There is another option to produce PDF actually, i.e., using pdftex. But let's omit it in this topic.

I usually do "latex -> dvips -> ps2pdf" to get pdf output. My LaTeX input file contains eps graphics, pstricks codes, and plain text. I have not tried to use "latex -> dvipdfm".

What are the pros and cons pertaining to both options?

2 Answers 2


The pros of one tend to be cons of the other, so I'll just list features. What comes to mind straight away:


  • More widely used (I think); has larger range of support in packages that use its \specials features.
  • The subsequent PostScript file can contain arbitrary PostScript code, which is a programming language; some packages require this feature to do subsequent processing on the output (e.g., pstricks, psfrag)


  • (successor to dvipdfm)
  • Aims to be feature compatible with the PDF features that are accessibly though pdfTeX. Such as PDF annotation, file attachment, and so on.
  • Started out at being better with multilingual/UTF8 processing for some of the Oriental-based TeX systems; I suspect it's required for some CJK packages but I don't know the details.
  • xdvipdfmx is the fork that allows XeTeX to produce PDF output; while Mac OS X has xdv2pdf, on Windows/Linux there's no alternative.

Using latex+dvips+ps2pdf works with pstricks , beamer and powerdot and gives you all the graphics features of xypic; using latex+dvipdfm does not. However the latter approach may give a much smaller file size.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .