Wednesday, November 29, 2006

No one cares about "average"

http://discuss.joelonsoftware.com/default.asp?joel.3.445505

No one wants to be average. We just want to treat everyone else like they are. That's why we generalize, form stereotypes. It leads to goth and emo, as kids try to be unique and different from average people ... but in exactly the same way as everyone else in their preferred group. (Oops, was that cynical?)

It also leads to being pretty spectacularly wrong in your conclusions, even when you're completely right in your premises. Take for instance what Joel Spolsky had to say about OSS development:

Development [on OSS projects] is much more likely to be geographically dispersed, which results in a radically different quality of team communication. It's rare in the open source world to have a face to face conversation around a whiteboard drawing boxes and arrows, so the kind of design decisions which benefit from drawing boxes and arrows are usually decided poorly on such projects.
His premises are correct: development is often dispersed; design can suffer. His conclusion, that OSS projects usually just clone existing (closed source) software is not technically wrong, but is irrelevant for the same reason traditional media is wrong when they discount blogs as "real writing": The average doesn't matter. If the best blog -- or the best OSS product -- is better than the leading commercial media or closed software, it will gain market share.

As Paul Graham says:
Those in the print media who dismiss the writing online because of its low average quality are missing an important point: no one reads the average blog. In the old world of channels, it meant something to talk about average quality, because that's what you were getting whether you liked it or not. But now you can read any writer you want. So the average quality of writing online isn't what the print media are competing against. They're competing against the best writing online. And, like Microsoft, they're losing.
The "average" application on SourceForge may be derivative and just plain bad, but you don't install averages. You install applications. I don't use an "average browser", I use Firefox.

This holds true for games, too. As a Slashdot poster says of open-sourcing a game:

we will see millions of variations of modifications that will be incompatible with each other and that will bring down the quality of the game.

Some will be much worse than the original, probably few will add high-quality content, but some may become very good indeed.

Software, writing, games, in fact all digital media, tend toward a winner-take-all outcome. In the age of infinitely-reproducible goods, it doesn't matter what process yields the best average outcome. What matters is what process yields the single best outcome.