[Rd] (PR#11509) rgb to cmyk conversion is wrong in

Jeffrey J. Hallman jhallman at frb.gov
Tue May 27 16:51:08 CEST 2008


ripley at stats.ox.ac.uk writes:

> Please do study the FAQ and posting guide: R 2.7.0 is not the current 
> sources, and this was changed a month ago in R-devel.  You were asked to 
> check for such changes before submitting a report.

I thought fixes were also applied to R-patched, which you can see I did check.
I also searched the bug reports and the release notes for R-devel.  I follow
the r-devel mailing list and did not see this bug mentioned there either.  Nor
did I find anything on CRAN's R site search.

> Also, there is no unique way to do undercover removal in conversion to 
> CMYK, and so it is wrong to call any method 'wrong'

More accurately, it is wrong to call any method "right", as they are all
"wrong".  But some are more wrong than others, and the code I reported
was very wrong.  Some of the CMYK numbers it produced were outside the
[0,1] closed interval, which is clearly an error according to PostScript
documentation.  And both the printed output and PDFs produced via ps2pdf
looked horrible.  When I substituted CMYK numbers produced by my R function
into the .ps file, the colors were, in fact, very close to the ones produced
using the default rgb colormodel.

> -- in fact the one used in earlier versions of R was a documented method and
> within the family described in the PDF reference manual.

OK, maybe it's just because I'm a bit annoyed by your tone, but why is deleted
code from an earlier version of R relevant here?  It doesn't make my bug
report any less valid.

> -- 
> Brian D. Ripley,                  ripley at stats.ox.ac.uk
> Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
> University of Oxford,             Tel:  +44 1865 272861 (self)
> 1 South Parks Road,                     +44 1865 272866 (PA)
> Oxford OX1 3TG, UK                Fax:  +44 1865 272595
>
> ______________________________________________
> R-devel at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel
>

-- 
Jeff



More information about the R-devel mailing list