diff --git a/tools/unicode/c/readme.txt b/tools/unicode/c/readme.txt
new file mode 100644
index 0000000000..c1e02a4178
--- /dev/null
+++ b/tools/unicode/c/readme.txt
@@ -0,0 +1,32 @@
+# Copyright (C) 2010, International Business Machines
+# Corporation and others.  All Rights Reserved.
+#
+# created on: 2010jun04
+# created by: Markus W. Scherer
+
+These tools parse Unicode Character Database files and generate
+data files (text, source code and binary) for use in ICU.
+They are used during the Unicode beta period and after a Unicode release.
+
+For a log of actions and changes for recent Unicode version upgrades of ICU, see
+http://bugs.icu-project.org/trac/browser/icu/trunk/source/data/unidata/changes.txt
+
+Since the Unicode 5.2 upgrade, these tools have been moved out of the ICU source
+tree to here. They have not been fully tested and probably need some more work
+and setup. They might benefit from additional shell or Python scripts.
+
+There are autoconf makefiles (Makefile.in) and Visual C++ project files (.vcproj)
+in the subfolders. They are copied over from the ICU source tree and will not
+work without modifications. However, I started to use CMake (CMakeLists.txt)
+which is much simpler, and if it works well enough then I plan to just
+delete the old makefiles and project files. The CMake files will currently
+work only on Linux, just because I hardcoded the ICU library filenames
+(e.g., libicuuc.so).
+I should use more variables to make the CMake files more portable, and should
+use ICU's installed icu-config or Makefile.inc to get the values for these
+variables.
+(If and when ICU itself uses CMake, we should be able to point to its modules.)
+
+Things will improve as I work on Unicode 6...
+
+markus