netterfield at astro.utoronto.ca
Mon Jul 12 23:59:45 CEST 2004
I think this may be a good way to go in general (I haven't looked at the code)
because of variable numbers of parameters on some fits (ie, polynomial).
BUT: we need to find a way to get at the parameters for putting in labels (ie,
slope and intercept for a line, or sigma for a fit, etc.).
i) some fits return scalars as well as scalar lists
+ self documenting (ie, scalar names can say what it is)
- not general
+ doesn't require changes to kst.
- requires changes to plugins
ii) we introduce syntax to access an element of a scalar
- Not obvious to user what parameter element 1 is (slope?, intercept?)
+ general solution for all fits
- requires work in kst UI
+ doesn't require changes to existing plugins
iii) add capability to fit plugin system to somehow produce a default label
+ easy to use when it is what you want
- not flexable (can't change default label)
- significant work
My feeling is we should
-keep the changes Andrew has introduced.
-Do option (i) for the fits where it makes sense.
-Do (ii) soon but not now.
-Think about how to add (iii).
On July 12, 2004 05:18 pm, Andrew Walker wrote:
> CVS commit by arwalker:
> Don't do some of the usual vector thing with parameter and covariance
> results from a fit.
> M +6 -4 kstfitdialog_i.cpp 1.17
> M +28 -5 kstplugin.cpp 1.53
> M +2 -1 kstplugindialog_i.cpp 1.75
> M +82 -55 kstvector.cpp 1.71
> M +8 -1 kstvector.h 1.47
> M +1 -1 kstviewfitsdialog_i.cpp 1.7
> M +3 -2 plugin.h 1.21
> M +4 -1 pluginxmlparser.cpp 1.18
> M +8 -6 vectorselector.ui.h 1.7
> Kst mailing list
> Kst at kde.org
More information about the Kst