[SciPy-User] Subsecond frequencies in Scikits.timeseries

[Fernando Perez]

On Tue, Aug 25, 2009 at 12:44 PM, Pierre GM<pgmdevlist at gmail.com> wrote:
>> As far as I know it is still in the development branch at
>> https://code.launchpad.net/~nipy-developers/nipy/trunk-timeseries
>
> Thx ! I'll get a look ASIC.

Great.  I should add that we have every interest in possibly
integrating with the scikit in the future.  But our needs were driven
by the analysis of experimental data, and there are several groups at
UC Berkeley working with code of this nature. We looked at the scikit
carefully, and decided that (considering the fact that we have
research work that needed this code to be functional in a reasonable
timeframe) the impedance mismatch between our needs and the scikit was
large enough to justify a separate effort.  And I think it has paid
off, because by working from scratch, we've been able to address some
design problems in a way that I'm actually happy with, and this would
have been harder to do within an existing codebase.

But, we certainly would love it if this can lead eventually to an
integrated TS codebase that fits everyone's needs, it's just that our
resources weren't sufficient to try to do everything in one shot.

I should also mention that in the next couple of weeks (I'm in
semi-vacation away from my office, Ariel and the others at Berkeley
working on these related problems) we intend to  make this code
available on its own, outside of the nipy trunk tree.  This will make
it  easier to install/use by anyone who doesn't need the nipy core
code (only those who want to use actual nipy-specifiic features would
need nipy, not everyone looking for the timeseries stuff).  We'll make
sure to post about that here so anyone interested can track it with
less dependencies.

Regards,

f