Un ptit troll ne vous fera pas de mal
-----
I run NetBSD on x86 instead of FreeBSD now for many reasons. The main reason is that FreeBSD is more like the Linux of the *BSD's now. FreeBSD is very easy to use and configure and less like true Unix. Just one small example: by default FreeBSD uses a csh shell clone tcsh.... csh is linked to tcsh. Often FreeBSD users will berate Linux newcomers to FreeBSD for the preference using Bash. Indeed, they miss the irony at play here. With this same logic NetBSD users could bash (pun intended) FreeBSD users. The reason being is that NetBSD uses the real csh by default. It's little things like this that make NetBSD much more like Unix. It is also why it will seem very newbie unfriendly. I also see little things such as FreeBSD removing perl from the base system as a last ditch effort to return to a more unix like feel by mimicking NetBSD.
The other reason why I prefer NetBSD on x86 and all other platforms is because of American/Internet culture. The original design of the Internet, Unix and the C language itself was around the concept of highly portable systems communicating with ease over common protocols on different archetectures. If you understand the OSI model of TCP/IP, Know the motivation Kernighan and Ritchies canonical C language, and also know the goal of UNIX then one must accept that NetBSD is the only OS today that can accurately give off the vibe of the one true computing platform.
Also, it makes a great platform for research environments. NASA uses NetBSD and some NASA scientists are actually on the NetBSD coding team. Nasa has saved the United States tax payers literally millions over the years by using NetBSD for its experimental projects.
It's for these reasons that newer cutting edge features are first implemented in NetBSD. For instance ipv6 was first implemented and tested on NetBSD by the Kame project ( see Itojun's cvs changes).
Another reason: If you want to seriously get into kernel hacking and general system hacking.. NetBSD is the best platform. The reason is because it is so clean and efficient. There is not alot of extraneous code to confuse and distract you. Instead it offers you only the best component code to cut your teeth on.
This may be mere speculation on my part but I would say this is the reason why the best OpenBSD hackers prefer NetBSD over FreeBSD. In fact Theo De Raadt and Miod learned how to hack on NetBSD not FreeBSD. Miod and many others have expressed their preference for NetBSD over FreeBSD on many occassions.
Last reason: Hardware upgrades in an out of control Keynesian occidental capitalist society without traditional laisse faire natural market conditions will invariably lead to software bloat (commercialism). In order to drive demand within an agreeable relationship with Moore's law and the subsequent market hype from the ubiquitous horizontal and vertical media.. Indeed, in a Laisse Faire capitalist system without fiat paper controlled by the Fed and consequent government meta controllers such as the New Deal : these useless software company parasites would wither and die. I am not railing against usefull Government measures like the Taft-Hartley act however.
The first thing a rocket scientist at NASA working on a donated platform or a serious hacker wants to do is escape this cycle. To get away from the noxious deleterious entropy created by the synergistic effects of the economic and scientific laws. Gresham's law and the second law of thermodynamics (being applied abstractly and metaphorically outside it's scope to software bloat here) respectively. Serious researchers must escape the useless sparkle and glimmer that distracts lesser simians and go for consistency and quality. The former will add needless stress not only to researcher but to their test archetectures as well.
NetBSD is the most uncommercialized research OS on the planet most likely. So it will not fall trap to the cycle mentioned above. Instead the code upgrades to NetBSD are based soley on proper engineering decisions. I would say FreeBSD is not totally immune from the cycle mentioned above due to its various ties with corporate entities.
NetBSD does have some corporate ties -- such as Wassabi systems. However, the embedded market does not lend itself to arbitrary useless bloating of software in order to force perpetual upgrades as much as the PC market does. This is due most assuredely to the constraints of such small systems. Therefore the negative effects on NetBSD are minimal and are instead mostly positive. For instance, they pay many NetBSD core hackers a living salary and many of their clients submit code to the project.
In leyman's terms: Instead of getting distracted by a useless shiny new widget and upgrading FreeBSD's portage system by CVS and recompiling to the newest version for the umpteenth time. In NetBSD you may get bored enough to start hacking TNRG algorithms based on quantumn probabilistics to try to work with the ostensibly damaged SMT zener diodes by removing the only protection layer-- a simple #define in the hifn driver for soekris pci card ported from OpenBSD. In otherwords: hardware driven cryptography engines. Or some other cutting edge experimental features. The resulting work would not only benefit NetBSD but also the other two BSD's as well. The FreeBSD user would be able to do the same but is less likely to-- especially a newcomer.
Besides, that, if a FreeBSD user was working on the NetBSD platform they could very easily hack a freebsd binary package handler. Jordan K. Hubbard added all the meta data in the FreeBSD .tgz packages one needs to create such a tool in NetBSD.
In otherwords: you could create a tool to install FreeBSD binaries under NetBSD's FreeBSD emulation with less than %1 overhead but since your average FreeBSD user is roughly equivalent to your useless Linux user this is not likely to occur.
FreeBSD is good if you just want to be an end user like Linux.
NetBSD is good for serious computer scientists and hackers.
---------------
Membre du Front de Libération de Datoune | Soutenez le FLD | A Tribute To Datoune