I'd venture that your professor isn't particularly well-educated if he thinks BSD is dead or dying from either a commercial or a pedagogical perspective. A considerable amount of literature on the subject of networking is written using the BSD codebase as reference (e.g. the Richard Stevens TCP/IP books), and I don't expect that anyone is going to turn around and tell you that the Linux people got to where they are by ignoring all of that literature and the code base around which it was written. Second, beyond the base of open source host networking stacks, the BSD code base has been extensively grafted into proprietary Unix implementations, not to mention serving as the foundation for dedicated network devices such as Junos. You might argue that Junos isn't as prominent in the market as Cisco, but there are a fairly considerable number of arguments against teaching using IOS implementation pedagogically, except perhaps as a long series of "gotcha" lessons. Third, BSD networking continues to be grafted into other systems. A perfectly good example of this is that Sun has ported BPF into the Solaris kernel to support firewall portability as one of recent extension and refactoring initiatives to improve its network performance and provide an alternate set of interfaces for portability of networking code (e.g. for kernel code, or as an alternative to write directly to DLPI or through libpcap for anything that can't be implemented via [*cough*] Berkeley sockets).
The crux here is that the wisdom of acting as though *nix networking is a monoculture completely dominated by Linux (which in my opinion can both fail to be a monoculture in the way it needs to be and succeed in being a monoculture in ways it needs to curb) or will become one doesn't seem the only possible conclusion from examining the history or contemporary dynamics (and that's setting aside the rather material question of whether such a monoculture would be desirable in any case, given how important cycles of divergence and convergences have been to making *nix what it is qua dynamic and open systemnot to say that Linux is a monoculture... or as dynamic and open as ). Sure, Linux can have its value as teaching material, but it's far less credible to do so if the premise is that this is the only open source implementation worth teaching. There may be valid reasons for focusing on a single implementation in course design, but dismissing the value of a comparative approach or of subsequent independent study of other systems strikes me as pissing away credibility as an instructor and being dishonest about course design decisions. As for the instructor, you can lead a horse to water and all that. Perhaps the more important thing to learn here is how and why he's mistaken rather than that he is or to push him to such concessions. If you can't push him so far as to change his decision, but you can perhaps offer sufficient judicious counter-arguments to make other students want to learn more and build some continuing study groups on top of that. Cheers, Bayard Am 13 Feb 2010 um 08:06 schrieb TS Lura:
Dear OpenBSD community, I'm a student for a MSc Advanced Networking degree. I have a little situation maybe you guys could give me some feedback on. The issue is that my module leader is refusing even to consider mentioning OpenBSD, or any BSD in introductory Linux course where the focus is on network services. DNS, iptables, Apache. It is a introductory course, with limited time. So it's understandable that one has to be level-headed on what's to go in as material in the course. My argument is only to have a reference to OpenBSD, PF, and maybe the jailing of named, when we go through the topics of iptables, and DNS. My professor (the module leader) argue that almost no one is using BSD, and those that does is probably 70+ and so it will soon die off, in a humours tone. In more serious tone, lack of applications. I'm a bit resigned by this attitude, because we are at a master level about networking. We learn about all the technologies surrounding routers, switches, wan, security, etc. As such I think that OpenBSD is really a bean to be counted when we learn about open/free software. So in relation to this, I would argue that OpenBSD is a excellent platform for networking services. I have said so in writing, and verbally only to be brushed off. I feel it's game over, at this point. But maybe you guys have some suggestion about good arguments that might persuade my professor? Cheers, TSLura. PS. This might be the wrong crowd, but I also argue for the documents on the internal web-learning facility to be published in PDF (ISO 32000 standard) (he insist on doc), and that Linux at least once should be mentioned as GNU/Linux.(system-tools/Kernel, to pay tribute). This is also met in the same way as my BSD arguments. Which I find strange, since my professor has developed a bit of stuff for the GNU/Linux platform.