Re: Adding an aggregate for variance

[Jose E. Marchesi]

Hi John.

> Much as I think this change is of a generally good quality, I wonder
> if it really belongs in recutils.
>
> Recutils is a database, and not a statistical analysis tool.  It could
> be the start of a slippery slope ... what would come next  ... ?
> Sample Variance? Sample Std. Dev?  Population Std. Dev?  Skewness?
> Kurtosis?  Covariance?  Correlation Coefficients? Chi-square  ....
>
> There are other GNU tools which do these calculations.  Rather than
> pretending that recutils is a statistical analysis tool, I think it
> would be a better idea to implement easier ways to interface recutils
> to the tools which are specifically designed for this job.
>
>
> Just my $0.02

I totally agree with you in that recutils is not a statistical analysis
tool.

Still, I see nothing wrong in having support for simple aggregates like
average, count, variance and the like.  In their simplicitly, they can
be certainly useful for cases where using a full-fledged statistical
analysis tool would be overkill.

It is a bit like the encryption support... %confidential is not really
up to the task of providing good (or even decent) encryption
capabilities for anything that really should be confidential at all
cost, but it is certainly good enough for situations where we want to
hide the contents of recfiles to cursory looks :)