Daemon News 199908 : The Penguin's Council

The Penguin's Council

Copyright © 1999 Aaron Sherman

Praise

Wes Peters made a wonderful contribution to the community in his article, Daemon's Advocate, in last month's Daemon News. The tone was reasonable, the content thoughtful and the points well worth consideration. This is worth pointing out because many such comparisons (of Linux and the *BSDs) do not go so... smoothly.

One interesting note is that the line is not clearly drawn in Wes' article between the OS and the community around it, so he happily throws his analysis at both, and lets the reader sort it out. In essence, I think that this was wise, but it does dodge a few points which I think were worth making (hence this article). I will address primarily the community, and not the OSes themselves.

Did Somebody Say Commercial?

These days, the definition of "commercial" (as in commercial operating system) is becoming a little fuzzy. Linux was not considered a commercial operating system when all that existed was Slackware and few other ad hoc distributions of the kernel and the most commonly used utilities. Now, Linux enjoys the financial backing of Intel, Netscape, Dell, NEC, Toshiba and others (mostly through investments in distribution vendors or the creation of dedicated development groups internally) as well as several announced or planned IPOs. This does not make Linux "better" than it was, per se, but it does foster an environment where businesses can treat Linux as a commercial offering which has an established role in the market.
Why is this important? Well, it's very important to the BSD world, because this is where BSD _could_ go in the next few years (especially with the success stories that Wes mentioned in his article). So again, as Wes points out, BSD has the opportunity to benefit from Linux "converting" the masses to the UNIXish way of thinking.
More importantly, though, this gives the BSD camps a chance to determine if this is where they _want_ to go.

Hardware Support

One thing that I don't think has been analyzed yet is what will happen if hardware vendors start determining that Linux has reached a critical mass, and begin writing their own drivers. Assuming that this follows a best-case scenario, and the drivers are all Open Source, the BSD community has two options: one, to port the drivers one at a time; and two, to write an interface layer which allows the BSD kernel to take advantage of a Linux driver. If there are hundreds (perhaps thousands) of drivers appearing in this way, it may not be reasonable to try to port them all, and the second option may come in handy.

Cross-Pollination

Wes mentioned the concept of cross-pollination between the Linux and BSD development efforts, and while I agree with what he had to say, I do not feel that this cross-pollination currently goes far enough. The Linux world has a lot to gain from more contact with the development community of BSD, and while the BSD camps may prefer to think of Linux' developer base as "teenagers" who write drivers for doohicky hardware, people like Linus, Alan, Maddog and the vast majority of the Linux developer world are well-versed professionals who have a great deal to teach to as well as to learn from the BSD developer world. This sort of environment, where experts can freely trade information without the interference of corporate IP-hoarders is rare, and a thing to be treasured.
Most importantly BSD should be keeping an eye on the technical advantages and pitfalls that come from massive commercial involvement. In the end, Linux may stumble (or be knocked over), at which point the UNIX community may have to recall the conversation between Yoda and Obi Wan in The Empire Strikes Back, as Luke leaves to confront Vader....