A bunch more doc fixes.

[dunshire.git] / dunshire / options.py

diff --git a/dunshire/options.py b/dunshire/options.py
index c6e5c4cb079edfc7149ee8179137fb7f61975e84..a66abb7bc626f9793ec21b622b3b719ad5aa38cc 100644 (file)
--- a/dunshire/options.py
+++ b/dunshire/options.py
@@ -9,20 +9,22 @@ ABS_TOL = 1e-6
 """
 The absolute tolerance used in all "are these numbers equal" and "is
 this number less than (or equal to) that other number" tests. The CVXOPT
-default is ``1e-7``, and we need a little bit of padding on top of that.
+default is ``1e-7``, but loosening that a little reduces the number of
+"unknown" solutions that we get during random testing. Whether or not it
+improves the solubility of real problems is a question for the
+philosophers.
 """
 
-
-FLOAT_FORMAT = '%.7f'
+DEBUG_FLOAT_FORMAT = '%.20f'
 """
-The default output format for floating point numbers. It has been
-chosen to match the ``ABS_TOL``, in the sense that if ``x != y``,
-then they would not appear different when printed.
+The float output format to use when something goes wrong. If we need to
+reproduce a random test case, for example, then we need all of the digits
+of the things involved. If we try to recreate the problem using only,
+say, the first seven digits of each number, then the resulting game
+might not reproduce the failure.
 """
 
-
-VERBOSE = False
+FLOAT_FORMAT = '%.7f'
 """
-The underlying CVXOPT solver can output debug information, and I
-guess, so could we. But we don't want to because it's ugly.
+The default output format for floating point numbers.
 """