Comments on: “Startup weekends” in science

By: Pawel Szczesny

Pawel Szczesny — Sat, 03 May 2008 20:10:44 +0000

Alexei, Neil – I hope to organise something like this in a near future (after I convince somebody to host it). Hackathons are good example – maybe on their already established stature we could make this idea come true? We can call them hack-a-thought-sprints .

By: nsaunders

nsaunders — Mon, 21 Apr 2008 08:09:08 +0000

I guess the closest thing to this in bioinformatics at least is the "hackathon". I don't know if they run any more, but the OBF people use to have intensive meetings aimed at fixing up code for e.g. BioPerl, BioPython. I very much like the idea of a similar process aimed at addressing biological problems with defined outcomes.

By: Alexei

Alexei — Mon, 21 Apr 2008 06:28:03 +0000

This is a really great idea! I know it works well when you can physically round up the collaborators all in the same room and there is a sense of urgency to keep everyone focussed. What would you need though to make it effective online? There are many more distractions available … maybe setting everything up in a virtual world and being really careful on how to invoke that sense or urgency? You should call them “thought-sprints”

By: Ectropy » Blog Archive » Thought sprints

Ectropy » Blog Archive » Thought sprints — Mon, 21 Apr 2008 06:18:35 +0000

[...] is a cool idea for startup weekends in science by Paweł Szczęsny. Why not combine the ideas of code sprints and distributed teams for doing [...]

By: jan.

jan. — Wed, 12 Mar 2008 20:20:17 +0000

This idea very much makes me think of what students of a 2-week perl course have to do at CSHL (http://meetings.cshl.edu/courses/c-info08.shtml). During the second week, little groups of them have to tackle a project and try to finish it in just those few days. It is incredible what they can achieve.
It’d be nice if that could be done outside of a setting like that as well. However, I fear that one of the major obstacles will be that the collaborators are not actually sitting together, and even more importantly: disconnected from distractions. The reason those students get so far in their projects, is because they cannot do anything else but work on that (from early in the morning till… early in the next morning). The same goes for biohackathons, although we don’t make such long hours there.
So my comment is: for real high productivity it’s best to lock the collaborators up in a room.

By: Pedro Beltrao

Pedro Beltrao — Thu, 07 Feb 2008 22:52:32 +0000

I am repeating a bit what I said in a comment to Deepak’s post but what I think is the biggest limitation right now is awareness. How to get the right people to join an open project even for a short period of time. It could work for those problems where there is already a sufficient number of people with the right expertise tuned in to respond to a call.

By: Jason Kelly

Jason Kelly — Wed, 06 Feb 2008 15:28:13 +0000

I really like the "bursty work" idea. I wouldn't rule out experimental work though. As an example, the iGEM competition is approximately what you are describing. Teams of 4-10 people, ~6 months, and a fixed deadline (the competition date). It's largely experimental work, and it's amazing how much gets done in an iGEM project relative to a PhD student working alone for 6 months. Many of the projects end up being written up in scientific journals as well. Some of the teams also use OpenWetWare to host collaborative open lab notebooks, and I think the wiki does a pretty good job providing the shared space you were describing in your post. So maybe we're not too far off from bursty work! The major problem, IMO, is that PhD programs aren't set up to encourage and reward collaborative activity -- it's all about the individual, which frustrates these efforts I think.

By: Pawel Szczesny

Pawel Szczesny — Mon, 04 Feb 2008 17:45:40 +0000

Thank you, Jean-Claude, Cameron. I’m very honored to meet you both on my blog .

Lets see how it works – I’ll try to post some more on the topic, and if time permits, make an attempt with a small project. Cameron, you are right about size/specificity of the project – I’m not aiming here at the most important/most difficult stuff around, but at these small things taking 90% of time of in-silico scientists – exactly these which are often passed on to students (not all of us have this privilege). Definition of the problem is not often an issue – I believe in many cases the execution is a problem by itself.

By: Cameron Neylon

Cameron Neylon — Mon, 04 Feb 2008 13:19:13 +0000

There is also a resource and dependency issue. You need to have little packets of people’s time available at the right moment to make this kind of thing work. Unless you have highly defined and independent tasks it is difficult to get the big time benefits. It could work for a well defined very specific problem, the kind of things we do in undergraduate projects perhaps?

By: Jean-Claude Bradley

Jean-Claude Bradley — Thu, 31 Jan 2008 20:33:13 +0000

Ludovic,
Thanks for the mention of UsefulChem. While it is true that we collaborate openly, the time crunch point that Pawel mentions is an element that we don’t often have. That is probably one of the most useful properties of proposals – they have hard deadlines and force people to finish things. I used that opportunity recently to put some ideas together that would have otherwise taken longer to crystallize. Even if it doesn’t get funded, I can now refer to those ideas on Nature Precedings.
It really is a good idea – the challenge is to make the deadline hard enough.