This is dvipdfm, a DVI to PDF translator. I wrote this mainly as
exercise to get at the features of PDF I wanted to experiment with. You're
probably wondering why I don't use PDFTeX. I am a bit of a purist and I
would rather use TeX unmodified as Donald Knuth left it.
You can get it or the manual from
or if you prefer FTP:
- TeX \special's that approximate the functionality
of the PostScript pdfmarks used by Adobe's
Acrobat Distiller. Links, outlines, articles, and named
destinations are supported, for example.
- Ability to include PDF, JPEG, and PNG files
as embedded images. For PDF files, only the first
page is included. Resources will be embedded from
the original file as necessary. File inclusion
does not work for PDF files that store the page
contents in several segments in an array.
- Support for thumbnails (with a little help from GhostScript).
- Re-encoding support for PostScript fonts. dvipdfm
uses the same .enc files as dvips. The mapfile is different.
- Virtual font support.
- Support for arbitrary linear graphics transformations.
Any material on the page can be scaled and rotated.
- An internal color stack accessible via \special's.
- Beginning of page (BOP) and end of page (EOP)
\special's for placing arbitrary PDF stream
graphics on every page.
- Partial font embedding and Flate compression
for reduced file size.
- Balanced page tree and dest tree to improve reader
speed on very large documents.
Limitations (at present):
- Contents streams consisting of multiple parts cannot be embedded.
Mark A. Wicks
dvipdfm – A DVI driver to produce PDF directly
The driver offers a wide range of \special commands
(including a colour stack), and supports compression of data
streams, etc. Note that the extended version
dvipdfmx will operate “as
dvipdfm” if necessary.