Skip to main content

All things to all users

What is scientific computing? This may seem a strange question for me to pose in the leader column of a magazine that has flourished for many years by providing articles on precisely this topic for its readers. If we don't know what scientific computing is, who would?


Well, it is of course a rhetorical question, but one prompted by the striking contrasts in the application of computing to science that the articles in this issue of the magazine demonstrate.


There is first the matter of scale (which, incidentally, poses the question of 'what is a computer?', let alone 'what is scientific computing?'). The Grid figures constantly now in scientific computational talk. In this issue of the magazine, we have a feature article discussing the UK's efforts to develop both a Grid infrastructure and to encourage industrial participation, while European efforts to develop the Grid are reported on the news page.


With the development of the Grid, not only do computers become global but the very idea of 'a computer' is called into question. It is no longer a material box that one can point to in some specific physical location; rather it becomes a diffuse network. There are, of course, physical machines still there, but the ones contributing to the Grid will shift and change by the second. The Grid becomes the computational equivalent of that Ancient Greek philosophical conundrum that one can never step into the same river twice - except that this river will gird the Earth.


In stark contrast, also in this issue, Felix Grant describes the experience of reviewing some of the latest scientific computing software from a canoe in a swamp in an impoverished country far from mains electricity, let alone international computational networks. And yet, the sophistication of the scientific computing - involving data mining to sift the tiniest signal of the source of environmental pollution from a mass of noisy measurements - is such that the successful conclusion to the project would have been impossible, even a decade ago.

So the answer to my question is that there is no single answer. Scientific computing is many things, depending on user and use.

Dr Tom Wilkie
Dr Tom Wilkie

Media Partners