[R] [OT] 'gv' and fractional points

(Ted Harding) efh at nessie.mcc.ac.uk
Fri Jun 15 18:29:53 CEST 2007


On 15-Jun-07 15:56:58, hadley wickham wrote:
> This doesn't answer your original question, and isn't much help unless
> you're on a mac, but there's a nice looking program that makes this
> kind of graph scraping really easy:
> http://www.arizona-software.ch/applications/graphclick/en/
> 
> Hadley

Thanks, Hadley! But (as you implicitly surmise) I don't use a Mac
(just non-psychedelic Linux).

However, as a follow-up, I've since found that one can (somewhat
tediously) do what I was asking with the GIMP.

If you start the GIMP on a PostScript file, you initially get a
"Load PostScript" window which asks you to choose (amongst other
things" the "resolution", initially "100". If you wind this up
to say "1000", then you get 10 times the positional precision
in the numbers shown as below.

When the image is displayed, there is again a little window
which gives you the GIMP coordinates of the mouse position.
With increased "Resolution", the numerical values vary
correspondingly more rapidly with position.

The tedious aspect (compared with 'gv') is that to get better
visual resolution you need to zoom the image. This enlarges the
whole image, with the result that you have to pan around to
locate different parts, and you can get lost in a graph with
lots of widely-spread points.

With 'gv', on the other hand, you can select any small part
to view in zoom in a sub-window, which is much easier to cope
with; and you also get a "rubber-band" rectangle which enables
you to readily align points here with points there -- e.g. if
you want to read off a curve the y-value corresponding to say
x=5.0.

However, this is at least a partial answer to my own question!

Ted.

--------------------------------------------------------------------
E-Mail: (Ted Harding) <efh at nessie.mcc.ac.uk>
Fax-to-email: +44 (0)870 094 0861
Date: 15-Jun-07                                       Time: 17:29:49
------------------------------ XFMail ------------------------------



More information about the R-help mailing list