Linux Today: Linux News On Internet Time.

More on LinuxToday

The Value of Choice or the Choice of Value?

Jul 27, 2007, 22:30 (31 Talkback[s])
(Other stories by Brian Proffitt)


Desktop-as-a-Service Designed for Any Cloud ? Nutanix Frame

By Brian Proffitt
Managing Editor

Let's face it, last week I was being a bit of a smart aleck.

When I wrote a passioned statement about how having a choice of distros was a good thing, I decided to list 359 other things that it was good to have choices for. I thought it was a good effort, and still do. But in the course of a week I have also come to realize that other points of view may have some validity, too.

On Monday, I tagged along for dinner with a group of people from Canonical and Sun Microsystems. It was one of those impromptu things where I went up to Stephen Walli at a bar and said, "Food. Now." Then he mentioned it some people, and so on, until suddenly there are over 20 people descending upon a seafood restaurant in downtown Portland. Whatever, I was hungry.

When the dust settled, I was pleased to find myself sitting across from Jeff Waugh, of GNOME and Canonical. Fan-boy that I was, I managed to strike up a conversation with him about things in the open source world. Along the way, we started talking about the notion of choice in the community. I mentioned the crux of last week's column, thinking myself quite the clever person. Waugh, not so impressed, as indicated by the frown that crossed his face.

To him, the notion of choice is something that is rather overdone in the open source community. In fact, he thinks that those who argue that the best thing about open source is the availability of choice are completely wrong.

He is not denying that choice is an inevitable outcome of open source development, but he is concerned that too many developers are pursuing new projects for the sake of choice itself.

I wanted to understand where he was coming from, so I said that while I would easily agree that too much choice was unwieldy, having different options was still preferable to just one. In the case of 359 Linux distributions, I said that yes, that was a lot, but very few people actually slogged through every single choice... typically, it comes down to deciding between a few distros that realistically fit a need.

At this point, Waugh shook his head. Even a few choices was not necessary, because ultimately it comes down to what's the best value for the customer. The customer chooses based on the best value, and having too many choices leads to a situation where none of the choices are capable of providing that value.

I leaned forward a little, prompting the other diners around me to chide me for wanting to get into an argument. I relaxed, because intellectually I knew that listening is always a better way to learn. But it wasn't easy to see my stance get deftly, but politely, whacked upside the head.

When you buy a car, I responded, you're going to want to have some choices, because it's a major purchase. Yes, Waugh responded, but when it comes down to it, you're still only going to choose the car that fits your best value model.

Essentially, Waugh was asserting that developers in the open source community should spend less time focusing on the value of choice and more on building what's the best value for all end users.

By the end of the dinner, it was clear to me that he has a good point about choice for choice's sake. If a developer is going to break off a project and build something to scratch an itch, why can't that itch be scratched in the original project to begin with? Of course, the GPL and many other open source licenses allow that new feature to get ported to the first project, but will the actual work get done?

There, I agree with Waugh. But I must also honestly state that I think on some level, choice is still good, because there are fundamental differences in application coding that don't seem to fit everyone's value system. The GNOME vs. KDE debate, for instance: GNOME is designed for and by users who don't want the ability to configure every little thing. KDE is built for those who want lots of choices on many elements of the environment. It seems that at this level, fundamental values that aren't compatible are getting addressed.

I am still mulling this one over. I sense that we are both right on this, but I am struggling to codify how choice should work in the open source community. Help, as always, appreciated.