Clouds and Waves … computing at the interface

So this has been an action-packed week in my science world:

On Monday, I discovered that I could potentially access the Amazon cloud (EC2) with an academic grant. http://aws.amazon.com/education/

This has a lot of potential and will allow us (hopefully, if the grant is successful) to do computing along the lines of more famous projects such as SETI@home and Folding@home. Although we will not be discovering either aliens or cures for cancer, we do hope to make some contribution to the scourge which is tuberculosis, a disease that kills millions of people each year, and is only set to get worse as the incidence of co-infection with HIV and multidrug resistant strains increases.  At any rate AWS shows great potential so we’re keeping our fingers crossed.

The other cool thing this week was receiving my Google Wave developer preview account.  Google wave is a tool set to revolutionise internet communications through new ways of utilising existing protocols.  Whilst it may not be for everyone (many people don’t like google’s email interface for example, and emphasis on searching, and this is similar), one place it should certainly take hold is in collaborative editing processes.  Hopefully we will be able to adapt the interface to electronic lab recording interfaces (for which it does seem naturally suited), and collaborative paper preparation and publishing (for which it is also ideal). I only bemoan the fact that the public preview is only due on September 30th, far too late really to integrate it into my second year undergraduate course.  And with the current instabilities of the developer preview (she says, to the sound of crashing waves …) it is going to be touch and go whether it would be safe enough for assessed materials.  So perhaps this can be a plan for next year, when people are then used to the new interface.

The next step is of course to get a robot up and running to start to integrate chemistry with the wave.  Already the ability to plot graphs (code), integrate chemical information from chemspider, and incorporate citations from services such as citeulike are available (see some of the things that happened at sci-foo).  Chemical converters are already under development … perhaps a periodic table is next (which will insert specific elemental details on request).

I am hoping to perhaps look at a way to grab files from finished computational chemistry runs (eg gaussian, gamess, nwchem etc) and convert at least some of the output into small graphics.  This seems like it may be possible using a mixture of the openbabel webservice (for translation) and molecule information retrieval services like the Royal Society of Chemistry‘s chemspider or NIH/PubMed’s PubChem.  Anyway, I need to learn python first, so watch this space.

Clouds and Waves

An alternate view of Peer Review?

Or why I’d rather be using google wave ….

The peer review system and methods of assessing science worthiness have been receiving quite a bit of attention recently.  Do we need peer review?  What purpose does it serve?  How can we have effective peer review that doesn’t carry an agenda beyond its primary purpose? Who should pay for the time and effort peer review requires?  How valuable is it anyway?  Shouldn’t we just let the web sort it out?

A number of these points have been eloquently addressed, with a couple of examples as starter points below:

Peer review of grant proposals …

The evolution of scientific impact …

So personally, what do I, as a jobbing scientist, get out of peer review, other than a backlog of extra work from the articles/grants in my inbox that I am expected to pass judgement on as to their scientific worth?

Well, firstly it is a great chance for me to see what other people in my direct field are up to, prior to normal publication, based on my listed interests at the appropriate journal/grant body.  That is, I gain some inside information on my ‘competition’.  The value of this will vary depending on the nature of how the science is communicated.  Firstly, if I am familiar with the person, I probably have already seen the preliminary details of the work as part of presentations or possibly informal discussions at a conference, provided I have been lucky enough to attend.  However, if we move to a more Open Science framework this advantage for journal articles will be lost, as all the details should be freely available to whomever wishes to see them.  The big advantage here is that I can have a preview of work from people whom I may have never met (but may now like to), and see work which I may never get to review before publication.  For grants, it is the ability to see ideas in action.

Is this window of pre-publication valuable though for journal papers?  In my field, many articles are turned around within weeks and are available on the web.  Sure, if I were waiting for the in-print version I might have to wait up to 2 years to see the hardcopy.  The inside track becomes only valuable in hyper-competitive, fast turnaround science, but even this much less so in a world where the press release is out even before the science has gone through peer review.

Increasingly the peer review is more for the benefit of others in this regard, whereby I validate their ideas and concepts on the basis of the evidence they have presented, as an expert validated by an external body (eg the journal editorial board).

And this is where I feel the main benefit for peer review comes for me – someone to look over my stuff, before I commit it worldwide and flag up any glaring errors/omissions/brain-fails or at least remove a couple of typos I missed.  But the question then comes, is why do I need such an elaborate mechanism for doing this as the current peer-review system?  In fact, what I would have preferred to do, for example for this blog post, would have been to type it up in google-wave, add a few of my trusted and (hopefully) interested friends, and finally, once they had given it the ok and made suggestions of points that could be either expanded or rewritten, added the bloggy tool to publish.  Alas, though, I haven’t yet got my account, so instead those reading will have to have everything unedited, and I will feel nervous about pushing the publish button.

But is my view reflective of the general scientist?  I have no idea, and I suspect that it will depend on the view of what publishing is for.  For me it is about having something that people can tangibly get hold of (eg print off), replicate as necessary from a detailed description, and as a paper-trail for finding other related science.  It is unclear to me that this cannot all be delivered through online and complete archives of scientific endeavour (ie open science) and in a form that is much more effective and accessible than current print media or many journals.  When added to a blog commentary, the expert interpretation is also available too.   And fundamentally this will mean that publicly funded science is available to the public – an obvious point raised by many others.  The main challenge will be in having this as acceptable evidence of output; for grant bodies, research bureaucrats, non-web savvy colleagues (yes, there are still quite a number in academia), job applications etc, instead of the usual and comparably succinct list of publications.  And the type of review needed will need to be demonstrably objective, because the outside won’t trust the ok to have come from my friends (which is the case in some fields, but if the reviewer is anonymous and selected by an external source this does not seem to matter).

So in the interim, what does peer review actually provide me with?  ’Objective’ confidence?

Confidence that my science is being done appropriately, and that my conclusions aren’t completely insane.  And confidence that what I read that is not in my specialist field is also not entirely crazy and has been validated objectively by a nominated source.  Will the web replace this?  Will we perhaps, with better communications, revert to a more old-school system of assessing work?  Time will tell, but the potential is there …