RELEASE NOTES

THE RAMPAGE IMAGE PROCESSING ARCHITECTURE
(c) 1997 Mayur Patel


---------------------------------------------------------------------------
version 1.0
9/97

RAMPAGE was written with the GNU C++ compiler v2.7.3 under Debian Linux 1.1.
The Makefiles are configured for the GNU compiler.  "make" will produce a debug
version.  "make release" will produce optimized object code.

You may get warnings "comparison is always 0 due to limited range of data type"
or something similar.  This is okay.  It appears because of a macro technique I
am using.

RAMPAGE has no external library dependencies and I have a strong commitment to
keep it this way.  Future development will likely require STL.  Multi-threaded
execution will likely use the POSIX thread interface (or pthreads).

#define RpSHOWSTATS if you wish the caching system to print caching information
to stdout during execution.  By default, "make" will print caching information.
"make release" will not.


---------------------------------------------------------------------------
version 0.0
12/96

I presented an early prototype of RAMPAGE in the July 1997 issue of Dr. Dobb's
Journal.  This article may be a helpful introduction to the current version
of RAMPAGE.

A more general discussion of efficient image processing can be found in:
Shantzis, M.  A Model for Efficient and Flexible Image Computing.  Proceedings
of SIGGRAPH 1994, (July 1994).  pp. 147-154.

Silicon Graphics has a commercial product based on tile-based, demand-driven
execution:  ImageVision Library (commonly called IL).  This library ships with
each SGI workstation.
