Mindswap Weblog

Science FOO Camp 2007 (Scifoo 07)

by James Hendler

Given my travel schedule I have no idea when I’ll actually get a chance to do justice to this meeting (probably never), but thought I’d do a quick shot to get some things online-

first, for those who don’t know what it is, Scifoo is a gathering of a relatively random collection of people, mostly scientists and technologists, to a sort of semi-formed meeting, to meet each other and discuss whatever they want (invites from Tim O’Reilly and Timo Hannay). I didn’t know what to expect, figured it might be interesting. Yow was I surprised - there were about 300 people there, ranging from post docs and folks from Google who helped with the logistics, to notables of science (Freeman Dyson, Carl Djerassi, Eric Lander,…), technology (Dean Kamen, Jaron Lanier, Charles Simonyi…), Science Fiction (Neal Stephenson, Kim Robinson, Greg Bear), some leaders of industry (Google represented by Sergei Brin and Larry Page), and lots of others (Martha Stewart, James Randi,…)

Lots of time for informal mixing plus good formal mixing

Lots of discussion was about the future of academic publishing — I led a session Sat evening (about 30 people) on the topic, asked it be kept to a technology focus (as opposed to arguing over tenure or peer review) - part of my newly started “Reinventing the journal” project (more about that when my thoughts are clearer - but see the piece “Science publishing on the Semantic Web” that Tim B-L and I had in Nature some years back).

The session included brainstorming some ideas on what we can/could/should be doing to explore this area — here are the notes from the whiteboard at that meeting (thanks to Richard Akerman for transcribing them and Nikita Bernstein for forwarding them to me) — these need to be written up and commented more - (I am thinking of doing that for an editorial in IEEE IS - which would of course be preprinted in this blog as they usually are.)

1) Downstream tracking / history (esp cross-domain)
- relationships between articles (automated? human?)

2) ID, logoing, review status tag, trust mechanisms
- other peer review status

3) semantic authoring aids

4) Data: process/workflow reuse; digital object provenance
- huge data sets

5) Review as a service that can be applied to an object

6) citation labeling (citation ontology)

7) authoring tools that add value for metadata

8) extensible geotagging / gis / unique ids for any domains

9) community ehanced - facebook for science e.g. Nature Network

10) unique ID for people (w/o necessarily revealing identity)
- persistent ID

11) annotation/metadata for non-text scientific objectat

12) “fantasy genetics” - prediction market for science

btw, this last one is a neat idea from Carl Bergstrom - idea is to allow participants to pick a set of papers from some particular conference, journal, field (or whatever) whcih they think will be the most influential over some period of time — ideas about how to evaluate include using Google Scholar citation counts, or some sort of voting scheme where the voters are different than (or controlled version of) the raters — idea is not to be a big force in the future of science, but a fun way to get a lot of people to think more about what gets published in their field. I think this is a great idea - hoping to get some undergrads from RPI to implement it next term

cheers and to all who were at scifoo - thanks! I had a great time


Leave a Reply

MINDSWAP is a W3C member