Evaristo › Likes

I used to be lower than 20, but I continue to add feeds, and am sharing the same %. - Louis Gray

Put a 100 cap a day on there, and we will have some serious issues. Don't limit me. :) - Louis Gray

heh. no one can put a cap on anyone but themselves. I want it for myself only. - Jenna Bilotta

submitted to Open Lab 09 - Bora Zivkovic

Shirley - this is an excellent essay on this topic, which is very close to my heart. Well done! - Peter Binfield

Shirley, excellent (LONG) post. Can't wait to read your books ;) - Ricardo Vidal

Thanks! This took me a couple days pretty much full-time to write (the luxury of not working a job right now); how do people with jobs do this?? - Shirley Wu

Tom Roud (?) writes an interesting rebuttal against article-level metrics - or at least cautions against some of the metrics that might be used (such as blog and media coverage). It's in French but I used Google translator to read it. My memory of high school French classes only allowed me to decipher the first paragraph, and get the gist that it was an argument against. http://tomroud.com/2009... - Shirley Wu

The post on tomroud.com is interesting. Both your response on his post and the comment by Mitch on your post address most of his concerns, methinks. Popularity contests are no good, but there are ways around them (Mitch), they are just one of many metrics to be used with caution (PLoS), and GlamourMagz are also popularity contests where quirky papers have no chance (you). But his thesis... more... - Bora Zivkovic

@shirley: how do we do it? In pieces over many days... - Björn Brembs

or get a job where writing this stuff is what you are supposed to be doing ;-) - Bora Zivkovic

Having read the piece, I actually have some less vacuous comments: 1. To my knowledge, Garfield introduced the IF to help librarians cut subscriptions, not for scientists to help them chose publishing venues? 2. As you point out, journal level metrics are mathematically inadequate for what they are used for now. However, Thomson's IF specifically is worthless because it is negotiable... more... - Björn Brembs

ONS would be more like: '(work, release)+, finish, publish'. - Egon Willighagen

Egon++ - Rajarshi Guha from iPhone

Ideally, publishing and releasing would be the same thing? - Duncan Hull

Egon :) - Jean-Claude Bradley

Yes publishing is making public. I am assuming you mean peer-review...? - Frank from iPhone

@Duncan: It would be great if Publishing and Releasing were the same thing (for example, the Journal of Machine Learning Research has done some things to tie the two together), but a Journal article condenses the operational details leaving enough to understand and reproduce the experiment. I think what I was trying to say is that we shouldn't be satisfied with human-readable summaries. The job of communicating results should include the release of machine-readable data & code separate from the paper. - Dan Gezelter

+1 Egon. - Bill Hooker

Egon nailed it - Deepak Singh

Love this: "“How can I possibly review this paper if I can’t see the code they were using?" -- I routinely suggest to authors that their paper would be stronger if it included raw data. I'm thinking about getting bolshier about it: "Four replicates, you say? 'Essentially identical' you say? OK, let's see the data." - Bill Hooker

Even if you see the four replicates there is still the question of the replicates that were run but not included. There is nothing wrong with excluding replicates - it is just that leaving out the "undesirable" runs leaves out a huge aspect how a particular experiment works in practice, especially the ways it can fail. - Jean-Claude Bradley