  X-BASED ABACUS
  ============== Abacus X Widget, V7.3

Primary site: ftp://ftp.tux.org/pub/tux/bagleyd/xabacus
Secondary site: ftp://ibiblio.org/pub/Linux/apps/math
Web page: http://www.tux.org/~bagleyd/abacus.html
Maintainer: David A. Bagley <bagleyd@tux.org>

  It has been tested on the following platforms:
    Sun Ultra 5     SunOS 5.9
    PC  Pentium 4   Linux 2.4.18
    PC  Pentium 4   Windows XP Cygwin
    PC  Pentium 4   Windows XP MinGW
    PC  Pentium 2   Windows 2000 Cygwin
    PC  Pentium 2   Windows 2000 MinGW
    IBM PowerPC     AIX 4.3.3.0

  Previous versions tested on:
    Sun Ultra       SunOS 5.6-5.8
    Sun SPARC       SunOS 5.5.1 (X11R5 & X11R6)
    Sun SPARC       SunOS 4.1.3
    PC  Pentium     Linux 2.0.9
    PC  486         Linux 1.2.13
    DEC VAX & ALPHA VMS 6.1
    IBM PowerPC     AIX 4.3.1.0
    HP  898         HP-UX 10.20
    HP  3xx & 700   HP-UX 9.01
    SGI             IRIX 5.3
    PC  486         Windows 3.1  Borland C++ 3.1
    PC  486         Windows 3.1  Microsoft C++ 1.51
    PC  486         Windows 95   Precompiled on 3.1

  If you have to do any editing to get this to work, please mail me
  the changes.

  NOTE: Installations before 7.0.2 will cause some incompatibility with
  the demo, so make sure you install over any previous versions.
  READPATH is the path of the demo files and sound files.

HOW TO BUILD?

  An ANSI C compiler is required for build.

  To build, it should be easy.  There are 4 types.  You may want to set some
    parameters like READPATH.

    configure:
      Use "configure -help" to see what options you can set using configure
      or parameters in Makefile.in .

      configure ; make

    imake:
      Sometimes this is not setup correctly by the distributor ... i.e. not
      my fault.  You may want to set READPATH in Imakefile .
      If you have Xpm you may want to do this quick edit in the Imakefile:
        Uncomment line "#define XpmLibrary".
      If you have Motif you may want to do this quick edit in the Imakefile:
        Uncomment line "#define XmLibrary".

      xmkmf ; make

    make.com (VMS users only):
      You may want to set parameters in make.com .

      @make

    win/Makefile (MinGW users only):
      You may want to set parameters in Makefile (like
      COPT=-02 -mno-cygwin).

      cd win
      make

      Then run "wabacus".  Your on your own on installing the wabacus.ini
      file and Abacus?.les files.

  Windows users only:
    xabacus.htm is included.  Most of the features are in wabacus with the
    exception of the command line options.

  X users only:
    Then just run "xabacus".  If you do a "make install" "READPATH" must
    be set correctly.

    You should copy Abacus.ad (or Abacusjp.ad, etc) to $(HOME)/Abacus or
    use "make install" to copy it to something like
    /usr/lib/X11/app-defaults/Abacus, especially if the background is
    similar to the beads. Edit this file for your preferences.

    The following installation options are you ran "make install".
    Normal installation usually in /usr/local.
    make install-png:
      This is used to install the "*.abacus.png" images for use with menus
      in Gnome and KDE.
    make install-xpm:
      This is used to install the "abacus.?.xpm" images for use with menus
      in CDE using system files (name changes from ".xpm" to ".pm" when
      installing).  The xpm images are the same the png images except the
      small image is 24x24 instead of 22x22.
    make install-xpm-home:
      This is used to install the images for use with menus in CDE using
      from your home directory.  I have this set up using the desktop and
      menus.  There are probably many ways but I have it set up using
      $HOME/.dt/types/dtfile.dt, $HOME/.dt/types/dtwm.fp, and $HOME/.dt/dtwmrc
      These files can be copied and changed from /usr/dt.

  Essentially, this is a proof by induction that a computer is more
  powerful than an abacus, since a computer program can contain an abacus.
  (But then again, you can simulate a computer within a computer, so what
  does that show).  Actually, with a real abacus, one can move more than
  one row at a time with 10 fingers.  But on the other hand, a real abacus
  does not have the current sum displayed as an integer.

  Try resizing the abacus. Notice the shape and size change of the beads.

  Try editing the file Abacus.ad to get a certain feature and then copying
  to $(HOME)/Abacus (where Abacus is the file name) and running xabacus
  again.

  Keep in mind that the Abacus is a widget that can be simply stuck in
  other X programs. (If you find a use for this, let me know).

  It has a built-in demo-mode that teaches the user:
  - how to count (from 1 to 20) using the abacus
  - how to perform simple additions
  - how to perform simple subtractions.
  - how to multiply... works best on Lee's Abacus

  In addition, there is a rudimentary facility for generating data that
  may be used to write customized scripts for teaching other concepts.

  The PostScript document, Abacus.ps, included in the distribution,
  describes every possible technique for performing additions and
  subtractions using an abacus.

  Refer to the man page for detailed command line options.

FOUND A BUG?

  Send bugs reports and fixes to the author. For bugs, please include
  as much information as to their location as possible, since I may not
  be able to duplicate them.

  TODO:
    1. The calculator portion does not work for big numbers, i.e. where
      one can type in math operations and see the abacus perform.
      Also abacus base should not have to equal display base for
      automatic calculations.
    2. When displayBase != base calculate with strings instead of
      "long long".  I.e. the number of rails can be large if displayBase =
      base but may error out if this is not true for large number of rails.
    3. Also allow for locale to allow "," for "." and vice versa.
    4. Redesign so that bead gets kicked over across rail or slot so that
      the gap between the beads is not necessary.
    5. Turning anomaly option on and off, the value should remain constant
      and the beads should change where necessary.  Allow automatic
      calculations with anomalies.

HISTORY

  [Feb 23, 2007]  V7.3:
    Html no longer corrupted.
    mono and reverseVideo update.
    Named fractions added for Roman numerals using 12ths and subdeck.
  [Feb 20, 2007]  V7.2.8:
    Roman Hand Abacus fix for subdeck part 1/3 of 1/12 should really be
      1/12 of 1/12 thanks to Ray Greaves <rayg AT base12.plus.com>.
  [Feb 20, 2007]  V7.2.7:
    Setting up Roman Hand Abacus from .ad file caused core.
    Fix for setting up twelfths and quarters from .ini file.
    New bump.au sound, db lowered, and bumpdat.au removed, thanks to
      http://audacity.sourceforge.net to help clean the sound up.
    Setting decimal a little off for vertical.
  [Oct 11, 2006]  V7.2.6:
    Option to format output to use commas so numbers can be read easier,
      shortcut use "g" for group.
    Fixed an overflow if number entered in display is too large.
    Fixed menu to assure the ok part of queries is handled correctly.
    Now use http://netpbm.sourceforge.net to generate images from xpm.
  [Aug 08, 2006]  V7.2.5:
    Roman subdeck to Russian error fixed (ffftb.04f).
    Roman subdeck error fixed (ffftb.04b).
    Roman fix for pieces and piece percents.
    Russian to Danish fix for twelfths.
    Added more debugging.
    Fix for decreasing number of rails with special rails by slider.
  [Jul 21, 2006]  V7.2.4:
    Fixed assorted memory leaks and uninitialized memory access reads.
  [May 14, 2006]  V7.2.3:
    Tan frame.
    Danish Elementary School abacus added (teaching aid).
      Do not know how its used, setup as:
        clears when pushed left
        lowest order beads at bottom
        kept simple so decimal point is at bottom
  [Mar 06, 2006]  V7.2.2:
    Fixed bug for signs and quarter percents and subdecks.
    Changed definition of CARRY buffer from 2 to 1, fixed display bug.
    Anomaly now works when carries are possible.  Although there is no
      historical abacus like this, its nice to remove this requirement.
    Pieces and piece percents now use "+" on bar except Roman Abacus.
    Right click clear fix (off by one).
    Made usage string compatible with versions of Motif and screen.
    Fixed a problem for quarter percents where beads would get stuck.
    Windows fix for printing of subdeck chars in Roman Abacus for Museum
      of the Thermae in Rome.
  [Feb 27, 2006]  V7.2.1:
    rngs.h left out of distribution, also some permission problems
      on installation scripts.
    Simplified code to switch to and from Russian abacus.
    Redesigned rounded beads to always account for slots.  I.e. a bead is
      either an overlay of 2 circles and a rectangle if a wide bead or an
      overlay of 4 circles and two rectangles (a cross) where there would
      always be a flat edge along the width of a rail or slot.
  [Jan 17, 2006]  V7.2:
    Added on 3 subdecks with 4 beads.
      ./xabacus -roman -rails 10 -tpiece 2 -bpiece 6 -subdeck 3 -subbead 4
      1/2,1/4,1/3 ounces column now work for a Roman Abacus.  This
      comes in two flavors, where possibly the simpler one, needs 3 decks
      (where a single bottom column would be) (one known example in
      museum in British Museum in London).  In the other flavor, the
      beads could stop midway on a single slot (in IMHO not a well defined
      way) (two known examples one in Museum of the Thermae in Rome and
      one in Cabinet des medailles, Bibliotheque nationale, Paris).
      Type "e" or use Museum option to cycle through museum pieces.
    Added older Roman Signs "|x| (((|))) ((|)) (|) c x |" on bar.
      Also allow these signs to be Roman Numerals.
    If anomaly, calculations are not allowed.
    Switching from/to Russian with a negative was broken.
    Added more sanity checks for configuration, or it could go into a
      weird state.
    Fixed for small beads and pc version, would not have rounded slots.
    Fixed top level room and spaces if top pieces or piecePercents
      greater than 2.
  [Jan 05, 2006]  V7.1.8:
    There were some cases where resizing the number of rails causes
      the calculation to change unnecessarily.
    Signed overflow fixes e.g. 50000000000*2=-5000 or
      -100000000000 Chinese converted to Japanese.
    Assorted tweaks discovered when porting code to Java version.
    Moved Anomaly from "a" to "l" to be consistent with Java version.
    Move bar problems with quarter percent.
    Quarter percent for Russian format would give an X error, fixed.
  [Jan 01, 2006]  V7.1.7:
    Sign, quarter, and quarterPercent code all rewritten to be less buggy.
    Quarter percents option for 1/4 Kopeks.
    Use of "piece" instead of "quarter" for more variability.  This now
      allows a "twelfths" capability (Roman ounces) for Roman Hand Abacus.
      Its pretty configurable (with the configuration file) so you can have
      a range from halves to sixteenths (stock market pricing, pieces of
      eight, etc), and possibly beyond if you want to get crazy.
    New symbols on the bar for "-" for sign, "O" for pieces.  If the base and
      piece and format in Russian style set weird it could otherwise be hard
      hard to read.  O used because it was used on the Roman Hand Abacus.
    Anomaly option for Mesoamerican Nepohualtzintzin Abacus.  Basically a
      JP Abacus in base 20 (3/4) where the columns left of the 2nd digit are
      multiples of 360 (instead of 400).  4th column of unit beads would be
      20x360.  Bar marked with an "X" where the anomaly is.  Note: this
      is turned off if there are extra beads as in the Chinese mode or
      Korean mode or this leads to even more unexpected anomalies.
    AnomalySq option to make more Anomaly more configurable, now can have a
      representation of time using hours, minutes, seconds.  Babylonians are
      responsible for that mess.  Though they probably did not use an
      abacus.
  [Nov 19, 2005]  V7.1.6:
    Vertical for proper Russian schoty (rotate xabacus clockwise 90 degrees).
      Use "v" or from the command line -vertical.
    The first bead thats of each 3 decimal places >= 1000 gets a darker color
      on Russian Abacus.
  [Oct 05, 2005]  V7.1.5:
    -version added
    Common option help for X and X-Motif
  [May 31, 2005]  V7.1.4:
    Separated out compatible components.
  [Nov 21, 2004]  V7.1.3:
    Bug fixes for auto-calculation and small base values.
    Thanks to Debian maintainer Florian Ernst <florian AT uni-hd.de> for
      esound fix.
  [Oct 22, 2004]  V7.1.2:
    Small changes synchronizing with Java code.
    Fixed goof introduced in previous which make quarter not work right.
  [Sep 25, 2004]  V7.1.1:
    Installation of sound files.
    bufferBead and colormap not properly initialized and could cause core
      dumps.
  [Aug 23, 2004]  V7.1:
    Thanks to Debian maintainer Florian Ernst <florian AT uni-hd.de> for
      pointing out that the demo breaks without Motif.  Fixed this and 'f'
      should now work better for X.
    Beads are now double buffered.  Faster and looks better on slow machines.
    Beads now slide in small increments.
    Got rid of turning the beads "inside-out" when pressed.  Too annoying.
    "<", ">" to speed up and slow down movement of the beads.  Actually,
      it changes "delay" inversely.
    Got rid of some drawing errors when beads are long.
    Sound added.
    Sort of real time using getttimeofday.
  [May 31, 2004]  V7.0.6:
    Demo now works with the other "standard" abaci.  Korean (or Pre-WWII
      Japanese) 1/5 (ko), Japanese 1/4 (or Roman) (jp), or Russian 0/10
      (ru).
  [Jan 23, 2004]  V7.0.5: Not released
    The idea with the calculator stuff is to see the beads
      move with input from the display line.  This will do a calculation
      but it is not meant as a demonstration of a mathematical operation.
      Uhhh... do not use this for your taxes yet...
      Has some special features... 2v = sqrt of 2, 2u = cbrt of 2, e for e,
        p for pi, ! for factorial (should do a gamma function if real?), and
        ^ for power.  Display Base must equal Abacus Base.  Eventually want
        to do operations on strings so abacus could be an "infinite" size.
      Answers that result in an error default to zero (maybe all the beads
        should fall off?).
  [Jan 13, 2004]  V7.0.4:
    Clear for Lee's Abacus is for each individual deck except for demo.
    Have the decimal point setting accessible to the demo.
      I treat the decimal point as the place setting vernier...
    Resources broken in 7.0.2 fixed.
    Decrementing with the decimal on the right fixed.
    Added the faster diamond shaped beads.  :)
    Color of decimal point the same as beads.
    Put the decimal point (or place setting vernier) at a post.  (To me it
      should be between posts but have not seen an historic example).
    Windows fixes: 'u' for quarters mode and rounder beads
  [Dec 31, 2003]  V7.0.3:
    The Roman mode forces a plate look where the beads move in slots.
    Fixed minor drawing errors in Windows.
    Changed order of additions in the 4th lesson to conform with the
      standard way of doing multiplication.
    Fixed bug where a new 5th lesson not installed.
    Fixed clean on auxiliary decks from demo.
    Fixed demo mode on auxiliary decks.
  [Dec 23, 2003]  V7.0.2:
    Lee's Abacus, default setup when using Motif.  Added a 4th lesson
      for multiplication.  Find out more about Lee's Abacus from
      http://www.ee.ryerson.ca/~elf/abacus/leeabacus/
    Type of abacus listed for nonMotif and Windows.
    Demo can now be set dynamically in Windows.
  [Dec 15, 2003]  V7.0.1:
    Increment when decimal point left most bug fix.
    The Hide functionality of the Escape key now iconifies or drops in
      task bar.
    Added help in Windows.
    Accelerator Keys (F keys and regular keys) in Windows now work.
  [Nov 08, 2003]  V7.0:
    Abacus??.ad* files removed, info is in Abacus.ad*.
    Integrated winpuz6.5 (Windows 3.1/95) into X source tree.
  [Aug 28, 2003]  V5.7.3:
    png for Gnome and KDE menus and xpm for CDE menus.
    Added xpm for icon.
    Reset a few colors for consistency.
  [Aug 14, 2003]  V5.7.2:
    man 1 to man 6 and DESTDIR added and suggestions for ad.in stuff
      thanks to Stanislav Brabec <sbrabec AT suse.cz>
    The demo should work if the demo files are in the installation directory
      (the default but can be changed by -demopath) or the current
      directory.
    Bug fix for xabacus -japanese -demo and a few involving "quarters".
    User interface now free of numbers for formats.
  [Aug 11, 2003]  V5.7.1:
    When changing base certain combinations would create too many beads
      in top deck, beads are instead set to zero.
    Reformatted with option menus.
    Help kicked over.
    Added Korean and Roman.  Roman is essentially the same as the Japanese
      version right now.
  [Jul 07, 2003]  V5.7:
    Sync up with xpuzzles
  [Apr 29, 2003]  V5.6.2: Not released
    Added callback for calculate... does not really do too much yet.
      Need to #define CALC to try...
    Bug fix when in Demo and Russian format is pressed
    Clear now checks if you really want to clear using left mouse button,
      'c' will not query
    'Esc' will hide abacus (actually osfCancel did the trick)
    Help & About expanded and pop-ups added
    Took out bitmaps since it takes too much room
    Demo can be activated from within window without needing command-line
      -demo option.  Removed leak in loadFont.  Added 'o' functionality.
  [Mar 21, 2003]  V5.6.1:
    Added runtime options -chinese, -japanese, and -russian
    Fixed shading if beads are black
    Installation fixes for Motif and X versions
    Help & About expanded and pop-ups added
    Roman Numerals added
    Fixed uninstall on Cygwin
    Took out toggles and bitmaps since it takes too much room.
      Now compile-time option.  There is real help and one can change
      the toggles through the menu.
  [Dec 16, 2002]  V5.6:
    configure if around 2nd AC_CHECK_FUNC
    removed Sun CC (char *) warnings
    removed hard coding of program name
    Cygwin changes ./, -lSM -lICE
    x and xm files combined using HAVE_MOTIF
    port back from Java (on the Java side I got some help from
        Sarat Chandran <saratcmahadevan AT yahoo.com>)
      colors changes
      frame color change on enter/leave
      background/foreground set by default
      3d beads
      reset maximum
      pixel width
      "," diamonds
      added signs and quarters (with smaller delay when moved)
      Russian abacus (schoty) (it fell over)
  [Sep 01, 2001]  V5.5.4: g++ 3.0 warnings removed.  Simplified format.
    Added demo error checking.
  [Aug 11, 1999]  V5.5.2: Fixed runtime library problems when using configure,
    added text to README and xabacus.man, fixed confusing demo start.
  [Jul 31, 1999]  V5.5.1: Removed HP usleep warning.
  [Jun 20, 1999]  V5.5: Combined the independent work of Luis Fernandes'
    xabacus 1.00, also started in 1991.   This includes all the features
    to teach the users how to use the abacus (-demo) and a more solid
    frame (its sturdier :) ).  -bars changed to -rails to conform with
    the terminology.  Abacus now defaults to the top beads up.  The
    picture on the cover of my abacus instruction booklet was deceiving.
    Abacus.ps and Abacus?.les are written entirely by Luis Fernandes
    <elf AT ee.ryerson.ca>.  (Also see http://www.ee.ryerson.ca/~elf/abacus).
    Added -rv and -mono options.
    Implemented -base option, can now go up to base 36, yeah I got carried
    away here (it should even work with EBCDIC though untested (really
    carried away)).
    Dynamic number of rails, with calculations saved with each change.
  [Jan 01, 1997]  V5.4: configure, man page updates.
  [Apr 22, 1996]  V5.3: You can now see beads move one unit space at a time.
    This is governed by a new resource, "delay".
  [Feb 12, 1996]  V5.2.1: beads now invert when pressed, wprintf name may
    conflict with other UNIX libraries - changed to
    motif_printf, keyboard may not work with motif - fixed.
  [Jan 31, 1996]  V5.2: Dynamic number of rails (well, the widget is dynamic
    but xabacus.c limits it to around 300 for the title text),
    fixed another bug when resized (negative radius circles).
  [Dec 15, 1995]  V5.1: Fixed a bug when resized.
  [Sep 30, 1995]  V5.0: Xt/Motif, your choice.
  [May 16, 1995]  V4.10: Warnings removed from Sun's cc and lint.
  [Mar 13, 1995]  V4.3: Removed lint warnings and added a VMS make.com .
  [Nov 08, 1994]  V4.2: Removed gcc -Wall warnings.
  [Jun 03, 1994]  V4.1: R6.
  [May 07, 1994]  V4.0: Xt version.
  [Feb 03, 1993]  V3.0: Motif version.
    I got some good ideas from Douglas A. Young's book:
    "The X Window System Programming and Applications with
    Xt OSF/Motif Edition", particularly his dial widget.
    I got some good ideas on presentation from Q. Zhao's
    tetris.
  [Dec 17, 1991]  V2.0: XView version.
  [Feb 14, 1991]  V1.0: SunView version.
