[et-cg] Feedback on IEEE Article

[John R. Boisseau]

I have only one "cosmic comment," though perhaps mine is really only galactic:

I would have two "cosmic comments"
a) The article uses model extensively and although that can include "statistical models", it is not a common way of describing data analysis which I believe is dominant part of eScience -- nearly all biology falls in this category as does analysis of LHC data. I would balance discussion of compute v data intensive aspects of eScience

I'm always leery of the compute vs data intensive characterizations of computational science. I recognize that there are simulation-based computational methodologies (heavy OUTPUT data, for vis/analysis), vs 'analysis-based' (informatics, data mining, statistical analysis, info vis, etc.) methodologies (often large INPUT data). But, the former has to analyze large amounts of output data, and the latter is often done where the former is not (yet) possible (not a solid enough theoretical foundation) and yet there is a huge amount of data that may yield insights-and eventually reveal the theoretical underpinnings.

So, I'm not disagreeing with Geoffrey at all (simulation will always be important; data analysis methods are growing in importance for areas in which we have more data than theoretical understanding; lots of bio fall is in this modality at present; etc.), just being careful not to let the importance of simulation be diminished, and not to forget that simulation generates lots of data. After all, we're lucky to live in a Universe that is governed by mathematics, and if we only knew every equation... So, yes, balancing the discussion is important!

Jay
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.ogf.org/pipermail/et-cg/attachments/20090323/21dd7715/attachment.html