This is release 1.3 of the GNU plotutils (plotting utilities) package,
including GNU libplot: a function library for 2-D device-independent vector
graphics.  The file CONTENTS lists the contents of the package.  Please
send bug reports and feedback on the package (suggestions for improvements,
etc.) to the principal author and current maintainer, Robert Maier
<rsm@math.arizona.edu>.

For information on installing this software on your system, first glance
through the generic installation instructions in the file INSTALL.  On most
systems, the installation process is simple.  You build the package by
typing `./configure' and then `make'.  To install the package, you then
type `make install'.  A documentation file, in `info' format, is produced
and installed as part of this process.  You may also produce a copy of the
documentation for the package, in dvi format, by typing `make dvi'.  This
assumes that you have the TeX document formatter installed.

By default, the installation process installs the file `plot.h' in
/usr/include, a system directory that some people do not like to install
files in.  To avoid this, you may wish to begin by doing 
`./configure --oldincludedir=""' instead of `./configure'.

On Silicon Graphics systems running Irix 6.x, when you build the package
you may possibly get messages about `GP overflow'.  If so, you should do
`make CFLAGS="-O -multigot"' instead of just `make'.

One common problem related to the X Window System should be mentioned.
Some machines have the X Window System, but provide Motif widgets instead
of Athena widgets.  On such machines (e.g. Hewlett-Packard workstations
running HP/UX, and some IBM workstations) you should do 
`./configure --with-motif' instead of just `./configure' to ensure that 
the software is configured correctly.

Note that many systems with Motif also have Athena widgets, but have them
in nonstandard places.  On such systems, `./configure --with-motif' may be
the best thing to do.

Older Hewlett-Packard workstations, running HP/UX 9.xx or earlier, require
special treatment.  They have Motif widgets, but the location of the Motif
files must be specified explicitly.  On such machines you should do

	make CFLAGS="-O -I/usr/include/Motif1.2 -L/usr/lib/Motif1.2"

rather than just `make'.

Some older Sun workstations [running SunOS 4.1.3, with OpenWindows and cc
rather than gcc] have a different X-related problem.  On such machines, you
may need to do

	make XLIBS="-lXaw -Bstatic -lXmu -Bdynamic -lXt -lXext -lX11"

rather than just `make', to work around a problem with undefined symbols in
the X Window System libraries.  This assumes that your `make' program
allows the values of variables to be specified on the command line.

If you simply can't successfully build the programs that use the X Window
System, you should do `./configure --without-x' instead of `./configure'.

To make maximum use of the X Window System support in this package, you
will need to have the 35 standard Postscript fonts installed on your X
server.  For instructions on installing Type 1 versions of these fonts,
clones of which are included in the package, see the file ./README.fonts.

Also, to make maximum use of the support for the `xfig' drawing editor, you
need to make a small change to the source for `xfig', and recompile it.
(The change makes it use the full set of standard Postscript fonts.)  For
this too, see the file ./README.fonts.

By default, this package includes support for drawing Japanese Kanji
(ideographic characters).  A library of 603 Kanji is available; for
details, see the file ./doc/kanji.doc.  Including this library will add
approximately 100 kilobytes to the length of each relevant binary.  If you
wish to omit Kanji support, you should compile the package by doing 
`make CFLAGS="-O -DNO_KANJI"' instead of just `make'.  If this is done, 
you will still be able to use Japanese Hiragana and Katakana 
(syllabic characters).  See the file ./doc/kana.doc.

The INSTALL file recommends that you do `make check' to run validation
tests on the plotting utilities.  Validation tests on NeXT machines have
been known to fail for more or less innocuous reasons (incorrect formatting
of floating-point numbers).  If any of the validation tests should fail on
any other system, please report it.
