Skip to Content

Open Source Science?

The Boston Globe has a piece on the open-source science movement. Many readers here will have come across the idea before, but it’s interesting to see it make a large newspaper. (Admittedly, the Globe is more likely to cover this sort of thing than most metropolitan dailies, given the concentration of research jobs around here).
The idea, as in open-source software development, is that everything is out in a common area for everyone to see and work on. (Here’s one of the biggest examples). Ideas can come from all over, and with progress coming more quickly as many different approaches get proposed, debated, and tried out. I like the idea, in theory. Of course, since I work in industry, it’s a nonstarter. I have absolutely no idea of how you’d reconcile that model with profitable intellectual property rights, and I haven’t seen any scheme yet that makes me want to abandon profit-making IP as the driver of commercial science. Of course, there’s always the prize model, which is worth taking seriously. . .
Even for academic science, open source work runs right into the traditional ideas of priority and credit, and the article doesn’t resolve this dilemma. (As far as I can tell, the open-source science advocates haven’t completely resolved it, either). There’s always the lingering (or not-so-lingering) worry about someone scooping your results, and for academia there’s always that little question of grant applications. There have been enough accusations over the years in various fields of people lifting ideas during grant proposal reviews or journal refereeing to make you wonder how well a broader open-source system would work out, given the small but significant number of unscrupulous people out there.
On the other hand, maybe if things were more open in general, there would be less incentive to lift ideas, since the opportunities to do so wouldn’t be so rare. And if someone’s name is associated from the beginning with a given idea, on some open forum, it could make questions of priority easier to resolve. A subsidiary problem, though, is that there are people who are better at generating ideas than executing them – some of these folks, once unchained, could end up with their fingerprints on all sorts of things that they’ve never gotten around to enabling. Of course, that might be a feature rather than a bug: people who generate lots of ideas are, after all, worth having around. And over time, there might well be less of a stigma than there is now for someone else to follow up on these things.
The thing is, science has already been a form of open-source work for hundreds of years now. It’s just that the information has been shared at a later stage, though presentations and publications, rather than being put out there right after it’s been thought up or while it’s being generated. That’s why I always shiver a bit when I read about how long Isaac Newton waited before writing up any of his results – if Edmund Halley hadn’t pressed him to do it, he might never have gotten around to it at all, which would have been a terrible tragedy.
And it’s why stories like those told of physicist Lars Onsager strike me as somehow wrong. Onsager was famous for only publishing his absolute best work – which was pretty damned good – and putting the rest into his copious file cabinets (example here). (A related trait was that he was also apparently incapable of lecturing at any comprehensible level about his work). Supposedly, younger colleagues would come by once in a while and tell him about some interesting thing that they’d worked out, and ask him if he thought it was correct. Onsager would pause, dig through his files, pull out some old unpublished work that the new person had unknowing duplicated, and say “Yes, that’s correct”. It seems to me that you don’t want to do that, withholding potentially useful results for the sake of what is, in the end, a form of vanity.
And although I’m not exactly Lars Onsager, this is as good a time as any to mention that my summer student, who’s finishing up in the lab this week, has been able to generate a lot of interesting data, and that I’m going to be trying to write it up this fall for publication. Readers may be interested to know that this work is based on more ideas I’ve had in the vein of the “Vial Thirty-Three” project detailed here, so with any luck, people will eventually be able to see some of what I’ve been so excited about all this time. And that’s about as open-source as this industrial scientist can get!

9 comments on “Open Source Science?”

  1. JSinger says:

    The article jumbles together a bunch of different schemes, united only by dubious analogies to source code. Open Access (which I was skeptical about at first but am now mostly behind) is very different from public notebooks (which I still think are idiotic, but probably I’m just old).

  2. HB says:

    What supposedly gives Open-Source software development (where the term was coined) its strength is the opportunity for vastly collaborative development, where anyone with an internet connection can help out with research. The other advantage is that the source code is available so that any end-user is free to modify the software to fit their own needs.
    Most science, especially wet-work, can only realize one of the qualifiers of the “open-source” description, since it is true that results and procedures are generally disclosed and others may play with the “recipes” as needed to modify or improve something.
    Still, I think outsiders ought to leave the “open-source” term to the computer people and not mash it into yet another wishy-washy buzzword.

  3. CMC guy says:

    Interactions and collaborations between scientists of same and different disciplines need to be strongly encouraged and supported however this open source concept may not spawn greater progress IMO. Besides some of the factors Derek mentions there is lots of noise and clutter out there now in the literature and if everything is not gated by good peer review (which not such how would be handled i n open source system) I think much more low quality stuff will overpower the good stuff so that there so it will be harder to connect the right pieces together. Likewise based on personalities alone some scientists who are strong thinkers may not wish to get overly involved in public forums/debates so lose valuable inputs. We need to have more mentors, peers and/or atmospheres in the current system that can work with, at times guiding, then start implementing the ideas generators out there.

  4. Morten says:

    Derek’s presentation of the open wet ware makes it sound a lot like the journal Medical Hypothesis. I f************ hate Medical Hypothesis. It is the stupidest thing I’ve ever come across.
    Hey I have an idea! I’m not going to do any science to try to validate or disprove my idea but I will publish it in a peer reviewed journal to give it a shimmer of credibility and confuse reporters.

  5. milkshake says:

    sharing and openly discussing work in progress is possible only among friends, not among compatitors, and giving stuff out selflessly is possible only in the absence of freeloaders – Cordova spolied it for many people in his field.
    When the need to crank out the publications for the tenure review at top university becomes acute, some people are perfectly willing to backstab their friends and step over bodies of their students and colleagues.

  6. Dana H. says:

    I agree that openly posting in-progress work has many problems. But one important principle of the scientific method is reproducibility of results. So even if academic and government researchers won’t share their raw data before publication, they should make it available afterward. This allows others to do their own analyses of the data to confirm that the data imply what the original researchers say they do. The availability of cheap storage and the web makes this much easier than it used to be.
    Yet many researchers still drag their feet when it comes to making data available, even when journals and funding agencies have explicit policies requiring this. I don’t know if this is worse in climate science than in other fields, but Steve McIntyre (co-debunker of the statistical errors behind the “hockey stick”) has had great difficulties getting researchers to share their data, and getting journals and agencies to enforce their policies. E.g., see here: http://www.climateaudit.org/index.php?p=66
    Post-publication data sharing is the type of openness that would really help advance science. It does not guard against outright faking of data, but it allows errors of analysis and interpretation to be caught much sooner than with the traditional peer review process.

  7. Collaborative research and Open-source research, in my opinion, are differing concepts.
    Collaborative research is a model where multiple individuals can contribute data to a central repository where, by concatenating this data, a broad research study may be conducted.
    Open-Source research, however, is a model where one individual (or group of individuals) conducts a study while maintaining an open line of communication (for example, a publicly-accessible research diary/log) where their peers may contribute their ideas to the researcher who is actively conducting the work themselves.

  8. Tuomas Pylkkö says:

    While open science in some manner is a direct threat to traditional science, so was Wikipedia to traditional encyclopedias once. Now, of course, nearly no one would chose a traditional encyclopedia over Wikipedia. As communication technology advances, I believe (and hope, apparently) that science will eventually evolve into that direction.
    One of the miraculous things attributed to Wikiepedia is that no single company in the world could afford to produce such a massive project. Perhaps the same applies to, say, drug developement, as it seems that there are less and less companies that have the resources needed to do drug development.
    Perhaps eventually the general increase in the efficiency of information sharing will lead to the crumbling of the tradional political economy behind science and new rewarding systems not imcompatible with open source ideaology will emerge. If this happens, maybe science and industrial development will not cease to exist, but will become only less redundant, more efficient, transparent and democratic.
    Of course, no democracy has ever been fully democratic. Also, many braches of science are so narrow that they really cannot expect to use “the power of masses” in the same way as Wikipedia.
    Even so, there is a lot of information in science that no particular agent even deisres to keep secret (like rudimentray r code, or lab protocols). In those situations, the amounts of people that could comment on them might be more substancial, and the risks low. Journals should require these and post them as electronic appendices.

  9. Following all I’m subscribing to your feed i we do hope you write but once again really soon!

Comments are closed.