I’ve just finished reading Nicholas Carr’s latest article, The Ignorance of Crowds, in the latest issue of Strategy & Business. (And it’s refreshing to note that Booz Allen Hamilton don’t appear to have constructed the traditional consultant paywall. For this relief much thanks.)
Carr makes a number of points succinctly and eloquently, and peppers them with relevant quotes:
Opensource tends to be “extraordinarily powerful way” to improve things that exist, but tends to be “less successful at creating exciting new programs from scratch”, and as a consequence “it’s an optimisation model rather than an invention model”
He quotes University of Michigan professor Scott Page as saying “when solving problems, diversity may matter as much as, or even more than, individual ability”, and extends this theme by saying “What an unorganised, fairly random group of people provides is not just a lot of eyeballs but a lot of different ways of seeing”
Eric Raymond, who is quoted extensively, is quoted as saying “debugging is parallelisable”;Â Carr builds on this, saying “all the debuggers have to do is to communicate their findings and fixes to some central authority, like Linus Torvalds
Based on optimisation not invention. Demonstrating the power of diversity. Depending on the capacity for parallelisation. Principal characteristics of opensource. All well-made points, not all new but definitely well structured and articulated, and worthy of further analysis.
Carr then goes on to look at what he considers some key limitations, again derived at least in part from Raymond’s work. He claims that “peer production works best with routine or narrowly defined tasks….. not well suited to a job that requires a lot of coordination among the participants… the crowd’s size and diversity would turn from a strength to a weakness, and the speed advantage would be lost”. He goes on to suggest that opensource “works best when the labor donated or partially subsidised”. He then asserts that “the opensource model — when it works effectively — is not as egalitarian or democratic as it is often made out to be.”
There’s a lot of good stuff in the article, go read it for yourself. I even agree with most of it, and will spend time masticating over the article at leisure. But. And it’s a mid-sized but.
I think there’s one key aspect he misses, or rather doesn’t do justice to. And that is this:
We shouldn’t dismiss lightly the propensity for opensource to innovate, to augment innovation and to accelerate innovation, for the following reasons:
The diversity inherent in the crowd creates long-tail effects, and this causes the bazaar to come up with stuff that the cathedral wouldn’t consider; in cases where the cathedral does consider the innovation, the bazaar is often faster and cheaper; and finally, while tight coordination by central authority seems a worthwhile thing, we should not forget the number of camels designed by committees.
In fact that’s one of the key stanchions of opensource communities. They don’t do camels.Â
Here is my rebuttal to Mr. Carr.
http://www.krishworld.com/blog/general/nicholar-carr-has-got-it-wrong/
I suppose it depends on one’s definition of innovation. If you take it to be the original idea then, in most cases, the innovators are few in number and that is at odds with the bazaar model. The bazaar may or may not optimise the innovation, but I’m not convinced that it originates the innovation.
Krish, I have not been able to get to your comment. Maybe my blackberry truncates the URL, but it looks okay yet doesn’t work.
I think if you see innovation not so much as a novel idea, but as a successful implementation or execution of an idea, then “opensource” is in a good position to be ‘innovative’, because the execution is a) more likely to happen, and b) more likely to succeed
I find it interesting that one of Carr’s key points is that Open Source needs good aggregation (or central management) in order to work properly. That is also one of the key points of Surowiecki’s Wisdom of Crowds. Surowiecki iterates over and over how important it is to have an aggregation method that maintains the value of the crowd without biasing the end product toward the will of the moderator or people doing the aggregation.
In short, the need for good aggregation or integration of the crowd does not mean that the crowd model is broken. I believe that aggregation issues are going to become increasingly important. Not just in IT and Open Source but also in Risk Management and Finance.
On the Wikipedia comments by Carr: So the measure of an online encyclopedia is the length of articles that he deems important vs. those he deems unimportant? WTF? All users of Wikipedia, or any online resource, should understand the advantages vs. limitations of the resource. These advantages and limitations are hardly captured by the length of the Homer article vs. the Flintstones. I want a bumper sticker that says “Wikipedia: It’s one resource out of many. Deal with it.”
I think most of Carr’s analysis of Wikipedia is plagues with the cognitive bias of Focusing Effect (prediction bias occurring when people place too much importance on one aspect of an event (or resource); causes error in accurately predicting the utility of a future outcome.) I knew those biases you posted would come in handy often!
-JD
plagues=plagued
too bad I only have a spell checker and not a rant checker.