18 Jun 2007

omg web 2.0 is kewl

I'm not going to try any sort of systematic assessment of 'Web 2.0' and science, or even bother with a definition. This is just a stream-of-consciousness post, prompted by my wide-eyed wonder at the explosion of social networking and 'user generated content'.

Nature have completely bought into all this Web 2.0 malarkey - they've got umpteen blogs, Nature Network, Connotea, a new aggregator thingy called Scintilla, Postgenomic... they've even got a group on the "social notworking" site Facebook and an island in Second Life.

Social networking is a huge part of Web 2.0. Myspace is a mess, Facebook is more for play than for work, but The Scientist recently profiled the use of the 'Facebook for professionals' site, LinkedIn, by scientists. Another site that I've yet to look at properly is SciLink.

Social bookmarking is quite the thing. As well as Connotea, there's citeulike, another shared record of articles people came across and liked (hence the name). All the BioMed Central journals have feeds on citeulike, and we're collaborating with them.

A hybrid of social networking and social bookmarking is Stumbleupon. This is a brilliant way to find random websites that might of interest
- each user can give a thumbs up or thumbs down to every website they visit, and write a review of it.

You specify the topics you're interested in (music, film, science), and then hit the 'Stumble' button, and a page in that area appears. It is tuned by the sites you've said you like or dislike. As well as 'stumbling', you can also view the sites a user has rated and reviewed, which acts as a kind of blog of their browsing. This is a glorious way to waste time, but the same functionality applied specifically to scientific literature would give a great way to serendipitously browse the literature.

Another great 'overlay' to the web is WOT, or the Web of Trust. Very simply, you rate how trustworthy a site appears, generally speaking or specifically as a business partner, in keeping personal information, and as a safe site for children. A little icon appears by hyperlinks, green for OK or red if warning you that you are about to visit a site others have deemed untrustworthy or unsafe. Genius.

JournalReview gets around the barrier of some publishers not taking comments and having restrictions on letters by letting anyone comment on any article in PubMed. It's got a bit of a medicine bias, could still be more user friendly, and could really develop into a mature post-publication review process if it insisted on named reviews (or at least pseudanonymous reviews), and used trust metrics. And could someone use Greasemonkey to reveal comments from Journalreview when viewing PubMed abstracts, or even publisher sites? David Rothman has profiled a plethora of similar sites, in a great series of posts titled 'Digg for the medical literature'. Another good aggregator of information about scientific articles is the brilliantly simple PublicationsList - researchers can use it to, well, list their publications. They're going to need a name disambiguator pretty soon though. ScientificCommons is doing much the same, and I like their layout more.

One of the most notable Web 2.0 pages is Youtube, the video sharing site. In a wonderful example of circularity, an anthropology professor, Michael Wesch, has posted a 5 minute video guide to Web 2.0, The Machine is Us/ing Us -
I've had it embedded at the bottom of my blog for a while now. You can also find out why the banana is the atheist's nightmare. Youtube doesn't seem to be of much practical use to scientists, but there is a 'Youtube for science', the Journal of Visual Experiments, or JoVE.

Anyone can now get involved in scientific research without being a guinea pig in a drug trial. Distributed computing first came into the public consciousness with SETI@Home, and now similar applications are popping up in biology. I've got Folding@Home running on my laptop; it's helping to calculate the folding of supervillin at the moment.

No mention of Web 2.0 would be complete without referring to Wikis. Wikipedia is the best known and biggest - the encyclopedia that anyone can edit. The accuracy of Wikipedia, or rather the lack of accuracy, has received quite a lot of flak, in particular the entries about science. I'm almost left speechless by these criticisms. If you spot an error in Wikipedia, and it exercises you so much that you complain about it in public - just edit it! You don't even need to register! Sheesh. Another criticism is that it can be written by experts to be impenetrable to the outsider. As someone noted on that blog post, "don’t bitch when your encyclopedia gives you too much information. That’s its freakin job". Knowing something about serpins, I checked out the entry, and it's frankly excellent. The entry for antitrypsin is not so great, so I gave it a quick edit and flagged it as needing more references.

Of course, blogs are now reviewing articles post-publication. There is a Greasemonkey script from Pedro Beltrao that adds trackbacks to posts from Postgenomic to articles on Nature. I just tweaked it very simply to also be active on www.biomedcentral.com/* , and it works on our sites too. Cool. And now I see that Noel O'Blog has extended it to automatically work on PLoS, PNAS and BioMed Central. Thanks Pedro and Noel! Once you have it installed (which takes seconds if you have Greasemonkey), you can see it in action. Doesn't it look pretty?

By the way, the ungrammatical title of this post owes a debt to Ratcatcher's blog (which itself owes a debt to this Wondermark cartoon) and is reminiscent of a recent (and entirely useless) example of user generated content: lolcats. In case this craze has passed you by, it involves pictures of cats, with deliberately poorly written and humorous captions in bold text. Otherwise sensible linguists on Language Log have been led astray by this 'meme', and there's even a programming language based on the unique grammatical style of lolcats, lolcode. The web doesn't have much to top this.


Pedro Beltrao said...

Thanks for the mention on the Greasemonkey script. I would hope that BMC would collaborate with Postgenomic (Nature) to show trackbacks from Postgenomic directly from server side. I think that a lot of people that might be interested in reading blog comments on papers will not be aware that they can get them via Greasemonkey.
Also, the same way we can display the blog trackbacks, if BMC had a standard way of communicating the highly accessed papers (or even hits per paper) this could be used for example to include it in Pubmed via Greasemonkey.

Noel O'Boyle said...

The Postgenomic Greasemonkey script works best when looking at journal Table of Contents pages. Unfortunately, BMC is one of the small number of major publishers who do not put the DOIs on the Table of Contents pages. That's so Web 1.0!

BTW, the extended script is not just a trackback - it also has a popup containing the first few lines of the blog article.

postmeta said...

Just wanted to mention a web2.0ish(?) text-mining app we developed:
Visual Medical Dictionary

Please check it out if you get a chance.