[Rivet] Some make-plots suggestions

Frank Siegert frank.siegert at cern.ch
Thu Jan 31 15:06:49 GMT 2013


On 29 January 2013 15:12, Frank Siegert <frank.siegert at cern.ch> wrote:
>
> Back when I wrote rivet-mkhtml, I added the "--full-range" for each
> call to make-plots. The reason was that the default automatic y range
> is typically too small for LHC plots, it does not even display four
> orders of magnitude (ymin=2e-4*ymax). Do we still consider this our
> best default? I wonder whether we could change this to e.g.
> ymin=2e-7*ymax, in which case it would also make sense to get rid of
> the default --full-range in rivet-mkhtml. If there are any objections
> with respect to backwards compatibility of make-plots, maybe we could
> make this a command line parameter replacing/amending the --full-range
> option?

No opinions? Hendrik, no objections because of compatibility? In that
case I would go ahead with the proposed change.

> On another note, I was wondering how complicated (and desirable) it
> would be to have the legend items right-alignable. I think it would
> make sense because the legend is positioned in the top right corner by
> default, and with a right-aligned legend there would be less overlap
> with the plot. In particular, the coloured lines/errorbands/markers in
> the legend should be on the right-hand side then, which would avoid
> even more the confusion with the actual data. And it's much more
> flexible with respect to long or short legend entries. Hendrik, do you
> think the pstricks for that would be very complicated, or have you
> possibly even implemented something like that before?

I have added this option to make-plots on SVN now with the hopefully
obvious syntax:

  LegendAlign={l,r}

I've left "l" (left) as the default for the time being, but if there's
a consensus for "r" as default then I'd be happy to change that.

Cheers,
Frank


More information about the Rivet mailing list