd certainly doesn't need Wayland. In fact, Wayland could use some changes for systemd to work better with it including for instance, making it more compartamentalized so that cgroups support is effective. If you provide a reference to systemd and X, that would help.
And that's actually one of my reasons for using FreeBSD on my server: Linux is huge and the amount of patches from one version to the next is overwhelming. More changes, more chances of bugs.
It's not just a kernel and you have to pick a distro and rely on that distro to get security updates. It a full system including userland.
Upstream supports multiple stable branches, security fixes go to all of them and security advisories get announced on freebsd-announce.
For Linux, there's the inofficial linux-stable, where patch submitters/reviewers have to take care of cc:stable. Linux distros use that as a base and throw their own (sometimes crappy) patches on top of it.
I find that alone way more compelling. Let alone all the neat features of FreeBSD.
It's is less bloated and I've been running it with an uptime of multiple years. Rock stable.
Sure, for most users its probably less useful as a desktop system compared to Linux. And I do use Linux on my desktop box, but for servers is find FreeBSD far more usable.
It's far from pathetic, and calling it that for not seeing *your* personal advantages makes you narrow minded.
Nice flameware going on here.
Of course, its irrelevant.
Fact is that Linux has a FAR greater reach than BSD, and it continues to advance at a breakneck pace while BSD isn't apparently going anyware very fast.
Yes, there are some nice features of BSD that Linux lacks. I think there are more nice features in Linux that aren't in BSD, but that's my opinion so fuck off if you feel like countering this point.
The facts are laid out above.
The reason for this is that there isn't much INCENTIVE to contribute to BSD.
That is caused by their license.
The unfortunate problem with the BSD license, is that for it to be successful, everybody who works on it or uses it HAS to VOLUNTARILY contribute back to it. It depends on the HONOR system, and we all know that MOST people and organizations LACK HONOR.
There are plenty of people and organizations who would happily contribute to it, IF they could trust that the OTHER people and organizations who take that code would behave as honorably as they did. I.e., contribute back their enhancements in order to build a strong open source codebase.
However, everybody knows that there are plenty of ASSHOLES who will take other people's HARD WORK, and keep it to themselves without sharing back their enhancements.
As a result, you have those who WOULD contribute, but don't in fear that others would steal their work, and you have those who don't contribute back because they're ASSHOLES.
That's what makes Linux much stronger with the GPL license.
With GPL, you are OBLIGATED to contribute back (hence solve the ASSHOLE problem),
with GPL, you don't have to be AFRAID of the ASSHOLE problem (because it is solved),
For this reason, Linux will always be more successful than BSD.
Yes, the BSD license is obviously more "free", but is, unfortunately, TOO free. It is SO free that it can be UNFREED. GPL is a license that FORCES freedom on you, even if you don't want it. I would classify GPL as a "balanced" license. It is not too free that it can be unfreed, and it is not unfree.