On 16 Sep 2009, at 12:59, Andrew Treloar wrote:
> 1. Often it is necessary to get from the reduction to the original
> (in order to verify that data points are real and not outliers, for
> instance)
Absolutely. Surely this is one of the key drivers for e-research
(whatever that is) - publicly accountable analysis and review and
reinterpretation and refinement...
> 2. Increasingly those summarisations will be 3D renderings or movies
> of time series visualisations. Not just text, and definitely not
> fractionally megabyty! (probably not a word, but I couldn't resist)
Mostly megabyty as opposed to predominantly petabyty?
--
Les
|