On Thu, Jan 22, 2009 at 3:52 PM, Pavel Kaderavek <kada@xxxxxxxxxxxxx> wrote:
I may try it,
in mf.py are two crucial things:
run of the process
data.jw_dipY[z] = data.calc_dipY_jw(data, params, z)
selection of the process
data.calc_dipY_jw = calc_dipY_S2_jw
so you want to exclude the second one and let there only one general
selection of the model. And the run will be done somehow like:
for (i in xrange data.xy_vect_num)
data.ci=data.ci_dipY[z]
data.jw_dipY[z] = data.calc_jw(data, params)
is it your vision? But it means, that we rewrite one variable by the
another. I do not know if it is correct to do it like this.
I will think about this design, and other alternatives. I need more
time to understand absolutely everything mf_csa.py is doing though.
But I was thinking of a more fundamental design change in the setup of
the 'data' data structure and the calculations using data in it and
placing data back. Don't worry about the complexity though, I can
implement this type of basic infrastructure change.
By the way, I do not want anything to hide, but is it really OK we are
discussing these things here? I feel a bit like a spammer of this mailing
list. (But if you say yes, I will forget about my feelings.)
Lol. There's nothing to worry about. This is exactly the type of
discussion that has been happening on this list since the beginning.
Just have a look around in the archives
(https://mail.gna.org/public/relax-devel/) to see this. These
discussions are very important (essential actually) and it improves
the quality of the code. The discussions here are usually very raw
and to the point.
Regards,
Edward
Re: cst branch development