Data types in the index

[Henrik Nordberg <henrikn@math.lbl.gov>]

Dear Challenged,

Right now the index is using ints and floats (single). Does anyone think
that this will be limiting in the future? There could be two reasons for
that, as I see it:
1) more precision is needed because the numbers are just too large to fit
into single precision
2) more precision is needed because that precision is needed when asking
queries. (e.g., 1.0 < prop < 1.000000000000000000001)

The first reason I think is unlikely to be valid, because the whole range
has to be large, otherwise we could just switch units.
As for 2), I don't know. If you ask such a query now, the numbers will be
truncated to single precision.

Note that this only applies to the index. The data stored in Objy could
still be double or long double, of course. 

Regards, Henrik
___________________________________________________________
Henrik Nordberg <HNordberg@lbl.gov> http://nordberg.lbl.gov
Scientific Data Management Research & Development Group
Lawrence Berkeley Laboratory