Re: moments.c

[Ben Pfaff]

Jason Stover <address@hidden> writes:

> I'm using moments.c in regression.q and I think there is problem: It
> requires two data passes. If we want PSPP to handle large data sets
> well, it needs to be able to compute moments in a single data pass.
> Does anyone mind if I modify moments.c to do so?

It already does support one-pass moments: use the moments1_*
functions instead of the moments_* functions.

> Second question: moments.c could also be made to compute correlations
> between variables in a single data pass. If changing it to do so
> doesn't make a mess of moments.c, does anyone mind if I make that
> modification, too?

Is there a logical or computational reason why correlations
should be integrated into moments instead of being a separate
module?
-- 
Ben Pfaff 
http://benpfaff.org