Open source refers to any enterprise where data (e.g. journal article, piece of software) may be modified by the relevant community and those modifications may be recontributed to the larger whole. There is therefore a very significant distinction between this and open access: open source data are mutable.
What is the advantage of such an enterprise? An open source biomedical research community that started in 2005, the Synaptic Leap (see below), has as its motto the quote "None of us is as smart as all of us" (This quote has an uncertain attribution. Some sources credit Robert Oppenheimer, some that it is a Japanese Proverb). The promise of open source lies in the massively collaborative efforts that may be undertaken, efforts that are effective only through the increased speed and scale of communication via the web. Stereotypically these contributors are unpaid volunteers, but a major survey of hacker activity found that 30% of those taking part in computer science open source projects were paid . Regardless, open source functions through the actions of many contributors from diverse backgrounds. There are two consequences of this. First, peer-review of the traditional kind (fixed duration, pre-publication) is not present - the peer review in open source is gradual and post-publication. Second, academic open source contributions tend to be of a higher quality and/or honesty than a cynic may suppose, a phenomenon known as the "gift relationship" .
Open source has delivered significant successes in recent years, and almost as much controversy. The number of people using Wikipedia may just be larger than the number who deny its usefulness. Wikipedia recently fared quite well in a head-to-head against the Encyclopaedia Britannica, but it is inevitable that such an enterprise contains errors (The original comparison was carried out by Nature, and a discussion of the ongoing argument between them and Britannica may be viewed on the Nature website) . As with all open source projects, the final product emerges gradually through a large number of iterative changes. In amongst reports of the thousands of spurious edits of the page for US presidents, it is worthwhile remembering that Wikipedia currently contains over five million articles, in 250 languages (1.5 million in English), has emerged within the last six years and is available free of charge. Britannica has been published since 1768 and contains approximately 120,000 articles in English in the online edition, and operates on a subscriber model. It is also worth remembering that the Oxford English Dictionary, when it was being originally compiled, relied on contributions from volunteers, including the notorious William Minor of the asylum at Broadmoor . Open source successes in computer science, such as Linux and Firefox, have been far less controversial, and have delivered high-quality products competing with those from major software firms.
Open source is also very active in Chemistry, though knowledge of these promising contributions is not widespread . Examples may be categorised as informal communities, chemical tools and collaborative research groups.
1) Informal communities
Blogs (web sites hosted by individuals, where readers may post comments) are informal environments where science can be discussed. Such sites will continue to multiply. While blogs have a reputation as not being serious science, useful scientific contributions do emerge. As an example, various experimental procedures have been described on long-running chemistry blogs Tenderbutton  and Org Prep Daily . Details of experimental procedures are described, along with extra content such as pictures of crystals from the experiments. User comments describe improvements and modifications. Anecdotal discussions such as these can only be useful to empirical scientists. That the web is so searchable means these discussions may easily be found. If a chemist has a problem with a reaction, they will typically ask their colleagues in the same group/lab/building for advice. Open source communities do exactly the same thing, but over much larger (geographical and social) distances.
2) Chemical tools
Several proprietary drawing packages are widely used, but open source alternatives exist. For example, free chemical drawing tools that are in development include Bkchem , and JchemPaint . Sophisticated tools exist for viewing molecules in three dimensions conveniently in web pages, such as Jmol . A related product, MDLChime, is free to use but not open source . A directory of open source chemistry projects may be found at the Open Science Project . The Blue Obelisk movement seeks to ensure interoperability in these applications by maintaining a set of open standards and, amongst other things, maintaining a list of algorithm specifications in chemoinformatics [32, 33]. Related tools of relevance to drug design are discussed elsewhere [34, 35].
3) Online Collaborative Research
As we saw above, informal blogging sites can be useful sources of advice on experimental methods. While the primary chemical literature remains the largest source of this kind of information, websites have the advantage that users may add or edit the information collaboratively. Organic Syntheses hosts open access, rigorously checked procedures , while Synthetic Pages is a website that enables informal user feedback . Open source protocols sites in the life sciences are also available .
Finally, several organisations have developed on the web recently that are looking into large, self-contained problems in chemistry, where profit-driven research has not delivered. UsefulChem posts the raw data on approaches to synthetic targets of interest, which currently include drug candidates for malaria . The possibilities of web-based collaboration in chemistry are clear here, in that scans of spectra and TLC plates, as well as video footage of reactions in progress, make it very simple for readers to contribute to the science.
We have recently started an open source collaborative group in biomedical research called the Synaptic Leap [40–42]. The organisation currently focusses on neglected tropical diseases, such as malaria, schistosomiasis and tuberculosis, and the site itself grew from the Tropical Diseases Initiative . The aim is to coordinate wide-ranging research projects in chemistry, biology and informatics. For example, a "gene wiki" concept is currently being explored as a way for the community to discuss and prioritise genes and proteins requiring further study. A current chemistry project on the site is the development of an enantioselective synthesis of the main drug used for the treatment of schistosomiasis, praziquantel [PubChem 4891] . The latter project is a perfect example of where open source can really deliver. The iterative improvement of the route to a drug that is of great importance to underdeveloped countries is of little interest to for-profit companies, but neither is it a priority for academia. We see open source collaboration as the only way to make research challenges like this tractable. Further, open source research communities could have great impact across drug discovery more generally as part of collaborations with more traditional big-pharma drug discovery programs .