October 24th, 2005

The better part of 'better, faster, cheaper'

Posted by Paul Murphy @ 2:39 am

Categories: Distributions, Enterprise Policy, General, Linux

Tags:

Last week talkback contributor John Zern called me on a casual reference to Linux as "better, faster, cheaper" than Windows. Here’s part of what he said:

How do you tell someone who has successfully used Windows for quite some time that Linux is "better"?

That part of the statement is purely opinion, as "better" is a relative term.

As an example (and it’s only one person, I will admit):

My brother-in law had set up a Linux server at home, his first server ever, about a year and a half ago, and had it running nicely after some time (he was new to it). When I purchased 2003 Server for what i need it for, I gave him the Server2000 to play with. He found it much easier to use and couldn’t believe everything it did right out of the box! He now feels that the Windows server is better for what he needs it for.

What I see here is that Mr. Zern is both right and wrong at the same time. He’s right in that there’s no reason to think that Windows 2000 server wasn’t better for his brother in law than Linux was.

We know this because he says so, and we have no other evidence. On the other hand the conclusion he draws, that Linux isn’t better than Windows, does contradict other evidence. Specifically, he refers to what Windows does "out of the box" but few people touch even a few percent of everything the typical Linux distribution ships with so what he’s really comparing here is the ease with which his brother in law was able to access the capabilities he set up for him - not the totality of what’s available "out of the box" in both cases.

Indeed if someone were able to list all the capabilities of both distributions, rate them according to their cost, reliability, and performance when used correctly, and then compute an appropriately weighted sum, most people would be utterly astonished if the result didn’t favor Linux by a considerable margin. Take a look, for example, at subsets of that comparison in areas like security, web services support, or multi-role capability and even a complete Wintel bigot would have to agree that Linux is better -and, indeed, what you typically see them do is retreat, as Mr. Zern does above, to that ease of access argument.

The problem with the ease of use argument is that if it really were the basis for the choice then you’d expect Mr. Zern to have helped his brother in law get a MacOS X server in place. Not only is that the leading OS in terms of ease of use, but you get both the stability of Unix and the power of the G5 processor for less money than a Dell. But he didn’t push MacOS X, and because he didn’t I feel entitled to conclude that "ease of use" wasn’t the issue in any objective sense at all.

So what was it? I think Mr. Zern did what he knows how to do - set up Windows servers, and thereby produced a comparison between Linux set up by a beginner struggling to get something working and Windows 2000 set up by an expert.

And that, I think, is the most common fallacy underlying these kinds of comparisons: we implicitly assume comparable expertise and comparable ancillary costs without stopping to acknowledge that we don’t generally find either one in reality. That’s why a general claim that Linux is better gets trumped by the reality of individual experience, it isn’t that Linux isn’t better: it’s that Mr. Zern’s brother in law doesn’t know Linux but did have an expert ready to give him a free copy of Windows 2000 server and help him get set it up right

There’s a moral to this story: that it’s the combination of expertise with technology that counts, not just the technology. The objective reality is that there isn’t anything that Windows Server OSes do "out of the box" that Linux can’t do better but the practical reality is that Windows set up by an expert works better than Linux set up by a beginner -and the scary thing is that most of the expertise out there is Windows only.