Dæmon's Advocate-BSD rules!

Greg Lehey <grog@lemis.com>

There's a new movement afoot: the governments of the world are gradually reassessing their software strategies and finding Open Source software useful. From Peru to Pakistan, from Denmark to Australia, governments are seriously considering basing large projects on Open Source Software (or ``Linux'', as they tend to call it).

Obviously government computing needs differ greatly from those of the typical reader of these articles. We're left with a number of questions:

A few months back, the Australian UNIX Users Group decided to issue a call the Australian government to accelerate its adoption of open source software. We specifically mentioned both Linux and BSD. Primarily as a result of this message, we became involved with the Australian National Office of the Information Economy, and we have learnt a lot.

Not stupid after all

Our first question was: how could we convince people that open source was the correct thing to do? How could we explain to them what it was? We found ourselves in the position of technical experts who suddenly had to reposition ourselves to see things from the outside. Well, so we thought. To our surprise, we found that the stereotypical idea we had of slow-moving, lazy bureaucrats was way wide of the mark, at least for the people we met. They know exactly what's going on. No need to sell open source there.

One of the results of the discussions is that we will participate in a seminar for government CIOs and CTOs to be held in February next year. We're looking forward to that with a mixture of excitement and apprehension: on the one hand, it's a great opportunity to spread the word where it can do a lot of good, but on the other hand it's an area where we have relatively little experience, and we have little idea of the attitude of the participants. We've already seen that we underestimated the people at NOIE. We don't want to make the same mistake with the seminar.

There are a lot more details of interest, but I can't mention them here, at least not yet. The details which have been made public are in a press release issued a couple of months ago. This article isn't about the NOIE seminar, of course: I'm looking at different themes here.

BSD or Linux?

If you follow the links above, you'll see a general equation of ``open source'' with ``Linux''. Even the NOIE Open Source page has a penguin on it. That's far from the truth, of course. Even if you totally ignore BSD, Linux is only a small part of the total body of open source software which could be of use to the government. The point is, it's the platform.

People get used to this equation almost without thinking about it. From the point of view of the AUUG I need to maintain strict neutrality, of course, but on occasion I have used the term ``BSD'' instead of ``Linux'' to indicate that there's more to open source than Linux, and it shocked some people who, though they were relatively neutral on the topic of Linux versus BSD, were unhappy that I mentioned BSD without mentioning Linux.

So how do we get more recognition for BSD in the government? Do we want to? If we do, it's not going to be easy. Even with open source, government software projects are commercial activities, and large computer manufacturers such as IBM are very active in the arena. None of these companies have any plans to supply BSD-based software. There's only one exception: Apple Computer. So far they don't seem to have been overly active in this arena, though it would make a lot of sense for them to do so.

The top end

There are a number of areas where open source software could be of use in the government. The ones which hit the headlines are the big ones, of course, like the deal with Centrelink, the unlikely name for Australia's Social Services department. Obviously they're relatively few and far between, and they're also the ones where BSD currently has no chance. That situation will remain until some company takes an interest in promoting BSD rather than Linux. My own efforts at IBM have shown no signs of success.

The bottom end

There's the other end to consider as well: everybody has a computer on their desk nowadays. Low end white box computer hardware starts at about $500 Australian ($300 US). A Microsoft operating system costs another $200, and and Microsoft ``Office'' costs $550 on top of that, increasing the system price by 150%. For some people, the functionality and self-compatibility of the Microsoft product may warrant using Microsoft, but for many users OpenOffice makes a perfectly good alternative, and it's free.

OpenOffice runs on Linux, of course, but it also runs on any BSD. That could seem to be a good place to target BSD. But why should any government department want to run BSD instead of Linux on such machines? I can't really think of any good reason. I can think of a number of reasons why not:

The middle?

So what's left? In the short term, I can only think of things like web servers, where BSD's superior performance, security and reliability are a great advantage. Even there, though, Linux is catching up. It's unlikely that we'll make many inroads there either.

What do we want?

This doesn't sound very promising, does it? But I still haven't answered one of the questions I posed at the beginning of the article: do we even care? What goals do the BSD projects have? In the FreeBSD handbook you can read that the prime aim of the FreeBSD project is to provide code to all comers. This is obviously a statement which could do with some revision, a thing I'll have to personally take in charge. NetBSD does better: paraphrased, they're to provide good, unencumbered, portable, standards-compliant software. OpenBSD quotes security, standards compliance and the best possible development platform as the main priorities.

One thing that all the projects have in common is that there is no mention of end users. Yes, OpenBSD wants to be the best development platform, which suggests that developers are the intended end users. But FreeBSD and NetBSD look more like self-contained projects intended to benefit the project itself. Is this still the case? There's nothing wrong if it is, but in that case we shouldn't be surprised if Linux gets all the limelight and BSD is used only by those in the know.

I believe that these goal statements all need reworking. The truth is that the BSDs are getting more and more exposure, and that more and more people I meet are actively using BSD. I have spent the last nearly three years working with a small group of people who at the beginning were purely Linux-oriented. Now at least three of them, including some famous names, are using BSD instead. On a more mundane note, I got a phone call a couple of days ago from somebody who had found my name on Google while trying to find a source of FreeBSD CD-ROMs in the Adelaide, South Australia area. I had them, of course, and told him where I lived, out in the sticks halfway between Echunga (population 436) and Meadows (population 560), and there must be about 30 houses on the 10 km road between the two towns. It turns out that the person calling lived on the same road, 2 km away. Clearly that's not statistically relevant, but it's certainly interesting.

So, in summary, what do we want? Do we want to have a BSD on every desktop or in every server? Do we want large government departments to use our software? Do we want anybody except ourselves to use our software? Until we know the answers to those questions, we don't know whether we're doing well or badly.