[Rd] Re: legend()s text color
Martin Maechler
maechler at stat.math.ethz.ch
Thu Feb 26 09:43:20 MET 2004
>>>>> "Paul" == Paul Murrell <p.murrell at auckland.ac.nz>
>>>>> on Thu, 26 Feb 2004 13:49:23 +1300 writes:
Paul> Hi
Paul> Marc Schwartz wrote:
>> On Wed, 2004-02-25 at 08:02, Uwe Ligges wrote:
>>
>>> Martin Maechler wrote:
>>>
>>>
>>>> {moved from R-help to R-devel}
>>>>
>>>>
>>>>
>>>>>>>>> "UweL" == Uwe Ligges <ligges at statistik.uni-dortmund.de>
>>>>>>>>> on Tue, 24 Feb 2004 19:05:46 +0100 writes:
>>>>>>>>
>>>>
UweL> Sebastien Durand wrote:
>>>> >> Hi,
>>>> >>
>>>> >> Is there a way to change the color of the text inside a legend, let say
>>>> >> I would like to use a black background in my legend, how can I get the
>>>> >> text to show up, it is black!!!
>>>> >> So for example how could I change it to white...?
>>>> >>
>>>> >> Sebastien
>>>>
>>>>
UweL> By setting par(fg = "white"):
>>>>
UweL> plot(1:10)
UweL> par(fg="white")
UweL> legend(2,2, col="white", legend="A", bg="black")
>>>>
>>>> or par(col = "white"),
>>>> as Marc has suggested.
>>>>
>>>> legend()'s source code makes it very straightforward to add an
>>>> argument 'text.col' for this purpose, and I have already done so
>>>> in my development version.
>>>> The question is really if the default should be
>>>> text.col = par("col")
>>>> or text.col = par("fg")
>>>>
>>>> I'd currently tend for par("col") , but the tendency is weak.
>>>>
>>>
>>> I agree, I would expect the lines around the legend's box in par("fg"),
>>> but the font in par("col"). And indeed, that's already the case if both
>>> values have been set in par().
>>>
>>> Uwe
>>
>>
>>
>> Martin,
>>
>> Thanks for taking this on.
>>
>> I'll throw in my vote and support the text.col = par("col") approach.
Paul> There is little consistency to the interpretation of these parameters
Paul> across different functions, but my preference would be to use "col"
Paul> Paul
Thank you Paul (and Marc and Uwe).
legend(...., text.col = par("col"))
is now committed to R-devel.
Martin
More information about the R-devel
mailing list