The previous calculation used an approximation which was only valid in
cases where the means being compared were similar; this resulted in very
odd claims being made, e.g. that 0 +/- 0 is a difference of -100% +/- 1%
from 100 +/- 1.
The new calculation scales sample standard deviations by the means, and
yields approximately correct percentage difference bounds providing that
the reference population is bounded away from zero. (In the case where
the values being compared are not sufficiently bounded away from zero,
the distribution of ratios becomes much harder to calculate, and is not
likely to be useful anyway.)
Note that when ministat is used for its intended purpose of determining
whether two samples are statistically different, this change is unlikely
to have any noticeable effect; in such cases the means will be similar
enough that the correction applied here will be minimal.
values when the data set has an even number of elements.
PR: 201582
Submitted by: Marcus Reid <marcus@blazingdot.com>
Reviewed by: imp
Approved by: bapt (mentor)
Off by default, build behaves normally.
WITH_META_MODE we get auto objdir creation, the ability to
start build from anywhere in the tree.
Still need to add real targets under targets/ to build packages.
Differential Revision: D2796
Reviewed by: brooks imp
The man page states that:
'-w width Width of ASCII-art plot in characters, default is 74.'
This is not entirely correct. The mini-help is more accurate:
'-w : width of graph/test output (default 74 or terminal width)'
In other words: the man page fails to explain that ministat will default
to the terminal width, not 74. It will only fall back to 74 if stdout is
not a TTY.
Submitted by: Ben Hutchings <ben@decadent.org.uk>
Approved by: philip (mentor)
Add "-C <column>" and "-d <delims>" options to chop up input lines.
Make '#' a comment character, rest of line is ignored.
Submitted by: Dmitry Morozovsky <marck@rinet.ru>
avg/median/stddev bars onto separate lines for readability if the
ranges overlapped. In 2005, ministat was extended to support more than
2 datasets, but the -s code was not updated. It will coredump if run
with -s and >2 sets.
PR: 82909
Submitted by: Dan Nelson <dnelson@allantgroup.com>
the limit is only the number of meaningful graph symbols available.
Statistical comparison is performed between the first dataset and
any further datasets.
No objection by: phk
of data from benchmarks etc. Implements "Student's t" for various
confidence levels, defaults to 95%.
If your benchmarks are not significant at the 95% confidence level,
we don't want to hear about it.
