On Sun, 09 Feb 2014 23:47:10 +0100, Thomas Neidhart wrote:
On 02/09/2014 11:39 PM, Gilles wrote:
[...]
Another thing to be concerned of: right now we have already *3*
places
in the library where e.g. the GaussNewtonOptimizer resides:
* optimization.general
* optim.nonlinear.vector.jacobian
* fitting.leastsquares
two of them deprecated. This is kind of crazy ;-).
The ML archive will provide ample explanation for the current state.
Well, I know how we arrived at this situation. My question was more
if
we should continue like this.
Another option would be to stop refactoring the 3.x branch and create
an
experimental branch where we do this kind of stuff and make an alpha
release of it to gather feedback.
The ML archive also contains this discussion.
The bottom line is that Evan Ward's proposal intended to satisfy all
current requirements.
Gilles
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
