[pdftex] pdfcrypt

[Thierry Bouche]

Given that pdftex contains primitives that almost none of his users are
aware of, I find it surprizing that the crypting code be eventually
removed, as it is desactivated by default, and will be usefull in some
obvious cases.

I'm writing this mail because I started the thread and would not like
that what I said therein might be interpreted as an offensive position
against Ricardo's work.

I'm currently working on quite a large digitization project, and doing
quite unusual things with pdftex, among which adding a title page to
bitmap-mode scanned papers. Because most of the papers are copyrighted,
I feel more fair to their publishers & authors to remember this to final
users through protection. Because the bitmaps are heavy and some papers
are 300 pages long I need byte-optimization for web delivery.

It appears that, *for this purpose*, pdftex without encryption is
perfect for adding a nicely formatted page to all my (2,000+) files in a
batch, pdftex+pdfcrypt makes a corrupted file when used in a single run,
pdftex+pdfcrypt does proper encryption when used as a postprocessor over
the previously assembled PDF file. But then linearization is
unavailable. If this feature was not in sight, I must say that pdftex
would have proven the most robust and versatile tool for cooking the
delivered PDF files, as it preserves rigourously the data of each page
included, which most other PDF utilities won't (pstill "optimization"
will convert CCIT G4 pages to ZIP, doubling their size, most tools would
destroy the hidden text, etc.). BTW, I also consider putting the hidden
text from an alternative OCR with rich XML output format instead of the
one I have, and for this too pdftex should be one of the rare available
tools.




-- 
 Thierry Bouche