This is one of a bunch of posts on science and the Web. Start here for an overview.
A while back I discovered a central theme to my blog posts -- better science through superior software. I've now realized this is one part of a larger interest: to change the culture and practice of science to be more effective by being more open and online.
Witness:
Posting grants publicly -- http://jabberwocky.weecology.org/2012/08/10/a-list-of-publicly-available-grant-proposals-in-the-biological-sciences/ and http://ivory.idyll.org/blog/grants-posted.html
Relatively few scientists do it, but by posting grants online you can connect to a broader community and attract collaborators that are interested in what you plan to do. It's hard to see a downside, frankly, unless it's that someone might steal your ideas. But most scientists have plenty of those; it's time and attention that are difficult to find.
Posting papers to arXiv -- http://pathogenomics.bham.ac.uk/blog/2012/10/all-the-cool-kids-are-on-arxiv-and-haldanes-sieve-why-you-should-be-too/ and http://ivory.idyll.org/blog/science-f-yeah.html
arXiv is a great way to communicate results, and I bet it will lead to higher citation rates and bigger impacts for papers. The quantitative side of biology is starting to realize this.
Pushing and enabling others to play on arXiv with me -- http://ivory.idyll.org/blog/rohan-on-weird-patterns.html
This kind of conversation is the future of science, and I feel privileged to have been involved.
Changing peer review and publication to better support remix culture -- http://ivory.idyll.org/blog/blog-review-criteria-for-bioinfo.html and http://ivory.idyll.org/blog/vms-considered-harmful.html
It's not enough to be open - you have to encourage building off of other people's ideas. This requires reproducible research, and review criteria that make sure that the research is reusable.
Using cloud computing and online course materials to both drive education and support better transmission of materials -- http://ivory.idyll.org/blog/ngs-course-with-aws.html
In some cases, these course materials are the best documentation of process there is. Strange.
Better software engineering in research -- http://ivory.idyll.org/blog/anecdotal-science.html and http://ivory.idyll.org/blog/automated-testing-and-research-software.html
If you're not writing good software, what's the point of doing it in the first place?
Open and reproducible research that can serve as a foundation, but only when it's ready http://ivory.idyll.org/blog/replication-i.html and http://ivory.idyll.org/blog/blog-practicing-open-science.html
I'm not a big fan of open-everything because I think the signal-to-noise ratio is a bit low. But once you're sure it's correct, go for it!
Trying to lead by example -- http://simplystatistics.org/post/29620679415/interview-with-c-titus-brown-computational-biologist
Pushing open data in large collaborations -- http://freethoughtblogs.com/pharyngula/2012/10/29/the-cephseq-consortium-has-a-strategy/ and http://ivory.idyll.org/blog/cephseq-cephalopod-genomics.html
Many fields simply do not get the idea that data is most useful in context -- in biology it's almost a necessity. Drafting guidelines that both protect individual scientists while making their data usable more broadly is difficult but important.
Scientist training -- http://www.nature.com/nmeth/journal/v8/n7/box/nmeth.1631_BX1.html and http://third-bit.com/blog/archives/899.html and Software Carpentry.
Scientists need to grok computing. Most don't.
--titus