Document information
Open Notebook Science Using Blogs and Wikis
Correspondence: (Login to view email address)
- Drexel University
- Document Type:
- Poster / Presentation
- Date:
- Received 11 June 2007 20:23 UTC; Posted 12 June 2007
- Subjects:
- Chemistry
- Abstract:
The first half is a summary of how the Bradley group at Drexel University is doing Open Notebook Science with the UsefulChem project to synthesize and test novel anti-malarial compounds. Graduate student Dave Strumfels’ code to compute kinetics from JCAMP NMR reaction profiles is then highlighted. Finally screenshots are shown of a building on Nature Island in Second Life where Beth Ritter-Guth and Eloise Pasteur have helped to set up a poster room with NMR spectra, molecules and an organic chemistry quiz that can be activated by clicking on an obelisk.
- Presented at:
- American Chemical Society, 27 March 2007
Discussion
- Votes:
-
12 votes
- Comments:
-
15 comments
Your analogy between open science and open source software development is right on the target. In regards to someone running away with your findings. This has not been a problem so far in open source software community simply because it would be hard for someone to run away with something that’s public. It’s like trying to steal the Mona Lisa and then put it on public exhibit in another museum. A few days ago I was pondering why the majority of researchers still follow a closed scientific approach (aka Cathedral research mode), some of the reasons probably belong to this group.
What about the fact that the time and money involved in a research project is no small investment? This cuts in both directions 1) open access to research findings may help save researchers from repeating others mistakes 2) more nefarious users will try to compile datasets and information to create “new” findings that they did not actually validate.
Don’t get me wrong I am in favor of this push to open science I believe that the long-term results will be very positive. Like all change there will be reactionaries against it, especially if it threatens the comfort and familiarity that established researchers have come to enjoy.Mario – the argument you make (and the Mona Lisa analogy you use is great) that it makes it more difficult to steal something that is public is very strong. If anyone can Google sentence fragments it will be very easy to demonstrate plagiarism. The current system of submitting your proposals and papers in private to your peers (aka competitors) can’t be more secure than transparency.
Craig,
I think the key point with Open Science is that the default license requires attribution. If someone uses or re-mixes my data, observations or ideas and I get credit for it, isn’t that a good thing?Egon,
The political consequences of Open Science are certainly going to be interesting to follow. (I’m glad we have yet another communication channel to discuss this!)Could you give us some links to the relevant pages?
Also, this is probably my own bias, but I think the Open Science approach would be particularly useful in neuroscience, with things like metastudies with fMRI data. However, one issue which comes to mind: When dealing with data from human subjects, how can one effectively share that data while still protecting privacy of the subjects? (For example, a subject may be identifiable via irregularities in their brain structure)
I think there is a subtle difference between the open source community and science community. The open source community is generally made up of enthusiasts who are eager to exchange and share ideas and code. Look at codeproject as an amazing example of the enthusiasm of the open source community, over 4,000,000 subscribers and 1000s of articles and code, all submitted by enthusiasts for no apparent gain. Science on the other hand has been heavily politicized, it involves money, careers, tenure track, etc. The science community has a completely different set of forces at work compared to the open source community.
The other problem we have is that we are very much wedded to the traditional journal peer-review model where a few gate keepers decide on whether a piece of work sees the light of day or not. As an alternative I very much like the codeproject peer review model where contributors submit an article/code and the community then votes and comments on the article. All articles are published under this model. As a reader, one can rank the articles by popularity and/or topic and just browse the top 25 percent. I am therefore very enthusiastic to see Nature deploy something similar in the form of Natureprecedings, it lets the community as a whole decide what papers are interesting and what are not. With the internet, publishing an article carries a negligible cost unlike the traditional model.
One possibility I would like to suggest to Nature is that every quarter or so why not publish in paper format the top ranked papers on the site?
Herbert Sauro, what you are proposing sounds like a “Digg” for science (“contributors submit an article/code and the community then votes and comments on the article”). I am a bit skeptic about it (on Digg, inaccurate articles are promoted sometimes). But last proposal about printing top ranked papers looks good and it is worth a try.
Sebastian, why are you concerned about inaccuracy? What is inaccuracy, anyway? Einstein was thought inaccurate by Planck in 1905, but by 1912 that had changed. Had Einstein’s 1905 work changed, or had Planck?
This focus on inaccuracy, not trusting professionals to be able to sort good ideas from the chaff, has led to today’s predicament, where he have an almost total suppression of radical ideas in the print media.
I started publishing in 1974, and I have seen the changes occurring slowly and steadily. Peer review has not evolved to encourage breakthroughs, but to discourage them, to limit the rate of change to that which randomly chosen scientists can deal with. Not the best scientists in their field, who have shown themselves competent to assimilate new concepts, but scientists purposely chosen by an editor as “peers.”
I have had the opportunity to thoroughly discuss my own work with my West Australian academic colleague, Nobel Laureate Barry Marshall. But Barry does not ‘do’ peer reviews. He has other priorities. I don’t accept peer review requests very often, either. However, Barry’s insight into how the publishing process worked (or didn’t work) for him, has been a huge help to me in getting my own work into the public eye. As have the insights of Laureate Peter Doherty (whose book “A Beginners Guide to Winning the Nobel Prize” should be on everybody’s ‘must read’ list).
You know, there have been no major breakthroughs in the Health Sciences for some time. The Salk vaccine was the most recent, IMO. Here is our current status: the US CDC predicts Arthritis will cripple 25% of the US adult population by 2030. Half of our seniors will have diabetes by 2050. Obesity is out of control. Where are the breakthroughs when it really matters?
The answer is that at least some of the potential breakthroughs are discarded on the floor of Journal editors. For example Boyana, at Nature Molecular and Structural Biology, who opined that a review proposal I sent her was just too controversial for NPG. Even Nature Precedings itself, refusing to publish a poster here which had already been presented at “Days of Molecular Medicine 2006” (which was a Nature Publishing Group conference co-sponsored with Karolinska Institute and MGH) because “our decision is based upon your clinical claim that vitamin D supplementation is immuno-suppressive.” Well, I am sorry, every expert on VDR transcription knows that, those in PhRMA who are working on Vitamin D analogs know that. Did anybody bother to contact peers who really KNEW what they were doing?
The breakthroughs will therefore come via other venues. Wiley picked up the review that NMSB turned down. Open Publishing has enabled a group of like-minded innovators to gather around us, and move forward as a non-profit organization, without the help of traditional scientific publishing. I personally like the idea of a ‘meeting place’ in second life, because I guarantee that 90% of the folk who currently write peer reviews won’t even know how to find the location when they are alerted to it on their Blackberries…
Neil,
As for a link to the UsefulChem project:
http://usefulchem.wikispaces.comConcerning human subjects, there would certainly be privacy laws to deal with. Luckily for sciences like chemistry that is not an issue.
Herbert and Sebastian,
Concerning the value of ranking systems, I think it depends on how the publication channel is used.For Precedings I see that there are at least two types of readers: browsers and those who come here via a search (i.e. on the site or Google) looking for something specific. A ranking system is probably a good idea for the browsers but I don’t see that it has much value for the Google searchers who just want answers (or at least ideas to solve their problem).
Maybe the editors can correct me on that but that is how I see Precedings, at least at this time.
Trevor,
Part of the problem you describe with the peer review system is that it really can’t spot most inaccuracies – or just simply mistakes -because there is not enough redundant information in a traditional article to even check for self-consistency. (Although some people, like Peter Murray-Rust, are coming up with some automated solutions to check for consistency in spectroscopy data as they appear in the article.)Articles tend to aim for conciseness – if the author says equimolar amounts were used in the experiment, it is stated once and you have to take their word for it. No reviewer can validate that.
That is the beauty of Open Notebook Science – because anyone has access to all the data, at the very least one can check for self – consistency. If equimolar amounts were used then a cascade of consequences should follow from the data (e.g. spectroscopic) that are subsequently obtained. If there is a problem then engage the researcher so that everybody can get at the truth.
Jean-Claude,
I agree wholeheartedly about the utility of open-notebook science. You are preaching to the choir here. Wherever possible, I have been using an ‘open’ approach in my own research for nearly 30 years :)There are two problems I have found. The first is that nobody seems to looks at my open data. ‘Supplementary files’ (eg PDB structure files) are rarely accessed. How do I know? Well, most of them have errors, and nobody ever asks me about those errors (small, errors, but errors…)
Secondly, in the Health Sciences, it is tough to be totally open with data. It is usually possible to trace the patient if enough circumstantial data is given about them. This makes the FDA, in particular, very concerned about privacy issues. I know, because I have been trying to get them to accept open study results, and it is very, very, tough. Consequently there seems to be a limit on how far any attempt to open up research data can be extended to the less precise sciences. We have the FDA commissioner saying that they want to be able to “data mine” raw data, but his staff scared to death about showing their ignorance while trying to do that. LOL. Oh well, good luck with your endeavors, in any case, this sounds like a very worthwhile project :)
Trevor,
Yes I know that I am probably preaching to the choir for the most part but there may be people following the discussion who have not seen all the arguments.I don’t know what number of visitors you need to be considered “successful”. Raw scientific information is really specific and unlikely to be useful to vast numbers. If one person (or machine) uses the information on our lab notebook wiki, that is one more than would if the notebook were private. Right now I’m happy with the type of web queries that are leading people to the wiki.
Good luck with your own efforts with Open Science! Working with human subjects it must be very difficult to push for openness. Luckily nobody cares about the privacy of molecules :)
- (Login to share with a colleague)
Additional information
- License:
- This document is licensed to the public under the Creative Commons Attribution 2.5 License
- How to cite this document:
-
Bradley, Jean-Claude. Open Notebook Science Using Blogs and Wikis. Available from Nature Precedings <http://dx.doi.org/10.1038/npre.2007.39.1> (2007)
- Version info:
-
Other versions of this document in Nature Precedings
None.
Other versions of this document elsewhere on the web
Egon Willighagen on 13 June 2007 10:14 UTC
Politics make open science difficult (“don’t want that department with lots of money to run away with my findings”), but I believe this may be an important step forward in scientific practices: open source has triggered a speed up and more efficient development of software; it should be expected to do the same for science.