Nature has published on software in the computational sciences. The column is entitled ‘Publish your computer code: it is good enough‘ and is itself a good read. There is an accompanying article (or is it the other way round?!) entitled ‘Computational science… Error‘.
Reading the comments and other blog responses (eg Anthony Fejes, Bruce Berriman) shows the considerable degree of consensus on the problem and form of solutions. Publishing code is critical, but so are the ongoing activities of further development, documentation, community building, etc. It also demonstrates yet again the inter-relatedness of openness, sustainability and preservation as concepts.
Our project perspective is very closely aligned to all that has been said. Both software engineering and openness support software preservation. We would like to see the good practice described become common practice.