Re: problem with fltk

[jacques.beilin]

Hi,

I tried to remove 55130 (MDJ) and the plot is ok.

So it is a real problem for me as I often have to plot earth 
coordionates (for example X=4400000, Y=200000, Z=4400000). I am worried 
because I can imagine that one day Octave may drop gnuplot support.

Best regards,

Jacques


Le 15/12/2011 12:25, marco atzeri a écrit :
> On 12/15/2011 11:45 AM, Michael Goffioul wrote:
> > 2011/12/15 jacques.beilin<address@hidden>:
> > > Hi,
> > >
> > > Here are the scripts. The script compare_fltk_gnuplot.m loads the 
> > > data and
> > > creates the plots using the 2 backends.
> > >
> > > The columns in deltaXYZ.txt are : column 1 : Modified Julian Date, 
> > > Columns
> > > 2,3,4 : X,Y,Z
> > >
> > > I suppose I am doing something wrong but I cannot see what !
> >
> > I don't think you are. My first idea would be that the problem is due
> > to rounding errors in OpenGL (which internally uses single floats, not
> > double).
> >
> > Michael.
>
> it is a rounding error.
> all the data on the first column are very near
>
>  5.51300201e+04 -2.02548504e-02 1.06003433e-02 2.65523735e-02
>  5.51300208e+04 -2.34114490e-02 1.00998580e-02 3.53220068e-02
> ...
>  5.51300889e+04 -6.01677597e-03 -5.80356270e-03 3.66309634e-02
>  5.51300896e+04 -7.13784248e-03 -5.65697253e-03 3.21426634e-02
>
>
> Jaques,
> if from the original data you remove 55130, you will see the expected 
> shape.
>
> Regards
> Marco
>
> _______________________________________________
> Help-octave mailing list
> address@hidden
> https://mailman.cae.wisc.edu/listinfo/help-octave

jacques_beilin.vcf
Description: Vcard