I’m at Participate ’08 today in Portland, OR (part of OSCON) and it’s worth a few thoughts. Although this isn’t specifically about science, I think it’s very relevant to what we talk about here at Nature Network. Microsoft’s Open Source folks deserve real credit for this discussion, btw.
We’re starting with a case study on Threadless (here is good background). Threadless is a t-shirt company that builds on a core community of t-shirt designers – their users submit designs, their community votes on the designs, and the winning designs get printed and sold back in limited quantities to their community. This is a canonical example of user innovation, which is why we’re studying it.
I love the idea behind this. I’d love to see something like it in science. It seems to be an implicit design goal for open science, science 2.0, [cliche meme + science of your choice]. But I think we should unpack some of the preconditions required for user innovation and take a hard look at how much work is to be done to create those preconditions in science…
There’s just a ton of stuff that sits underneath Threadless, none of which was created originally to help people make and sell shirts.
There’s the basic stuff: the Web, the Internet, computers, and bandwidth.
There’s the software stuff: browsers, web programming systems, t-shirt design / paint software, shopping cart systems, vote systems.
There’s fulfillment stuff: on-demand small batch printing, low-cost shipping logistics.
There’s the user stuff: lots of people with the skill to use the computers and software, lots of people to vote and buy (thus, low barriers to entry as designer and buyer, and low transaction costs).
And then there’s the more metaphysical stuff, particularly the shared language of a community that is implemented in all the stuff listed above. The users were able to repurpose all those tools without a lot of fooferaw and lawyering to create this community.
I am not sure we have all of this for the sciences yet. We certainly have the basic stuff, but the software stuff is a long way away – most databases are unintegrated, and science software’s utility is well below the average utility of software for web users. Fulfillment of orders in the sciences is lousy unless you’re lucky enough to ask for a product that’s actually in a catalog – which kind of defeats the point of the user innovation thing, as 99% of tools for science remain un-cataloged.
The problem of a small number of users is a big barrier. There are so many millions of people on the web, and most of them wear t-shirts at some point, so the percentages dictate that some set of them will be into Threadless. But the way we do science remains a guild culture, so those millions of people tend to never interact with science the way they interact with shirts.
The lack of shared languages and names is a metaphysical problem that has dire real-world impacts. On obscure mailing lists, every day, people are fighting over languages and names in sciences. Until we get to a shared language, or even a DNS for science, Threadless isn’t going to port very easily.
In the end, my obsessions with openness and infrastructure, which can get pedantic at times, come from my desire to get us towards this participatory culture for science. I want Threadless for drug design. But I think that if we are not clear-eyed about the systemic process problems between today and that future, we will fail to get there.
Participatory culture is not accidental. It is the result of years of boring investment in infrastructure, the result of the slow accretion of tools, many of which were designed for other purposes.
We have to look hard at the foundations of science on the network and advocate ceaselessly for the necessary upgrades – to science as well as the network – that will allow us to get millions of new eyes on science, asking millions of new questions. Until that happens, we won’t really have a digital science culture. We’ll have simply made the old problems into digital problems.
More later.
Last updated:
Monday, 21 Jul
2008 - 17:03 UTC
OK John, here’s a model. I’m not pressing to see this implemented, but it is a relevant and possibly workable real-life example of a Threadless-style application in science.
You wrote:
“Fulfillment of orders in the sciences is lousy unless you’re lucky enough to ask for a product that’s actually in a catalog – which kind of defeats the point of the user innovation thing, as 99% of tools for science remain un-cataloged.”
I work for Gene Tools, where we produce hundreds of custom-sequence Morpholino oligos every week and, since they are custom-sequences, we don’t have a catalog. Our entry-level order is $400 for 300 nanomoles, far more oligo that most users need (but the smallest our synthesizers can reliably make). For a while, a company (Open Biosystems) bought out oligos, repackaged them smaller and resolg 50 nanomole aliquots. This didn’t generate the revenue they expected, so they gave it up (to the disappointment of many users).
I can see a Threadless model working for custom-sequence oligos, but there are some barriers. For instance, who would announce their interest in a gene they consider hot when that might result in being scooped? Who splits up the order and sends it to the participants? Can enough users become interested to buy up all of the custom oligo aliquots?
I like the idea of a democratic group arranging orders to keep their own costs down. I don’t think Gene Tools will set out to implement it, but there are certainly user groups, such as the Zebrafish Information Network, who could do so.
This is a fascinating example. Please contact me offline (my last name at creativecommons dot org) and let’s chat. I would like to get you into a user study if possible…
This NYT article may or may not be of relevance to this discussion
Craig, that’s very relevant. I’m going to muse later on why Innocentive doesn’t work so well for biology…
Hi John,
No I don’t expect it to work very well for biology, however this site is very straight-forward in it’s approach which is nice.
I look forward to reading what you think…