Re: ranges

From: Alberto Micol <Alberto.Micol-at-eso.org>
Date: Fri, 02 May 2003 14:51:46 +0200

Since VO is imagination, let me be imaginative (I just had a beer for lunch:-)

Let me resurrect the idea of histograms instead of ranges.

Providing access to an histogram might give more precise results and even provide the likelihood of success for a given query.

After all a range is just the simplest of all histograms (1 inside the range, 0 outside). It could work for all kind of quantities, wavelengths (1d histo), sky coverage (2d histo), etc.

The implementation is VERY simple, and provides more flexibility.

Example 1: Wavelengths

 if my dataset (e.g.) offers U and J bands I will have an histogram with 2 non-null bins.

Example 2: Wavelength + resolution

 One could offer a lookup table for a combination of wavelengths and wavelength resolutions: a bidimensional histogram would do the job in a simple way.

Plus, as said, the query coiuld return the probability of success.

Alberto Received on 2003-05-02Z14:54:51