Reasons Why You Should Not Use FreeBSD
Phoronix: Reasons Why You Should Not Use FreBSD
After a few days ago sharing a list of why you should use FreeBSD as said by FreeBSD developers and community members who use the BSD operating system, here's a list of reasons for why not to use FreeBSD or missing functionality...
Drivers and user programs are important for users!!
*cough* stable api's, you too Linux *cough*
Ha-ha-ha - last point, is valid for almost any other Desktop OS, unless you are forced to use it by your applications work set.
I wonder if OSX has all those features that FreeBSD has AFAIK OSX is even more limited than FreeBSD
Examples: ext4 only with FUSE...
ZFS support will cost you 20$ or 40$ for full featured support
Most of the mentioned (and too early reported herein!) issues with FreeBSD are neglegible. ISDN/SIP is important for many people. A GLUSTER (Linux terminology) filesystem is also important, but only for a very specific group of users - in most common areas clustered and distributed files are not yet common as this lack of feature may imply. But if we come closer to HPC (and this is a very important whether a OS is applicable or not), then FreeBSD is definitely NOT the first and even not the third choice! Missing fast compiler (no Intel compiler, no NAG compilers, no PGI compilers, no PathScale compilers) support, a complete lack of GPGPU support which is very tight bound to the lack of drivers for graphics cards. Also, highly optimized and fast mathematical libraries are missing. NAG droped FreeBSD support when FreeBSD moved to 64bit. I think it was no coincidence that the change from kernel 2.4 to 2.6 on Linux made the gap between FreeBSD and linux smaller in terms of the reliability and speed of the network stack.
Politics is a very important aspect. There was no serious reason why people should have used Linux or even Windows when both of those OSes where instable and forced people to be more engineer than user. I'm now for approximately 20 years with BSD systems, since 1996 with FreeBSD. Although the BSD/MIT like licensing model is very developer friendly, people tend to prefere Linux with its sight-blocked GPLv3 licensing. I can not understand that many departments especially at German universities rely on Linux and not on a more academic OS like the *BSD. I reject the possibility that people are getting more and more stupid these days, so the conclusion might be that it is all about politics.
It has been mentioned in the list of PROs that FreeBSD has a very nice port system. Yes, it has. But the way it is maintained is a bit dangerous for those who rely on ports. UPDATING is very often updated AFTER critical port updates rushed into the tree and its then too late to back off. Such a situation occured just now with the new PNG library version 1.15: large ports like OpenOffice and LibreOffice are affected by this and on all of our remaining FreeBSD systems with LibreOffice installed LibreOffice isn't working anymore, rejecting compilation due to an error (I NEVER managed to compile OpenOffice or LibreOffice on a maintained, running and working system and most of the complains in the list where commented by people which installed OpenOffice/LibreOffice in a virtual machine or on freshly installed boxes and so they "compensated" complains about this lack of maintenance - this can be very frustrating and even negates the very good help of those who are willing to help).
We also run into harshg problems with Linux when it comes to maintainance - but I have the feeling, that on Linux, Ubuntu and Fedora especially, the availability of ports is better and somehow they manage it better to deal with different library versions.
As much I would prefere FreeBSD, there is a no-go when it comes to work and science. We rely on OpenCL these days in many projects, since our TESLA systems and even the ordinary GPUs in the desktop boxes deliver a tremendous performance boost I never achieved before with multiprocessor systems. We can use those platforms for nearly every "business" of n-body simulations, fluidal simulations and for our PR stuff we prepare with Blender, which is now capable of handling a present GPU via CUDA or OpenCL. GEGL library is about to move also towards OpenCL and speed up GIMP - a tool, many UNIX users won't miss and I bet many professionals are appreciating a speedup as seen in Adobe's Creative Suites utilizing a present GPU. Linux can do - the *BSD, especially FreeBSD, can not. And FreeBSD is even more loosing ground in science and other high performance areas and there is no political "will" to order the helm to apply an alternative route.
Who ever have seen the tremendous performance AMDs new Graphics Core Next architecture is providing using double precision calculations in OpenCL, or even with the more moderate recent TESLA, will never use anything else than a platform which is capable of using GPGPU.
PathScale was once offering/promising something that could make life easier on FreeBSD, called HMPP, but I never saw a product or anything useful coming out - since two years now. What happened? No market?
One can not harvest if he doesn't seed and plant.
Mac OS X is even easier to run on Hackintoshes....
I specially register to reply this thread. In my views, it's even easier to set up OS X on (illegal) Hackintoshes (i.e. non-Apple PCs) than F*BSD.
OS X has a beautiful and elegant installer interface while ONLY ASCIIS are on BSDs.
Here's the best reason why not to use FreeBSD: It's not Linux.
If this would be the best reason why not to use FreeBSD, I would use it.
Last edited by LightBit; 06-02-2012 at 02:58 PM.
It's fragmented, uses crappy license, its file system sucks, it's slow, it's old, legacy, lacking man power and automation, it's MS and Apple friendly, people from foolBSD forums are trolls etc. and Linux is stable, reliable, fast, has greatest community and license etc. If people have adblock turned on then such stupid articles won't bring you too much profit.