Le mardi 13 mars 2007 à 22:09 -0500, Arnaldo Leon a écrit :
> Gnumeric is great, but I think I think it badly needs the following
> (in order of importance):

Some work has been done on a time series analysis plugin which should
solve 1 and 4. I just need some time to finish it (the original author
has stopped working on it).

> 1) a built-in linear/spline interpolation function.
> 
> 2) 2-D interpolation (bi-linear and bi-cubic)
> 
> 3) a transfer function/linear filter. This could work like
> "=filter(numer,denom,data,time)" where "numer" and "denom" could point
> to single cells or 1-D arrays and represent your transfer function,
> "data" points to the data you want to filter, and "time" could be
> either a time step or an array of time values. The output would be an
> array of the same size as your input "data" (via CTRL-SHIFT-ENTER)
> 
> 4) Fast-Fourier transforms!
> 
> 5) eigenvalue and SVD matrix decompositions
> 
> Interpolation and filtering is something engineers do all the time. I
> think this could be a huge win for bringing in more engineers to use
> gnumeric. I used to work for an aerospace company and you won't
> believe the amount of stuff that gets designed solely on Excel.
> 
> Also - I don't know if the number of data points on a plot is limited
> (in Excel the max is 10,000), but it would be great if we could plot
> millions of points on a chart, just like matlab can.

in gnumeric, the only limit is the number of rows in the sheet. Millions
of points might take some time to draw.

> If you need a volunteer to code this stuff up, I might be able to find
> some time :)

You are welcome. You might meet the team on irc.gimp.org, channel
#gnumeric, at north america daylight time.

Cheers,
Jean

_______________________________________________
gnumeric-list mailing list
[email protected]
http://mail.gnome.org/mailman/listinfo/gnumeric-list

Reply via email to