Re: [Gnucap-devel] Storing data in Gnucap

[al davis]

I am still catching up on mail after being away for a few 
weeks...

On Thursday 12 April 2007 13:50, Xavier Romeuf wrote:
> However, while looking at the source code of Gnucap, I found
> in s__out.cc the SIM::store() method which is not yet
> implemented. Its comment is "will store data in preparation
> for post processing".
>
> Does it mean that this method would be able to store the
> results of an analysis in memory (and so the results could be
> directly read by a wrapper without using a file)?

Yes.

> If so, how would this method be used? Called from another
> method (SIM::outdata()?) or after using a command like:
> gnucap> store tran v(4) v(5)
> gnucap> transient .00005 .0001 0
> ?

Yes.

> If SIM::store() does so, I can try to implement it...
> A possible way would be to extend the probes contained by
> PROBE_LISTS::store[_mode] by making them to store the
> analysis results. For instance, they could contain a
> std::vector member filled by SIM::store().
> Would it be suitable?

Use a "WAVE" object.  It takes care of things like non-uniform 
spacing and interpolation.  It also has some math operations, 
and will get more.  The intent is to add a full set as needed, 
and some signal processing functions like convolution.

The SIM::store() function should have one loop, with one line in 
the body.  I can do it easily, so I will.  You can use it as an 
example of how it works.  Give me a day or so...