README file for the Database Reader

Created by Douglas M. Auclair, dauclair@hotmail.com
Copyright (c) 2000, Douglas M. Auclair under the GPL 
(www.fsf.org/copyleft/gpl.html)

DESCRIPTION
===========
The situation often arises in companies with which I contract that they
have a system that stores data in files, either in binary or in text
format.  The other situation is that their system is broken somehow,
and they need to read or compare that data for a report or for
debugging purposes.  Now.

Dylan to the rescue!

The database-reader reads binary data files (currently) and outputs
that information as text.  To do this, one must configure it
(internally) and recompile.  It assumes the binary data file named on
the command line is in the same directory as itself.

A previous and unavailable version compared two
dissimilarly-structured text data files and output difference of
information. 

Nothing flashy here; just a tool that works well.

INSTALLATION
============
$ gzip -dc database-reader.tar.gz | tar xvf -
$ cd dbreader
$ make

USING database-reader
=====================
* Included are two sample data tables that will work with the 
  database-reader (book.dat and xml.dat).  To have the application scan
  those files, pass in the name of the table, without the extension:

  $ ./database-reader xml

* The app will concatenate ".dat" to the table name and print out the
  table's contents

CONFIGURING database-reader
===========================
"All well and good", I can hear you think, "but how do I get it to read
my data tables?"

Fortunately, this is a simple process:
1. Add your data table's meta information to the *database* table in
   function setup -- note that the #"column names" are the leading 
   integer data of the record.
2. Add a do-read/print method that parses the tail end of the record --
   this is where you add columns that are strings or mixed integers and
   strings (or single characters or bytes or booleans).
3. Recompile and go.

TODO
====
Add floating-point support.

ISSUES
======
As I've worked on mostly big-endian machines, this code assumes the world
is big-endian.  I haven't tried it on a little-endian machine.  For the 
examples this code will work, but your database may have a reversed endian
storage scheme.  If you get really weird numbers, check out read-integer.