[Pkg-cups-devel] Bug#658004: cups-pdf: Generate more accessible PDFs

[Martin-Éric Racine]

2014-09-10 12:44 GMT+03:00  <behr at physik.uni-wuerzburg.de>:
> Martin-Éric Racine <martin-eric.racine at iki.fi> wrote on 09/09/2014 22:02:26:
>
>> On Thu, 01 Mar 2012 15:40:43 +0100 "Volker C. Behr"
>> <behr at physik.uni-wuerzburg.de> wrote:
>> > On 03/01/12 14:03, Brian Potkin wrote:
>> > > Followup-For: Bug #658004
>> > >
>> > >
>> > >
> https://bugs.launchpad.net/ubuntu/+source/cups-pdf/+bugs
>
>> > >
>> > > The really interesting discussion is in #820820 but you'll want to
>> > > read
>> > > #942866 too. The essence is that desktop applications output in PDF
>> > > and
>> > > jobs from lp/lpr are converted to PDF by CUPS. The extra processing by
>> > > cups-pdf with Ghostscript may produce a less than optimal PDF.
>> >
>> > I am aware that some desktop applications output in PDF - but not all
>> > (CUPS-PDF is not developed specifically for Linux). So I cannot assume a
>> > PDF workflow. Furthermore, not re-processing the PDF doesn't allow to
>> > change e.g. the PDF version.
>> > Finally, if you already have a PDF you like as native output of your
>> > application - why not simply use that one?
>>
>> I think that the main issue with our pdf2pdf usage is that it tends to
>> systematically convert PDF documents containing text into one gigantic
>> image, thus losing the possibility to copy/paste/search text content. If
>> there is a pdf2pdf option that we can leverage to avoid this loss, then
>> we really should use it by default.
>
> I remain with my view that if your application already supports PDF output -
> why not simply print to a file? Furthermore, one feature of CUPS-PDF is the
> ability to use the full scope of GhostScript to generate custom PDFs -
> simply passing PDF through would hamper that option. Lastly, as stated
> above, CUPS-PDF is meant as a universal UNIX-tool, not just for Linux
> distributions implementing a PDF-workflow.

Volker,

I fully agree with your general idea that if something already is a
PDF, then why not just save it.

The larger issue here is, is there any way to avoid converting PDF
documents that contain text into a large image that can no longer be
searched or copy-pasted?

Martin-Éric