[Rd] minor oddity in pdf() help page

[Roger D. Peng]

Okay, it might be the early morning hour---when I read it a second time it made 
sense.

-roger

Prof Brian Ripley wrote:
> No, it means what it actually says.
> 
> If you include R's PDF in another application, the latter will usually 
> compress *if you asked the application for compressed PDF*.
> 
> On Thu, 2 Mar 2006, Hin-Tak Leung wrote:
> 
>> Roger D. Peng wrote:
>>> The following paragraph from ?pdf struck me as a bit odd:
>>>
>>>       'pdf' writes uncompressed PDF.  It is primarily intended for
>>>       producing PDF graphics for inclusion in other documents, and
>>>       PDF-includers such as 'pdftex' are usually able to handle
>>>       compression.
>>>
>>> Should that be "...and PDF-includers such as 'pdftex' are usually 
>>> _un_able to
>>> handle compression" ?
>>
>> Hmm, I think the documentation is correct but incomplete - pdftex *can*
>> handle compression, but compression is not implemented in R's pdf
>> output device. So it should say:
>>
>> "... PDF-includers such as 'pdftex' are usually able to handle
>> compression, but R's pdf device does not utilise that feature of pdf."
>>
>> (I have checked a pdf generated by R, and it doesn't compress, and I was
>> using pdflatex this morning to include a compressed pdf, so both
>> parts are correct).
>>
>> There is a caveat: the PDF specs (and the postscript language standard)
>> actually defines a few stream compression schemes - LZW and deflate
>> are two I know of from the top of my head, I think there are more.
>> But LZW used to be tangled up with the Unisys patent until recently
>> when the patent expired, so most open-source softwares won't do
>> it. deflate is implemented in zlib and ghostscript-written pdf
>> usually have stream compression on. i.e. For some purposes such
>> as getting smaller pdf's, it may be better to output from R
>> postscript and use ghostscript to do ps2pdf rather than doing
>> it directly from R, and to be pedantic, pdftex can only handle
>> deflate encoded compression, AFAIK, for the reason I outlined above,
>> but it is sufficient for most purposes, since most tools cannot
>> generate LZW-compressed pdf's.
>>
>> HTL
>>
>> ______________________________________________
>> R-devel at r-project.org mailing list
>> https://stat.ethz.ch/mailman/listinfo/r-devel
>>
>>
> 

-- 
Roger D. Peng  |  http://www.biostat.jhsph.edu/~rpeng/