TeX - LaTeX Stack Exchange is a question and answer site for users of TeX, LaTeX, ConTeXt, and related typesetting systems. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I thought PDFLaTeX was simpler and eliminated useless steps as well as possible font issues?

In other words what is the difference between a pdf generated using the LaTeX command and DVI's and the one generated using the PDFLaTeX command?

EDIT #1: As an example, for the submission of a thesis, my university requires the .tex file and all the associated files which are required to compile it and a .dvi file.

It's in French but here it is anyways:


Forme du mémoire ou de la thèse

Lorsque le mémoire ou la thèse a été rédigé à l'aide de LaTeX, il faut acheminer à la Faculté des études supérieures le fichier .tex avec tout autre document nécessaire à sa compilation (images, fichiers de bibliographie, fichiers d'index, etc.) et le fichier DVI. Pour acheminer le mémoire ou la thèse, il est recommandé d'utiliser un format de fichier compressé.

From the official webpage.

EDIT #2: Further research shows that Jukka Suomela's answer is right. Here is what is listed on the webpage of a conference I am writing for (right after the part that explains how to compile successfully through DVI-PS-PDF):

Another alternative is to use the pdflatex (pdftex) program instead of straight LaTeX or TeX. This program avoids the Type 3 font problem, however you must ensure that all of the fonts are embedded (the pdffonts utility mentioned above will yield this information). If they are not, you need to configure pdftex to use a font map file that specifies that the fonts be embedded. Also you should ensure that images are not downsampled or otherwise compressed in a lossy way. If you are knowledgeable about how pdflatex deals with included images and how to ensure that they are not compressed or downsampled, please email us at submissions@vgtc.org so that we might improve our support for this program.

share|improve this question
Can you point us to a place where this requirement is given? – Joseph Wright Aug 19 '10 at 20:29
Thanks for the edit. Sounds to me very much like something written before pdfTeX was invented, so they just assumed that there must be a DVI file. – Joseph Wright Aug 19 '10 at 20:44
If you, or anyone, is about to advise the conference organisers, or the library, about what they should be requiring here, then there should probably be at least some mention of PDF/A. There's a mention of that elsewhere, and also on stackoverflow. – Norman Gray Aug 20 '10 at 8:57
@Joseph: I still don't see why they couldn't use a postscript file, though I suppose PDF makes it a lot easier to verify that none of the fonts are Type 3. (DVI, of course, doesn't include the fonts in the first place.) – SamB Sep 16 '10 at 21:02
I wonder if it's because DVIs are smaller (as they lack fonts and graphics)? All seems very mysterious. – Joseph Wright Sep 16 '10 at 21:04
up vote 14 down vote accepted

In practice, in situations like this, there isn't a good reason.

The instructions are written 10 years ago. By people who haven't really used Latex since 1990s.

To give a stupid example, many ACM conferences require that you submit the PostScript file produced by dvips. The reason for this is that earlier people didn't know how to use latex + dvips + ps2pdf to produce a valid PDF file with all fonts embedded (including those used in EPS figures). Of course, nowadays we could simply use PDF illustrations + pdflatex.

share|improve this answer
Probably, hence my question above! – Joseph Wright Aug 19 '10 at 20:29
So I take it that if the conference just requires a pdf file, I shouldn't mind any talk about DVI whatsoever and just build it the way I prefer? Is there any difference about the way images are embedded (if I use, say, eps and png images with eps2pdf). – levesque Aug 19 '10 at 20:44
@JCL: There is a difference for raster images; if you convert them to EPS and the EPS to PDF, ghostscript will (by default) do JPEG compression on them, and if they were already JPEG compressed, it will decompress them (this can't be reconfigured; there's a note on the Ghostcript "projects" page about this), losing data. – SamB Sep 16 '10 at 21:09
@JCL: To summarize, if you use pdfLaTeX, it's better to just include the raster images as such (or use pts.szit.bme.hu/sam2p to convert straight to PDF). – SamB Sep 16 '10 at 21:12

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.