[open-science] Making science more accountable and efficient

Sat Feb 12 19:56:42 UTC 2011

Hi all,

I just subscribed to the list, so please forgive me if you discussed a 
similar topic already earlier.

I'm regularly doing reviews for journals and conferences in computer 
science and I notice, that still many papers describe approaches and/or 
systems, but the underlying software and data is not (publicly) 
available. From my point of view this substantially hinders 
reproducability, reusability, and peer-reviewing of scientific results 
and thus efficiency and accountability of science in general.

I know there are the Panton Principles, but I think (although its good 
to have them) they are too strict to be applied in the general case for 
the following reasons:

1. its not only about data, also algorithms and their implementations in 
terms of software need to be made available

2. sometimes it is acceptable to have this data, algorithms, software be 
releases publicly, but not under an open license.

I imagine some kind of manifesto, which stresses the importance of 
making scientific artifacts (i.e. data, software, models etc.) 
available, maybe in different gradations similar to the LOD start badges 
[1]. The gradations could be:

0. The scientific artifacts underlying a scientific publication are not 
made available.

1. The scientific artifacts underlying a scientific publication are made 
available in a limited way, which do not yet enable complete 
reproducability.

2. All scientific artifacts underlying a scientific publication are made 
available to the public, they enable complete reproducability but the 
license they are made available under imposes restrictions potentially 
limiting the reuse.

3. Data is published according to the panton principles and all software 
underlying a scientific publication is made available under an 
open-source compatible license.

Once such a manifesto with these gradations is produced, journals and 
conferences can link to it and require submissions to classify 
themselves according to these criteria (and maybe that precedence will 
be given to more works).

Has someone of you seen something like this already? If not do you think 
the creation of such a manifesto is useful and would you be interested 
to work with me on that?

Have a nice Sunday everybody,

Sören

[1] http://lab.linkeddata.deri.ie/2010/lod-badges/