Cameron Neylon

On Tuesday I was able to sit in on a conversation that is regularly held
within the Computer Science department at University of Toronto that
focuses broadly on what can computer science bring as a discipline and a
skill set to the sciences more generally. The conversation is lead by
[Steve Easterbrook](http://www.easterbrook.ca/steve/) so there is a
focus on climate science but we also roamed much more widely than that.

A key question that was raised, one which many of us have been
struggling with for some time, is how to describe and publish
descriptions of the progress of research projects in a way that provides
a route in for non-specialists. Blogs provide a great way to do this,
either as a generic journal with more or less detail, or as an overlay
over a more detailed open notebook. Jean-Claude Bradley's
[UsefulChem](http://usefulchem.blogspot.com) blog is a great example of
the latter, and the [blogs of Rosis Redfield's
group](http://www.zoology.ubc.ca/~redfield/who.html) an example of the
former.

The conversation was interesting for me in that it pinned down the idea
and necessity of creating a narrative. This contrasts with the kind of
(largely incomprehensible) detail found in a notebook which is usually
fragmented and often distributed. One of the things that researchers are
quite poor at in my experience is actually recording the why of an
experiment; the question of how it fits into the wider context. Again
blogs are a great format for doing this but where is the motivation?
Writing that narrative in any form is hard work, a classic example of
work that "takes me away from the bench" so how can it be justified?

One reason is that it raises the profile of the research, always an
important issue in today's research environment. But this is more
important to some people than to others. Another very valid reason is to
take personal notes, to create a personal narrative of what you are
doing and have done that you can return to and use as an index to your
own work. In a later discussion with [Alicia
Grubb](http://www.cs.utoronto.ca/~amgrubb/) she mentioned that her
supervisors insisted on her blogging about literature that she
collected. Equally taking notes in a bookmarking service could provide
the same functionality. But understanding the context in which you
bookmarked something is valuable. [Brent
Mombourquette](http://climatetooldev.blogspot.com/ "Brent's blog"), an
undergraduate student also demonstrated a nice Firefox plugin that he
had developed which captures and displays browsing history as a directed
graph which is an interesting tool to think about in this context. I'll
write more about some of the fabulous student demos I saw later.

For me though, the biggest benefit of making your research accessible,
is that it provides an entry route for new people to come in and help.
The story of [Galaxy Zoo](http://galaxyzoo.org) shows how by placing a
question within an understandable narrative you can enable people to
come in and help out. No-one is going to come in from the outside and
comment on my lab notebook unless they are already a specialist with a
specific question. I've often thought I should start another blog to
discuss more generally what is happening in my "real" research. Maybe
this is the time to do that.

But you can also take this one step further. There was a [debate in the
comments on the RealClimate
blog](http://www.realclimate.org/index.php/archives/2009/02/antarctic-warming-is-robust/)
a few months ago about making the details of data and analysis publicly
available. One real and valid concern was that denialists would dig into
the detail and mis-represent problems or mistakes to advance their own
agenda. Dealing with this kind of thing takes up valuable time, time the
average researcher, particularly if they are committed to taking time to
engage with a wider community, doesn't have. My question was whether you
could configure that public release of data and process in such a way
that even those who are working against you are helping you. If people
are searching your code for bugs then surely there must be a way of
taking advantage of that?

The argument that releasing data costs you time sounds compelling when
it comes from a researcher. But equally the same argument sounds
dangerous when it is made by, for example, a government. As Steve said
tongue in cheek, perhaps channeling some recently removed political
leaders, "clearly we're not going to release this data because it would
take us time to deal with public complaints, and that will costs
taxpayer. In fact, we're not even going to run the consultation because
that would cost money. It's much cheaper and more effective use of your
tax dollars if you just trust us to do the right thing." The situations
clearly aren't exact parallels and resources for communication are much
more limited in a research setting but it would be interesting to think
about parallel cases in different domains, such as government and
research, and how the domain effects the credibility of the argument. If
you believe in the value of sunshine as disinfectant for government data
then you need a strong case to argue the same doesn't apply to research
data.

But if you decide that you want to make that narrative public, or even
better the narrative along with the underlying data, it does take work
to make it comprehensible. As I've discovered recently such posts don't
translate easily into papers so making the argument that you can re-use
the text doesn't really work, at least for me. To make this worthwhile
you either have to be required to do it: [JISC](http://jisc.ac.uk) in
the UK basically requires that all funded projects have blogs; or you
have to believe and work towards the benefits it can bring you. In a
sense this is actually just recapturing the idea of the research
notebook that many historical scientists kept, and which make such rich
pickings for modern historians of science. Somewhere along the line we
lost that. There are lots of tools around that can help you create that
narrative, from clickstreams, records, environmental capture, but these
remain only an *aide memoire*. Making the story is something that will
probably remain a purely human activity. It is something that we seem
highly evolved to do and it remains the most effective means of human to
human communication. The computers can help, and they can provide the
detail to dig down into if desired, but the story itself will remain
ours alone for a while yet I think.


On Tuesday I was able to sit in on a conversation that is regularly held within the Computer Science department at University of Toronto that focuses broadly on what can computer science bring as a discipline and a skill set to the sciences more generally. The conversation is lead by Steve Easterbrook so there is a focus on climate science but we also roamed much more widely than that.

Messaggi di Rogue Scholar

Telling stories…