Unification and plurality
Posted on January 21, 2003 @ 08:50 in Research
Wired reports on developments in "grid computing" or how it will pretty soon actually be feasible to not only connect super computers and data centers, but to have them work in unison, effectively forming a much bigger 'virtual' computer. Examples benefitting from such computational power are earthquake research, a unified database of mammograms or a simulated 'virtual' universe for astronomers. The authority quoted on grid computing in the article is Dan Abrams, a professor of civil engineering at the University of Illinois at Urbana-Champaign, who thinks that
The assumption is that people will buy into this and we'll be one unified community. [...] We could basically have pseudo labs -- let labs at a distance do our work.
And that will counter some of the troubles besetting science in Abrams' view:
It's all done in an ad hoc manner. [...] Everybody's doing things independently of each other, and some don't share at all. If they do, there's no standard data format.
Even though I can see some clear benefits of hooking computers up in a grid, I'm a bit worried about scientific diversity. What if we end up running experiments within the same simulation models, what if we plough through the same unified data sets? In the end it will become harder and harder to think outside the big distributed box that we've built for ourselves. I firmly believe in standards and don't think we can do without them, but I also think that there is an inherent benefit in plurality. Sometimes you just need to start from completely different premises to get ahead.
Comments and Trackbacks
No comments or trackbacks for this entry yet.
Post a comment
Comments and trackbacks have been closed on this site. My apologies.
Since MT-Blacklist inexplicably stopped working I had no other recourse than close comments and trackbacks to stop the spam. I've been meaning to correct this for quite a while, but life got in the way... in a good way I should add.