If I understand you well this means that you export the whole
Inkscape document to PNG and then transform it to PDF with OpenOffice
Impress (you could do the same with the Gimp or ImageMagick's convert
actually). If this is what you do, you could keep it in PNG instead
of turning it to a PDF, it won't change a thing.
PDF is a vector based format and PNG is a raster format:
http://en.wikipedia.org/wiki/Vector_Graphics
http://en.wikipedia.org/wiki/Raster_Graphics
The advantage of exporting to PDF from Inkscape's SVG (which is also
a vector format) is that you keep the vector information (i.e. when
you zoom in, the lines stay sharp). If you export the SVG to a raster
format (PNG here), no matter in which format you convert the PNG
afterwards, you won't get the vector information back, you will still
end up with just a bunch of pixels.
In fact, the raster/vector difference is complicated by the fact that
you can include some raster images in a vector based format. For
example you can include some raster images in an Inkscape document.
And what you do in your PDF conversion is exactly that also: you just
include a raster image (a bunch of pixels) in a vector document (the
PDF), but that does not turn it into vector in any way.
I hope this helps. Maybe the FAQ in Inkscape's wiki need to be
expanded a bit about this:
http://wiki.inkscape.org/wiki/index.php/FAQ#What_is_vector_graphics.3F
Cheers,
JiHO
Yes, well not exactly... I'm conscious of the differences of Raster and
Vector graphics. I only export those elements I want to fine tune. If
anyone uses Linux, then using Scribus is another possibility, which
means a better PDF exporter than that of Inkscape's native one, which
then means no need for GIMP, and direct manipulation of the SVG within
Scribus itself, as it supports the format (I ignore if inkscape
specifics are lost, like blur effects and the like) when manipulating
directly the SVG within scribus. Another way to do this, is to export
image portions to a separate raster format