[Clam-devel] dynamic types

[David García Garzón]

As Xavi says, we all mostly agree with your analysis, Greg. DT have been the 
more persistent over design on CLAM from the very beginning. Pau implemented 
some time ago a simplified TDDed version of dynamic types not depending on 
the Add/Remove hell (CLAM draft repository) but we are even considering to 
drop them all.

The problem here is that some CLAM processings heavily depend on such 
mechanisms and it would be hard to change it all without a heavy rewrite.

The good news: Pau and me are somehow starting such revolution at the 
PluginExample folder. We already have two independent MagPhaseSpectrum and 
ComplexSpectrum classes, FFT/IFFT working with ComplexSpectrums, conversors 
between Complex, MagPhase and the current Spectrum, and a 
MagPhaseSpectrumProduct.


On Monday 16 July 2007 12:49:07 Xavier Amatriain wrote:
> Greg Kellum wrote:
> > Any one have anything to say about this?
>
> I think that the CLAM team mostly agrees with your analysis. Moving away
> from DT has been in
> our Roadmap for some time now but it is a big change that we had
> reserved for the 2.0 ;-) Actually
> before removing DT there are some previous steps like simplifying some
> Data classes like the Spectrum
> you just mentioned.
>
> _______________________________________________
> Clam-devel mailing list
> Clam-devel at llistes.projectes.lafarga.org
> https://llistes.projectes.lafarga.org/cgi-bin/mailman/listinfo/clam-devel