On 8/4/07, Fernando Perez <[EMAIL PROTECTED]> wrote: > Just a minor comment... >
> This entire argument, which is critical to the notice and probably to > much of SAGE's motivation, has been heavily defended and discussed > already in the literature. To my knowledge, the first to push really > hard in this direction were Jon Claerbout and D. Donoho, both at > Stanford. From Donoho's wavelab at > http://www-stat.stanford.edu/~wavelab : In addition to Bioconductor, I forgot to add the links to the 'children' of Claerbout's original SEP effort, which has certainly borne lots of fruit. The Magadascar project http://rsf.sourceforge.net/Main_Page is squarely aimed at the problem of reproducible research in the geosciences. Their workflow is based on Python and SCons, and this page has lots of references and details on the whole issue of reproducible research: http://rsf.sourceforge.net/Reproducible_computational_experiments_using_SCons I don't know who funds Magadascar, but the related effort at http://www.geodynamics.org/ is a large, multi-year, well-funded by the NSF project to establish an open source computational infrastructure for the geosciences. One of the key institutions there is CACR at Caltech: http://www.cacr.caltech.edu/[EMAIL PROTECTED]/project.cfm?ID=9 and CACR is a big Python shop. The won the award (from the NSF, if I recall correctly) for the DANSE project: http://www.cacr.caltech.edu/research/project.cfm?ID=27 is the data analysis platform for the Spallation Neutron Source, the largest scientific project built in the USA in recent years (it's a $1.6 billion toy at ORNL). DANSE is fully open source: http://wiki.cacr.caltech.edu/danse/index.php/Trac_access MADNESS is a DOE Scidac project: http://code.google.com/p/m-a-d-n-e-s-s/ http://www.scidac.gov/matchem/petachem.html which is GPL licensed from the get go. Hopefully this provides some evidence that yes, the NIH (re. Bioconductor), the NSF and the DOE are already deep in the business of funding, to the tune of quite a few million a year, the development of fully open source scientific computing platforms where the questions of research reproducibility are at the forefront of the discussion and not an afterthought. Whether any commercial vendor is happy or not about it is besides the point: much of the computational developments in the applied fields is already fully engaged in this effort. The difference with your notice is that most of these projects have a specific 'scientific' goal, and that tends to be easier to 'sell' to funding agencies. Getting funding for something which in some ways is 'pure infrastructure' such as SAGE (or the underlying building blocks, say numpy or scipy from the Python side) is often difficult. I've always been of the opinion that this is pure and unfortunate short-sightedness from the funding agencies, since I think that a few well-invested dollars (well, a few million) in such key projects would easily pay for themselves many times over. But nobody is asking me how to run the NSF... ;) I hope this is useful to you as ammo in what I think is a very important debate. Regards, f --~--~---------~--~----~------------~-------~--~----~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~----------~----~----~----~------~----~------~--~---
