Re: Max Size Datafile in 10g

pagesflames_at_usa.net (Dusan Bolek) wrote in message news:<1e8276d6.0409150508.6e7947e1_at_posting.google.com>...
> "Howard J. Rogers" <hjr_at_dizwell.com> wrote in message news:<4147d493$0$23893$afc38c87_at_news.optusnet.com.au>...
> > > Cern, for instance, will generate around 10 Petabytes of info a year
> > > when they fire the accelerator up. It could run for 10-20 years.
> > It was a good answer. But I have to say, a little cynically, I doubt the
> > world will be a better place for all this data.
> >
> > Seems to me that in the rush to capture everything, we've forgotten how to
> > precis or discriminate between the worthless and the priceless.
> >
> > But that's a topic for another day.
> >
> > Thanks for the link, by the way.
> >
> > Regards
> > HJR
>
> Just one small remark about CERN to point out that sometimes huge data
> doesn't mean worthless heap of everything. I had a small discussion
> with the guy working for CERN on this collider business and if I
> understand him correctly (probably not) the reason for this huge
> storage is this:
>
> The very principle of research with colliders is to bounce various
> particles against each other and waiting what is going to happen. It
> can be described as some odd kind of lottery, because almost all
> collisions are not winning (e.g. nothing interesting happened) and
> only just one in millions can be a catch. The problem is that you
> can't tell if something happened or not just by "looking" at. So the
> way how to deal with this is to store data about all collisions and
> then use complicated program analysing certain patterns to get the
> winning ticket. The probability of getting the game winner is
> increasing with increasing of data available for analysis. So they not
> only want to store really huge amount of data, but the also need them.

Ironically, the patterns they look for tend to be like "right-handed spirals emanating from a certain area," which is, of course, the essence of visualization - human visual pattern recognition is a very parallelized operation. Whether some analogy of that (like neural computing) or something else will be useful to sort through the data remains to be, uh, seen.

jg

--
@home.com is bogus.
And so, my undergraduate degree and Oracle will merge sometime after
my lifespan.