[ESIP-all] FW: Dear Colleague Letter - Data Citation in the Geosciences

[Joe Hourcle]

On Mar 29, 2012, at 5:26 PM, Brian Wee wrote:

> From: National Science Foundation Update [nsf-update at nsf.gov]
> Sent: Thursday, March 29, 2012 4:09 PM
> To: Brian Wee
> Subject: Dear Colleague Letter - Data Citation in the Geosciences



Last week, I had a poster / handout / filler talk (done really last minute as 2 of the 3 speakers for my session dropped out at the last minute, so mostly expanded/cribbed from the poster) at the Research Data Access & Preservation summit on a recommendation for what I called 'Landing Pages', but in talking with other people, it's got a few different roles depending on how you look at it:

Basically:

	1. We need a record of the data publication.  

	   Think of that page in every major book that tells you who the contributors were (author/editor/photographer/layout/illustrator/etc.), gives the title, date/place of publication, ISBN number, printing history, etc.  This page should have sufficient information to be able to generate a citation.

	2. We need a consistent target for data citation DOIs.

	   That target should (a) persist for the long term (even if the data is moved/modified/removed/etc.), (b) be updated if knowledge of the data changes, (c) link to the data or otherwise tell you how to obtain it, (d) give you sufficient documentation/software/whatever to be able to use the data.


The problem w/ citation is that it's really a pain in the ass to cite something when the 'something' is really, really ambiguous.[1]  The groups distributing the data should tell us how they want us to cite the data, and give us nice little targets to point to.  They even can give them multiple names/identities if they want.

For more info, see:

	http://docs.virtualsolar.org/wiki/Citation

(and I have some other stuff that I've written up on the flight back, but the back gate closes in 10 minutes, so I'm going home now)


[1] Wynholds, 'Linking to Scientific Data: Identity Problems of Unruly and Poorly Bounded Digital Objects', http://www.ijdc.net/index.php/ijdc/article/view/174


-Joe