I have a friend who is considering moving his research development tools from Matlab to SciPy (and other Python modules). The guy is into image and signal processing. From what I understand from Zope 3, "products" will be much closer to standard Python modules, so that code developed for Zope 3 has a better potential for reuse in other Python apps. That goes the other way around. All of this to say that something like SciPy could benefit a lot from a CMS to store data, algorithmic definitions (in some declarative form maybe ?) and above all system users, who would be granted different data drilling or analysis capabilities as a function of their membership level. I know "killer app" usually refers to something the general public can be hooked on easily and what I am proposing, that is to try and unite the CMS capabilities of Z3 with the computing/analysis capabilities of SciPy (or any other Python product that targets a specific problem domain for that matter), is not a "general public killer app". But it would be a scientific (that is university, research centers ...) vertically integrated killer app and knowing how much skilled resources there are in universities, bringing Zope in research/academic facilities through the "science" door might pave the way to its use on a wider scale.