Research papers as software

I just stumbled upon this post from Daniel Lemire.

Research papers need versioning: the authors should revise their most important work, to fix bugs and improve the presentation. Important research papers should be perfected as much as possible.

This would be oh so nice ! I love that idea, although the implementation sounds totally unrealistic. Like everybody else, I guess, my research have been and is very incremental. We come up with new ideas and hypotheses, and a couple of months (years) later we improve it.  Thus there are two ways you can work: publish a new paper every time you’ve made substantial progress, or wait until the paper is,indeed, as perfect as possible. In which case, eventually, you never publish, and you might be in trouble if you don’t have a tenure position, since the publish or perish mantra is pretty much the major constraint today.

Combine this with the idea I found the other day (I can’t find the reference anymore, though) that you are only allowed to publish 5 or 10 papers in your entire career, so that you have to focus on the most important ideas and most innovative work you can come up with. That would cut the background noise considerably.

Leave a Reply

Please log in using one of these methods to post your comment: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.