Re: [boost] [ANN] Boost.Time_series: feedback requested

From: Austin Bingham (abingham_at_[hidden])
Date: 2006-12-18 09:31:13

> I'm afraid I don't follow. What do you mean by a time series 
> "discretized on range"?

This is a bit of me thinking out loud, so I apologize. In some 
time-series for physical processes (i.e. radar returns) it is
sometimes easier or more natural to think about things in terms
of range rather than time, although both are directly correlated
and effectively interchangeable. So the notion was that I could 
choose a discretization value that represented some distance 
rather than some amount of time. 

The more I understand your library, however, the more I see how
this really would break a lot of things. I think my thinking 
really just comes from years of using matrix and vector classes
which didn't force me to think purely in terms of time.

> clip() does *not* copy the data. It returns a clipped view. If you want 
> to copy the data, you should assign the clipped view to another series.

OK, I see that in your documentation now. I had based my 
assumptions on this:

   http://boost-sandbox.sourceforge.net/libs/time_series/doc/html/boost/time_series/clip.html

which doesn't mention views or non-copying. I figured those features 
would be mentioned if they were being used.

--
Austin Bingham
Signal & Information Sciences Laboratory
Applied Research Laboratories, University of Texas at Austin
10000 Burnet Rd., Austin, TX 78758