Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!pacific.mps.ohio-state.edu!verber
From: verber@pacific.mps.ohio-state.edu (Mark A. Verber)
Newsgroups: comp.text
Subject: Re: Which dvi->PS driver is best for including PostScript?
Keywords: dvi PostScript include figures driver
Message-ID: <958@pacific.mps.ohio-state.edu>
Date: 27 Oct 89 03:06:29 GMT
References: <11100@phoenix.Princeton.EDU>
Reply-To: verber@pacific.mps.ohio-state.edu (Mark A. Verber)
Distribution: na
Organization: Ohio State University, Physics Department
Lines: 101

There are a variety of dvi-->ps.  Which is best?  Good questions.

(i) by Larry Denenberg, available from june.cs.washington.edu by ftp,
under the name   tex/ld_dvi2ps.tar.Z
 
Main features: support for builtin fonts, psfig, TeX relection commands,
reasonable directory organization for fonts and environment variables,  
single pages, different printer resolutions. PXL or GF fonts.
 
Problems: The  psfile  special was disabled when the support for 
psfig was added. No PK fonts.   

(ii) by Peter Damron, available from june.cs.washington.edu by ftp,
under the name   tex/pdvi2ps.tar.Z 
 
Main features: Improved font substitution. Well-documented code.
Supports psfile specials and macintosh graphics. PXL or GF fonts.  
Problems: No builtin fonts, no support for psfig.  
 
(iii) by Jing-bai Wang, available from june.cs.washington.edu by ftp 
under the name    tmp/dvi2ps.tar.Z 
 
Main features: PK or PXL fonts. psfig. builtin fonts. some control of
virtual memory. PostScript scaling of substituted TeX fonts. On-line help.
Full page specials, psfile.
 
Problems: Rigid directory organization of fonts, poor search for substitutions.
Help program probably should be separate. Reimplementation of psfile looks
dangerous, but I haven't tested it.
 
(iv) by Van Jacobson. Available by ftp from lbl, but I don't remember the
precise address right now. It was posted fairly recently in either comp.text
or comp.lang.postscript.
 
Main features: psfig, builtin fonts, PK or PXL fonts, virtual memory control.
Problems: Poor search of font directories, no psfile special.
 
(v) by Greg Lee. He will email the shar files. Send him a message at
        lee@uhccux.uhcc.hawaii.edu
 
Main features: PK or PXL files, builtin fonts. More VM control. Command 
line arguments processed by getopt. Extensive use of relative movement
in PostScript, and of widthshow whenever possible to optimize PS part
of the process.
 
Problems: No psfig.
 
(vi) by Piet van Oostrum. Available from oddjob.uchicago.edu by ftp
under the name  dvi2ps.2.45.tar.Z

I just got a copy of this one, and haven't looked at it. 
Has correct positioning algorithm and improved selection of pages
to print. PXL or PK fonts, VM control.

(vii) dvips by Tomas Rokicki.  from labrea.stanford.edu in pub directory.

This is currently my favorite:

Single version which uses PK, GF, or PXL fonts (specified at compile time),
uses builtin fonts, supports the psfile special, psfig, and macintosh
graphics, has a flexible font directory organization and a reasonable search
procedure with PostScript scaling of substituted TeX fonts, some control
of VM, the additional specials introduced by Wang, and some enhanced
page specification options.

According to Tom Rokicki:

This is the latest and greatest version of dvips.  It supports:

	* Automatic generation of fonts via MF whenever needed
	- EPSF PostScript files complete with automatic determination
		of graphics size (see epsf.tex)
	* Literal PostScript includes complete with definitions
	- Non-EPSF PostScript graphics with user-specified scaling
	- TPIC specials
	* Correct budgeting of printer memory
	- Output at 300 or 400 DPI (or any other DPI)
	- Use of PostScript fonts
	* Use of popen on output file.
	* Optional generation of `compressed' output---greatly
		reduces final output file size and VM requirements.
	* Reduced VM requirements even without compression

(The starred features are unique to my implementation; no other translator
that I am aware of has these features.)  In addition, mine meets all the
standards that many others do not, such as use of the standard maxdrift
algorithm, even with PostScript fonts, and generation of a correct set of
ligatures and kerns for PostScript fonts.

And there are many, many other features, such as collated copies in
addition to uncollated copies, and EPSF-2.0 conformant output, etc.

Mine is virgin code, though, not derived at all from the more global
dvi2ps, so it will be harder to `integrate'.


-- 
Mark A. Verber
System Programmer, Physics Department, Ohio State University
verber@pacific.mps.ohio-state.edu
(614) 292-8002