[lisplab-cvs] r4 - src

Jørn Inge Vestgården jivestgarden at common-lisp.net
Sun Mar 1 11:13:48 UTC 2009


Author: jivestgarden
Date: Sun Mar  1 11:13:48 2009
New Revision: 4

Log:
Minor README documentation

Added:
   Makefile
   README
Removed:
   src/README

Added: Makefile
==============================================================================
--- (empty file)
+++ Makefile	Sun Mar  1 11:13:48 2009
@@ -0,0 +1,14 @@
+
+first:
+	echo "Plase specify target."
+
+touch:
+	touch system/lisplab.asd
+
+lispclean:
+	-find . -name "*.fasl" -execdir rm \{} \; 
+
+clean:	lispclean
+
+distclean:	clean
+	-find . -name "*~" -execdir rm \{} \; 
\ No newline at end of file

Added: README
==============================================================================
--- (empty file)
+++ README	Sun Mar  1 11:13:48 2009
@@ -0,0 +1,47 @@
+INTRODUCTION
+Lisplab is a mathematics library for common lisp.
+Directory structure:
+  src/         Lisplab source code 
+  src/matliap  From the Matlisp project. Rewritten for Lisplab matrices.     
+  shared/      Unmodified code from other projects.
+  doc/         Documentation.
+  system/      Package and asdf-system definitions.
+
+
+NAMING CONVENTION
+The levels in the names of source code is dependency levels:
+level 0:
+  Methods that work on non-index objects,
+  but  the methods here are also typically reimplemented 
+  for matrices.
+  Exampels: copy, .+, .*, ... 
+level 1:
+  Basic matrix/tensor methods for indexing, element reference 
+  and dimensionality. In order to make a new kind of matrix, all
+  mathods on this level must be reimplmented.   
+  Examples: mref, dim, rows, cols, new, ...
+level 2:
+  Basic functionality related to matrices.
+  Examples: mmax, mmap, diag, ...
+level 3:
+  Linear algebra and anything else based on the matrices. 
+  Examples: minv, m*, ...
+
+  
+INSTALLING
+Lisplab is asdf-installable. It has only been tested on SBCL on Linux,
+but should be fairly portable to other platforms, as soon as some minor 
+dependencies of the package sb-ext are resolved. 
+
+The Matlisp linear algebra depends on externally libraries and these must 
+be specified in the variable asdf:*lisplab-external-libraries* before loading,
+as seen in start.lisp. The order of the libraries matter and Blas must be 
+before Lapack. 
+
+On Linux the Blas and Lapack libraries can often be installed by the operating 
+system package system. In Ubuntu, to get the Atlas build of Blas/Lapack, type
+  % aptitude install libatlas3gf-base 
+
+ 
+
+




More information about the lisplab-cvs mailing list