[tex-k] header special extension proposal

Paul Vojta vojta at Math.Berkeley.EDU
Tue Aug 23 07:39:51 CEST 2005

On Fri, Aug 19, 2005 at 09:18:01AM +0200, Hendri Adriaens wrote:
> > 	perl fixeps.pl < figure.ps > figure.pro
> > 	tex test1
> > 	dvips -o test1.ps test1
> Ok, tried that, see below.

[bug description omitted]

Fixed.  The first step should now be:

	perl fixeps.pl figure.ps > figure.pro

(i.e., no '<').  As usual, the files are located in:


> Although it is a nice alternative to modifying dvips, the obvious remark
> would be that this requires an additional step with an external program that
> most users won't know about even where to get it...

Anyone concerned with this issue is probably informed, or at least motivated.

Also, it's not just about modifying dvips, it's also about modifying
previewers, etc.  And, for dvips, it is crossing a threshold that has not
been crossed before, namely modifying input files.  I feel very strongly
that one should hold the line on adding features, otherwise we end up
with bloated software.  See "cat -v considered harmful":


And, what would it gain us?  dvips has been around for some 20 years,
and I haven't heard of this issue coming up until recently.  (Of course,
Tom would know better than I.)  But, usually TeX is used for typesetting
books and articles, which rarely have repeated figures.

The original e-mail mentions presentation packages as a possible application.
However, the web page


surveys TeX-based presentation packages.  The vast majority are pdf-based,
and none using PS are immediately apparent.  The two presentation packages
that I regard to be the most popular, namely prosper and latex-beamer,
use pdf.

This brings me to my last point.  If someone is concerned about file size,
then why aren't they using pdf?  Pdf is a more compact format (40% smaller
in one case that I checked).  There are times when you have to use
PostScript (e.g., for printers), but the days of slow serial links to
printers are long gone.

So I hope *I* don't sound unfriendly in this, and I do appreciate your
efforts in this direction.  I'm just not convinced.


Paul Vojta

More information about the tex-k mailing list