Original URL: https://www.theregister.com/2005/12/09/mattson_open_cookie/

The sophist and the open source baking farce

Don't take the biscuit

By Ashlee Vance

Posted in Channel, 9th December 2005 22:58 GMT

Into the Valley One of the biggest threats to open source software has arisen from a most unlikely place - the food and beverage industry. Steve Gundrum, CEO of food engineering house Mattson, has teamed with sophist and celebrity author Malcolm Gladwell to bastardize the fundamental concepts behind open source software development, turning the OSS idea into nothing more than another term for inefficient collaboration.

Worse, the two gentlemen have opened a huge front for proprietary software advocates to rally around the cry that open source is just a trendy fad. And they did all this simply by baking a cookie.

In 2004, Gundrum set out to do something he claimed was revolutionary. Instead of having food scientists, chefs, marketeers and executives guide a new food product to development in the traditional way, Gundrum devised an experiment to try three "new" methods for creating a cookie. Gundrum, a software fan, centered one method around "open source" development, another around "extreme programming" and a third around more traditional, hierarchical product development. The goal at hand was for the three groups to come up with a cookie recipe that would deliver a yummy product that had 15 per cent better nutritional value than the average of the best-selling, major cookies.

An even larger goal was to create a quicker system for bringing new food and beverage products to market. According to Gundrum, it takes a little more than two years for the average new product to go through development, testing and then hit the store shelves. Food makers have relied on trusted models for developing products and don't have the freedom to tinker because of the constant push to roll out fresh edibles.

All of this seems fair enough until you start digging into it a little deeper. It's here that open source advocates should worry, as new businesses try to hijack a key philosophy to the software industry. The business types don't understand how open source software really works and when they try and fail to apply it elsewhere, it makes the FOSS crowd look inept.

Gladwell didn't do the OSS crowd any favors by publishing a puff piece on the "open source cookie" project in the New Yorker. The first clue that Gladwell's knowledge of open source is shallow is that he misspells Linux creator Linus Torvalds name throughout the entire piece, referring to him simply as Torvald, and classifies him as Norwegian when he's a Finn.

In the context of Mattson's Project Delta, "open source" development simply meant assembling an elite group of food scientists, chefs, marketing experts and food packaging gurus and having them collaborate on the creation of cookie.

"The roots of Project Delta are really in the development of software," Gundrum said, during a speech yesterday at the PARC facility in Palo Alto. "I started looking at code and thought to myself, 'Oh my god, this is a recipe.' You would essentially compile all the ingredients of a recipe and then you would send that mixture to a piece of hardware and you would end up with the product.

"Obviously, I had to take some license. It doesn't translate in a straight line at all."

No, it doesn't.

In fact, neither Gundrum nor his hagiographer Gladwell ever explain how a group of experts collaborating on a project resembles open source software development. They suggest that each expert would provide input in the necessary areas and share ideas. They compare this to a more traditional, hierarchical model where one person works on part of the project and then passes it up to the next level and so on until the end result is reached. The obvious problem, however, is that the hierarchical model is a myth: people have been collaborating on projects in a non-hierarchical way since the dawn of time, and just because you call this open source doesn't make it so. Microsoft has experts that work on projects together too. So that isn't much help.

Perhaps sensing weakness, Gundrum tried to extend the open source metaphor in a direction that made more tangible sense. At the end of Project Delta, Mattson would turn over the cookie recipes to its clients and make the product formulas open for free industry use. In exchange, cookie makers using the recipes would need to turn over 1 per cent of their sales to a non-profit started by Mattson that would use the money to feed US citizens living below the poverty line. In addition, supermarkets selling the products would need to waive their exorbitant stocking fees whenever the non-profit Helpings-branded goods were involved. While noble enough, we'll get to how this plan undermines open source as well in a bit.

On the extreme programming front, Mattson adopted the ideas of programmer Kent Beck to create small, nimble product development teams. For Project Delta, this meant having just two "food engineers" and a team coach. Again, we're left wondering how calling this "extreme programming" really makes sense. In the old days, you would call it a team. Just ask Hewlett and Packard, Ben and Jerry or the Disney brothers.

For the third method, Mattson employed a much more traditional "managed process" structure that had nothing to do with software at all - not the other two have more than branding to do with software either.

To hear Gundrum or Gladwell tell the story, this experiment was a marvel of human interaction that unearthed new truths about the nature of individuals and groups.

In his story, Gladwell spends pages giggling over how difficult it is to create a cookie that tastes good and is still good for you.

"Over the years, there have been many attempts to resolve these contradictions - from Snackwells and diet Oreos to the dry, grainy hockey pucks that pass for cookies in health-food stores - but in every case, flavor or fluffiness or tenderness has been compromised," he writes. "Steve Gundrum was undeterred."

Nor was Gladwell deterred by the challenge of producing a vast quantity of banal prose, something he excels at. This is a man with a 6,300 word disclosure statement on his website.

Gladwell hyperbolizes the "open source" group by calling it the Dream Team.

"It is quite possible that this was the most talented group of people ever to work together in the history of the food industry," he writes. Although, the food industry stretches back quite far, and the likes of Louis XIV might object to Gladwell's proclamations.

The most amusing part of both Gundrum and Gladwell's accounts is that they promise that the "open source cookie" experiment will produce revelations about how product development can be done more efficiently. But unfortunately, that didn't happen as Gladwell expected.

"I was expecting open source to run away with it," Gladwell said in a podcast. "In fact, I didn't even think it would be close. The thing that surprised me was . . . how well the other two groups did."

After surveying 300 families, Mattson found that the favorite cookie was the strawberry-cobbler concoction developed by the traditional managed process group!

This unit had stuck with a single design for 90 per cent of the project but then had to take a different course at the last minute because it couldn't meet the health objectives. In the end, the group came up with a stunning new way of adding flavor to the cookie and won the challenge.

The open source group tried 35 different approaches before settling on an oatmeal, caramel, pecan cookie that finished second. And the extreme programming unit developed its cookie at the quickest clip but finished way behind in third place. Why? Well, because the cookies were over-baked and rushed out the door without a second try.

"We chose to send them out anyhow," Gundrum said. "Malcolm had to get his story out."


So, what we seem to learn is that experienced folks working with a proven method won the contest, while the scatter-brained Dream Team tried a ton of ideas and ended in second, and the quick working small unit was screwed by a reporter who ruined the entire experiment he was studying.

(Nowhere in Gladwell's 6,300-word disclosure statement does he disclaim ruining this experiment.)

None of this would hardly matter were it not for Gundrum and Gladwell running around the country telling groups about this amazing new "open source" method of food creation. Their propaganda has seemingly sensible people groping to find deeper meaning here.

"Although Steve Gundrum's experiment was hardly rigorous or conclusive, it does touch on a real issue with open source," writes InfoWorld's Jon Udell.

"Striking a balance between an architecture of participation and an architecture of control is a central concern for all kinds of product development. In the case of most successful software applications that ride above the commodity stacks, the balance we've struck so far typically locates participation outside the API boundary. Whether that can change, and whether it should, are two very interesting questions."

Architecture of participation?

During his talk, Gundrum went on to speak about how opening up the license of these cookies could create a new business model for the industry. Again, there is a lot of fact shielding going on at this particular API boundary.

Let's not forget that these cookies will only make it to market because the grocery chains wave their slotting expenses required to get goods on the shelves. In addition, the charity aspect of the project would seem to be a rather unfair variable added to the fierce experiment in capitalism.

In the end, all Gundrum has managed to do is promote open source as a cute gimmick. It doesn't work very well. Hell, it's not even open source. It's just people working in a group. But it sounds good for a staid food company to be on the edge with the software powerhouses.

It's also a heck of lot of fun to play around with opening up your licenses when you're trying to feed the needy as another gimmick. Never mind that entire nations, groups of citizens and corporations take the proprietary versus open source battle very seriously and that many of the items being debated can have a real impact on how business is done.

Don't make a mockery of open source by getting the concept wrong, and then wrapping a disaster of a business model around it.

Even worse, Gladwell seems to want to admit that collective intelligence proved less impressive than he hoped - but can never quite bring himself to make this confession. Instead, he fumbles in and out of negative language dulled by his constant Dream Team admiration. All the while, he co-opts the open source movement to try and give an otherwise boring piece more significance.

Gladwell fails to realize that Linux kernel developers are professional, not amateur, and that kernel development is hierarchical with strict judgments made as to an expert's skills in a particular area. It isn't some kind of free-for-all.

For some reason, Gladwell is paid tens of thousands of dollars for flimsy stories about companies taking "unique" approaches to solving problems when they aren't actually doing anything of the sort. In his corporate pep-talks, Gladwell describes a utopian vision that can never be realized, while doling out pithy gems such as, "ability plus experience equals expertise."

Who would have thought it?

Such a cheerleader does little to help the open source crowd, which must deal with reality every day, and face hard-nosed opponents who will cheerfully use Gladwell's drivel to undermine it.

You'll find Gladwell's disaster here and Mattson's self-congratulatory web site here. ®