Now <i>this</i> is something whose time has come.<p>The journal itself is (well be) at <a href="http://www.openresearchcomputation.com/" rel="nofollow">http://www.openresearchcomputation.com/</a> - the articles will also be open-access (I checked that first thing).
I think that the goal of 100% code testing will exclude most research using software not written by professional programmers (i.e. most software in science and engineering research).
Rather than a new journal, it'd be better if established journals included special sections/topics on OSS related to the journal's area.<p>Journal of Machine Learning Research already does this [<a href="http://jmlr.csail.mit.edu/mloss/" rel="nofollow">http://jmlr.csail.mit.edu/mloss/</a>].